Ok, obviously us programmer-NS-fans are a minority, otherwise this blog would have the usual 120 comments instead of 30... That being said, being a minority doesn't mean we should be ignored all of the time. Giving us less attention makes sense, neglecting us completely doesn't. So to everyone else out there, when we get a random technical podcast, please let it slide.
schkorpioI can mspaintJoin Date: 2003-05-23Member: 16635Members
<!--quoteo(post=1666641:date=Jan 8 2008, 04:11 AM:name=Domining)--><div class='quotetop'>QUOTE(Domining @ Jan 8 2008, 04:11 AM) <a href="index.php?act=findpost&pid=1666641"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Its fairly hard to listen to amateurish podcasts, especially when they contain almost nothing interesting. That wouldn't be very exciting, unless you're a programmer I'm sure.<!--QuoteEnd--></div><!--QuoteEEnd--> no one is forcing you to listen to them......
I feel obliged to reduce the negativity by indicating I liked this podcast. Sure, not as exciting as the others, but I enjoy listening to what amounts to a generalization of the development process they're going through.
hotd0g: Guess what? You'll have no idea whatsoever how development is going if they don't do this.
Domining: I'm not sure what constitutes a professional podcast, but I don't see what's wrong with the current one. This isn't a radio show, it's an audio commentary for people who don't want to read paragraphs and paragraphs of jotted down development notes.
Perhaps small news digests should be released with or around the time of the podcasts for people without the time to listen to them. Perhaps people feel obligated to listen if they want to know what's going on, when some of it can be condensed down to bullet points. I personally enjoy listening to the entire audio commentary clip.
Remember, it's called a <b>development</b> podcast/blog. You don't make games by synthesizing how they should work, then magically produce them.
TheAdjHe demanded a cool forum title of some type.Join Date: 2004-05-03Member: 28436Members, NS1 Playtester, Constellation
<!--quoteo(post=1666695:date=Jan 7 2008, 07:50 PM:name=IronFist)--><div class='quotetop'>QUOTE(IronFist @ Jan 7 2008, 07:50 PM) <a href="index.php?act=findpost&pid=1666695"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I feel obliged to reduce the negativity by indicating I liked this podcast. Sure, not as exciting as the others, but I enjoy listening to what amounts to a generalization of the development process they're going through.
hotd0g: Guess what? You'll have no idea whatsoever how development is going if they don't do this.
Domining: I'm not sure what constitutes a professional podcast, but I don't see what's wrong with the current one. This isn't a radio show, it's an audio commentary for people who don't want to read paragraphs and paragraphs of jotted down development notes.
Perhaps small news digests should be released with or around the time of the podcasts for people without the time to listen to them. Perhaps people feel obligated to listen if they want to know what's going on, when some of it can be condensed down to bullet points. I personally enjoy listening to the entire audio commentary clip.
Remember, it's called a <b>development</b> podcast/blog. You don't make games by synthesizing how they should work, then magically produce them.<!--QuoteEnd--></div><!--QuoteEEnd-->
I'd rather see an IRC discussion than a podcast, I really don't feel like having to listen to people talk when I can read it and get it done faster. Wouldnt' be bad for an easy Q&A as well.
Well, he like some of us, is probably just a little frustrated with the lack of information across all "aspects." While they did a god number on topic X last podcast, this time it was on topic B. Saying stuff like "the podcast is ######" won't get you anywhere of course, but I think it's important to acnolidge the frustration from some, since I feel it too. I did enjoy #15 a little, but not as much as others because the information seemed very focused <i>yes</i>, but lacking with regards to where they are at.
I know more now with regards to what they were talking about, but I also feel like I totally missed everything else, like other parts were left out. I still don't know where they are in terms of maps, models, testing, answers to forum topics, and a few others probably. <img src="style_emoticons/<#EMO_DIR#>/sad-fix.gif" style="vertical-align:middle" emoid=":(" border="0" alt="sad-fix.gif" />
I do wish they would still do recaps concerning the rest of the game such as the topics I stated just a second ago [maps, models, and where they are on everything else]. but comments like "the podcast sucks" probably pisses them off a tad. I know I would be a little peeved...
Could you guys maybe do brief recaps or short blurbs on other topics in future podcasts? I was personally curious where you guys were with maps and models progress. Are you making maps as you go, or do you actually have a planned out list or the maps you will be making [i.e. A through Z]? And I don't remember hearing an answer on the customization of models, will you allow it beyond Male/Female? I know I would like to at <i>LEAST</i> be able to change either my face, or something unique! <img src="style_emoticons/<#EMO_DIR#>/tsa.gif" style="vertical-align:middle" emoid="::tsa::" border="0" alt="tsa.gif" />
You make a good point, corps. Noone really knows anything about the bigger picture of NS2, of its development. Personally, my enthusiasm for contributing one or even two maps to its release is starting to wane. They say the words NS2TR, but then let the months pass with no public progress. Apparently there was an alpha, but I heard nothing of it (since im on the mapper mailing list). Feels like its just a matter of time before I give up on NS mapping and find some source mod to contribute to. At least then I'd be working on a more up-to-date engine...
locallyunsceneFeeder of TrollsJoin Date: 2002-12-25Member: 11528Members, Constellation
It was interesting to hear you talk about unit tests. At first I thought you mean regression tests, but I don't think there's any real way to automate regression tests with a game GUI. I just finished doing unit tests at my work and it would be interesting to see the kind of test code you guys would produce. Your test units would probably encompass more code than ours do. We use both UT and regression tests so we test units at a pretty low level. As someone else said, I would UT the more critical areas and not worry about line coverage outside of the areas where the bugs most commonly occur.
<!--quoteo(post=1666721:date=Jan 8 2008, 07:22 AM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Jan 8 2008, 07:22 AM) <a href="index.php?act=findpost&pid=1666721"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->They say the words NS2TR, but then let the months pass with no public progress. Apparently there was an alpha, but I heard nothing of it (since im on the mapper mailing list).<!--QuoteEnd--></div><!--QuoteEEnd--> they said the initial alpha release would only be sent to mappers that have contributed significantly in the past, mappers they know and love, mappers that will make the official maps for NS2. so unless you made or updated maps like eclipse, bast, machina etc, you'll have to wait until the full NS2TR is released, whenever that is.
<!--quoteo(post=1666721:date=Jan 8 2008, 07:22 AM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Jan 8 2008, 07:22 AM) <a href="index.php?act=findpost&pid=1666721"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Feels like its just a matter of time before I give up on NS mapping and find some source mod to contribute to. At least then I'd be working on a more up-to-date engine...<!--QuoteEnd--></div><!--QuoteEEnd--> err... they're using the most latest release of source - ep2. you wont find any source mod out there with a more up-to-date engine than that....
<!--quoteo(post=1666431:date=Jan 6 2008, 12:27 AM:name=MasterPTG)--><div class='quotetop'>QUOTE(MasterPTG @ Jan 6 2008, 12:27 AM) <a href="index.php?act=findpost&pid=1666431"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->To da Devs:
The sooner you get the gameplay done, the sooner the -final- maps are going to be done. #$%# unit testing. You have a huge f'ing testing base. Even if the game is buggy to hell (but fixable), I'm sure you can get peeps to help you for free. GOGOGOGO FFS. Work 80 hours a week TESTING and CODING the f'ing SKILLS and movement and stuff. Slap it into a random CS map and go at it. Whatever, it doesn't even matter. Get the gameplay and skills roughly cut out and slap it into your orange-skinned boxy maps. Then 'take it up a notch', and get the maps done. Seriously, GOGOGO.<!--QuoteEnd--></div><!--QuoteEEnd-->
Even though he's making an ass out of himself, MasterP raises a good point - the core reason your game has endured is the gameplay, and while I have no reason to believe you haven't worked more on that than anything else (since we wouldn't see it yet), if, indeed, all the team has done so far is an emulated release of DI and some (very impressive) art assets, or perhaps the port of the desired NS data structures into Source, then you may consider pulling back and rethinking how things are going along.
The gameplay is the infrastructure, the skeleton of the game. While it's necessary to design a complex system from the top-down before it "evolves" through iteration, it's crucial that you build the "polish and frosting" (eg blast radius vars, welding, functions of DI) <i>around</i> the core gameplay that a player spends 70-90% of their time doing, which is combat oriented, and by this I mean the attack and (perhaps even more so) movement systems.
NS is a deathmatch game. Please maintain that mantra in NS2. I would hate to see such a great fantasy simulation become as trivially point-and-click as CS or TF2 in its re-iteration.
As for the alpha, i meant that we merely heard there was one. Thats it. Regardless of being on the mapper list or not, there's been absolutely nothing for mappers to even think about. And im pretty sure the alpha TR was simply to do some preliminary bugchecks for the full TR.
<!--QuoteBegin-'Buzzou'+--><div class='quotetop'>QUOTE('Buzzou')</div><div class='quotemain'><!--QuoteEBegin-->err... they're using the most latest release of source - ep2. you wont find any source mod out there with a more up-to-date engine than that....<!--QuoteEnd--></div><!--QuoteEEnd-->
Yes, they're using the newest version of Source, but 99% of NS mappers right now are not. We're using HL1. Because thats all we have. There's no point whatsoever in even pretending to map for NS2 on source yet, since the critical numbers and details have not been given. Why make rooms for gameplay that we don't know, with doors possibly too small for the largest creature or DI, with too many rooms or simply the wrong types of rooms given then new gameplay, and with no textures and no props?
Its a much better idea to find some other interesting Source mod and create maps for those, if you want to learn Source mapping for NS2. I began on source so I already know all the bells and whistles, but i'm pretty sure most NS mappers don't.
The problem with this approach is that NS will lose its mappers. At least some will learn to love those other mods, or simply be midway into their new map and not want to abandon it just because the NS2TR finally came out.
Oh, and you mentioned that <!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->they said the initial alpha release would only be sent to mappers [...] that will make the official maps for NS2<!--QuoteEnd--></div><!--QuoteEEnd-->
It sounds like the Alpha went out to like 5 mappers, and I'm sure some of them are simply helping out with the debugging without planning on putting in the huge time requirement for a Source map. The devs have explained that they're letting the community make the maps, and they'll buy the cream of the crop. So no, the majority of the mappers who will be making the official maps are sitting here on the forums (or not) and just watching the months go by. Hell, even the Official Map Pack hasn't gotten a post in about 4 months, despite the fact that it could have been released within a few weeks, 6 months ago.
<!--quoteo(post=1666801:date=Jan 8 2008, 05:41 PM:name=Buzzou)--><div class='quotetop'>QUOTE(Buzzou @ Jan 8 2008, 05:41 PM) <a href="index.php?act=findpost&pid=1666801"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->they're using the most latest release of source - ep2. you wont find any source mod out there with a more up-to-date engine than that....<!--QuoteEnd--></div><!--QuoteEEnd-->
Oh ok! That kinda threw me off. I hear y aknow, "HL source," but then someone always pipes up and says something else like NSTR2, or encoda, or something, so I get confused. You rsaying the game will be made using the Source engine? I fugred it would maybe be old looking because they were using some other older, game engine. I am confused again.... <b>Arg!</b>
<!--quoteo(post=1666804:date=Jan 8 2008, 06:32 PM:name=Radix)--><div class='quotetop'>QUOTE(Radix @ Jan 8 2008, 06:32 PM) <a href="index.php?act=findpost&pid=1666804"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->NS is a deathmatch game. Please maintain that mantra in NS2. I would hate to see such a great fantasy simulation become as trivially point-and-click as CS or TF2 in its re-iteration.<!--QuoteEnd--></div><!--QuoteEEnd-->
NS is more than DM. Any game where it is possible, however remotely so, to actually win the game without killing another player, is not a Death Match. The elements that make up Natural Selection make it far more than simple Death Match. So, if you are going to maintain a mantra, at least make it accurate. NS is a team objective based game: Kill the last hive or kill the Marine IP + CC + entire team... Actually, why are we maintaining that mantra? They are already changing the resource model so they can have a more flexible game play prototyping process ... and they have their key stones / pillars for the game ...
Actually, the formula they put forward reminds me something of a bride getting ready for a supposedly traditional white wedding: "Something old, something new, something borrowed, something blue." I like the mantra(s) they have much better than your mantra. Come to think of it, if all I wanted was Death Match, I would rather be playing Unreal Tournament and its extras to that than NS. Yeah, NS is definitely more than just Death Match. Even Combat is more than just Death Match.
NS:combat is death-matchy I guess... But we're talking about actual NS here. I suppose some of you already know, but I'd still like to point out that unit testing isn't about finding bugs, it's about preventing them. If you ever saw the mantis bug-tracker for NS1, you'll know why this is important.
Having 10,000 testers to find all your bugs doesn't change the fact that it would take too long to fix them all. It's better to not create them in the first place.
<!--quoteo(post=1666827:date=Jan 9 2008, 04:49 AM:name=Revenge)--><div class='quotetop'>QUOTE(Revenge @ Jan 9 2008, 04:49 AM) <a href="index.php?act=findpost&pid=1666827"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->NS:combat is death-matchy I guess... But we're talking about actual NS here. I suppose some of you already know, but I'd still like to point out that unit testing isn't about finding bugs, it's about preventing them. If you ever saw the mantis bug-tracker for NS1, you'll know why this is important.
Having 10,000 testers to find all your bugs doesn't change the fact that it would take too long to fix them all. It's better to not create them in the first place.<!--QuoteEnd--></div><!--QuoteEEnd--> Totally agree.
In addition, some bugs may only be fixed by changing the overal design. And 9 out of 10, you wind up creating a big amount of new bugs.
I would say that the usual bugfixing is a process where the coder evaluates his design. He did not think about some structural design problems in his code and tries to clean his code during the bugfixing process.
Where test driven development makes you think about what you are about to build. You design the application before you write your code. Therefore their "should be" less bugs afterwards.
This week’s podcast was different in that it was about game development in general and covered NS2 only very briefly at the end. It wasn’t very entertaining, but it does provide a view at the game industry from the perspective of a small Indy game development company. I’ve added the times at which new topics start in the podcast. Thanks <a href="http://www.unknownworlds.com/ns2/forums/index.php?showuser=30793" target="_blank">Crispy </a>for your PM with corrections.
Charlie: Welcome to the 15th unknown worlds podcast. Today is Friday January 4th 2008. This is Charlie Max: And Max Charlie: And welcome to the New Year Max: yes Charlie: We did it Max: I don’t know what we did. We made it. Charlie: I don’t know, we got through. Max: That was a close one. We almost didn’t make it Charlie: actually, yeah, that wasn’t too difficult. Max: That’s true, well, no I was thinking it actually is almost true. Charlie: Really? Max: Well I don’t know. Who knows? Charlie: Financially? Max: There is ups and downs. Charlie: That’s true. Well we made it intact. We got back from our families Max: Yeah, we survived that as well. Charlie: Yes. Mom, in case you’re listening, it was awesome. Max: Since it’s the first podcast of the new year, we figure we would do some resolutions. Our game development resolutions.
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> UNIT TESTING (1:13)<!--sizec--></span><!--/sizec-->
Charlie: Right. So we’re gonna talk about a bunch of things that we want to change about ourselves in respect to game development and maybe also what we think the industry should change for this year. It’s kind of a slow week so. ‘Cause we just got back from vacation so indulge us. So some of the things we’re thinking about: well we always bring up the concept of <a href="http://en.wikipedia.org/wiki/Unit_testing" target="_blank">unit testing</a>, which is a big pain in the butt to implement. But basically what that is, is testing every aspect of your code. Often you write a test before you even write the code itself and then every time you make a change you run a whole sweep of hundreds or thousands of tests. And if you break anything along the way, you see little red lights show up and tells you… Max: … the siren goes off, people start running around. Lights flashing everywhere Charlie: That would be nice. That would prevent me from making bugs Max: That’s contrary to… well traditional testing is not automatic. You have <a href="http://www.smackall.com/viewarticle.php?title=Quality-Assurance-For-Dummies&article=1882" target="_blank">QA [Quality Assurance]</a> people who run the game maybe they have a test plan with a bunch of things they’re going to test out and try to break it. Whereas the unit testing is automatic: you have some small piece of code which gets executed and then the test checks the results to make sure everything works right Charlie: right Max: So the nice thing is that you constantly running these checks. You’re not limited by man power. Or women power. Charlie: That’s right: people power. And with Natural Selection of course we just relied on you guys to test everything, but this time around, seeing how we’re selling the game, I don’t think that’s quite gonna cut it. Max: Well I would say that… Charlie: well for balance and stuff the community will do it. Max: Well, just general QA: my experience on Natural Selection, I thought the QA was just as good as it was. Charlie: isn’t that weird? Max: Working on professional products with publisher and their big QA team. Actually that doesn’t seem to matter. Charlie: in air quotes “big QA team.†It’s like their big marketing budget. Max: Yeah think we probably did have a bigger QA team in Natural Selection Charlie: Right. We definitely had a much wider install base probably. Or I’m not saying, not for Titan Quest but… I mean we probably had 30 testers, right? Something like that. So we had a couple hundred thousand. Max: Oh well I mean in the closed betas as well Charlie: Right. Oh that’s true. Meaning after [NS] 2.0 Max: Obviously with the mod it’s kind of an ongoing beta you know. Charlie: right Max: There is not as much pressure to get all the bugs fixed. Charlie: that’s true Max: Oh yeah even just the closed beta - we probably had a comparable sized beta team and you know <a href="http://www.unknownworlds.com/forums/index.php?showuser=8487" target="_blank">grepdashv</a>. Charlie: Right, he did an awesome job Max: as the lead QA guy he was great. Well he was kind of professional QA guy isn’t he? Charlie: I don’t think so. I can’t remember now. It’s been a while. I know he works at Intel but it sounds like we’re talking ourselves out of a resolution is what we’re saying. Cause do we want to automate it or do we want the community do it. Max: That’s the thing. I mean I’m not sure that unit testing is a good idea. Charlie: I think if we were writing a web app, something that was “text in text out,†I think unit testing would be fantastic. I think when we have a graphical and interactive product, highly interactive, that’s also a complex system, and plus our key heuristic is fun, or engagement, that’s really difficult to get a set up where you can… Max: Yeah, but there is the whole other level: just does it work? This is a problem we ran into and this is kind of the reason why I think unit testing is potentially useful is the problem that we would run into was: it’s the eleventh hour you discover some bug, but you cant fix the bug because Charlie: …the proper fix is risky Max: you don’t know what it’s going to affect. You don’t now, if I fix this is it going to break something else? Charlie: Right Max: If it’s the eleventh hour, you don’t have the opportunity to run through all of your tests to make sure that everything works. Charlie: eleventh hour happens all the time: publisher demos, investor demos… Max: it’s pretty much non-stop Charlie: … internal play test. It’s like every patch, we’re going to have a lot of those. Max: So that’s why I feel unit testing would potentially be helpful. I imagine a lot of cases even with unit testing people would probably still feel it would be a bit too risky for a lot of things, but it’s even when it’s not the eleventh hour. In general, you don’t know: “Oh, I found some bug, what is it going to break?†Presumably with a unit test you can immediately find out what change in that code would break if anything. But that assumes ideal unit tests that actually find bugs. You could have a bug and you write a test which doesn’t find the bug. Charlie: True, but it’s never a justification for not putting it in. Max: Presumably having it is better than not having it, assuming you don’t sacrifice, you don’t scale your other QA efforts because you have it. Charlie: right. Well without going into too much detail we should… Max: well I don’t know. I’m not putting that down as a resolution. Charlie: Ok, I don’t know why I wrote it down. Ok. I think it would be a great goal and I think it should be something we should keep our eyes peeled for. ‘Cause I would love to press a button that says “game is perfect†or “game failed two tests.†But the amount of work involved… Max: We could make a button that has that, and it has a big green light to just say "This game is perfect", whenever you push it. Charlie: It would make me happy. We could read up one of those orbs so it’s glowing on our desk [Max inaudible] I think it’s something we should continue to look out for, cause that would make us a better studio… Max: Ah, the ambient orb.
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> MAKING ESTIMATES (6:50) <!--sizec--></span><!--/sizec-->
Charlie: Let’s see so I think another thing is: we need to get better at estimating. We are really bad at that. Thought it’s not really clear to me if that actually is an attainable goal. ‘Cause if you look at the best developers out there, like Valve, Blizzard, they’re not used to it. They don’t make estimates and then hit them. It kinda seems opposite to making a good game. Max: Yeah. I don’t know if that’s true or not. Charlie: I don’t either. Max: That’s actually the one criticism I have of a lot of these kind of practices. I mean estimating isn’t really in the same class as unit testing, because everyone has to estimate. But some companies really espouse to these techniques and I’m not sure that it actually translates to better products. Charlie: Well estimating may not, well it could either lead to a better products because you say “I would rather have feature A or a feature B and C because the estimates are both the same†Assuming they’re accurate you can make that choice. Or the bigger thing it just helps your company run cause when you think that you… Max: That’s what I mean Charlie: it’s risky to not be able to estimate Max: the best company practices Charlie: …may not give you the best game Max: …not necessarily. It seems that way. I guess I don’t know that much about the practices of Blizzard or I guess we know a little about Valve. Charlie: I’m guessing they just don’t estimate that much, cause they know that it’s always off so they just assume it’s going to take five times as long as Max: I don’t think that’s true Charlie: No? Max: I think, no, I think that the reason why things take a long time at companies is ‘cause they do stuff and then decide it’s not good and they throw it away. Charlie: I know Blizzard does that a lot Max: But that doesn’t mean you don’t estimate Charlie: Right Max: That means you need have to have a flexible process, so at the start of the process you have your estimates, that says you have all your tasks broken down and how long they’re going to take. And when you get through the first tenth of that you realize that everything is wrong so you redo it. Estimating is still useful you just need to constantly be revising. Charlie: right. It’s like in military, they say a plan is essential even though it is irrelevant. <a href="http://www.unknownworlds.com/forums/index.php?showtopic=103563&st=20&p=1666698&#entry1666698" target="_blank">What is the saying?</a> “Planning is essential but the plan is useless†or something? I forgot how it’s worded. Max: I think it’s a Patton quote. Charlie: So basically it’s the planning that goes into formulating it. Also a horrible reference. I’m sure you can find in Google or something. So we would like to get better at that, ‘cause we’re pretty bad about some things projects, the website and stuff. And <a href="http://www.unknownworlds.com/decoda" target="_blank">Decoda </a>and <a href="http://www.unknownworlds.com/zen" target="_blank">Zen of Sudoku</a> and everything. Max: Well maybe. I mean I’ve always been bad at estimating the time that something takes. I know because at <a href="http://www.ironlore.com/" target="_blank">Iron Lore</a> we would track all that stuff. We had a schedule it was pretty detailed... Charlie: Everyone has a multiplier Max: we had a programming schedule in terms of days. I think art actually was scheduled down to the hour, but programming we would do on days so it’s fairly detailed. And pretty much my estimates were always about one third of what they should have been. But I guess that’s ok, so long as you know there's a common... Charlie: If it’s consistent, you come up with the factor. Max: the Max factor of three. That’s good to know. I don’t know if I stand by it. The problem is that it wasn’t like it’s going to take the third of the amount of time. What would actually happen is I would spend two thirds of my time doing something else. Like helping people with problems they have or stuff like that or scheduling or whatever. So there is that scaling factor which is probably based on underestimation of how long a task is going to take and then also how much time you actually spend working versus watching you do videos or whatever. Charlie: Right. Max: But there is another piece of it I want to talk about. There is figuring out how long a task is going to take and then figuring out what the tasks actually are. And I’m not as concerned about how long the tasks take, because I guess there is a scaling factor you can usually figure out, but figuring out what exactly those tasks are that’s what I found to be a bigger problem for me personally. Once I list 10 tasks I need to do, I’ve done the first task, it’s like… Charlie: Now you have fifteen tasks, Max: … well all the other tasks don’t even make sense anymore. Charlie: Oh really? Huh Max: I guess it’s the matter of actually thinking through of what you’re going to do Charlie: Right Max: I’m more of a seat-of-the-pants kind of guy I guess. Charlie: right. I’ve made little to-do lists but I usually double the task in the to-do lists. Max: You don’t figure all out Charlie: Yeah. This is going to require something else first so you add that in. So you need to do two things before you get to the first item on the list. So that slows you down right there. So we should take measurements and record them in a spreadsheet and check them every month and just a little bookkeeping to see if we can get better. Max: Well <a href="http://www.fogcreek.com/" target="_blank">Fog Creek</a> has software that does that, which I guess is an interesting idea Charlie: it’s expensive. Max: Well yeah, I don’t know if we necessarily need to get the software, especially because the technique they use to estimate is pretty simple. But yeah, they basically take the historical data and they automatically make predictions about the future, which I think is not the traditional way. You don’t do it automatically. I don’t know how you do it - with your hands or something. I think it’s an interesting idea; maybe we can try and play with it. Charlie: There is something else there too that I thought is really amazing that’s slipping my mind now, but which I can really bring up right now I guess. But I think they said, what was it? Oh I know what it was: the average amount of time you think something would take you is biased not with an equal fall off on both sides but it actually is slanted to the – it’s an exponential falloff the average amount of time. So if you think a task is going to take you a day, the amount of time it’s actually going to take you is not evenly distributed around the day. It’s biased after the day. prediction of the future. Max: Oh yeah. I guess that’s built in into the Charlie: Ok. Which seems like a totally key realization. Like that kinda blew my mind when I saw that. This is in <a href="http://www.fogcreek.com/FogBugz/" target="_blank">fogbugz </a>they have a video you can watch for their QA product. So estimating is something we should get better at. We should come up with some next action for that. Max: Perhaps. My original approach was just to eliminate the necessity of estimating and scheduling from my life, but… Charlie: How? Max: By working at Unknown Worlds Charlie: oh really? Charlie: I don’t think it works that way. We still have lots of constrains still Max: No but the smaller your team, the less important that stuff becomes to the point where if you’re just one person, it’s not nearly as important. Charlie: I don’t know about that. Because when you have X amount of dollars in the bank, you need to get something done before anyone else. Max: I guess there are different reasons for scheduling. There is that, there is also working with other people. They need something at some date to do their work. So I need to deliver Charlie: That’s true. But we even have that with out tiny team. That happens Max: I’d say that’s reduced. It is reduced because it’s reusing an existing engine. 99% of the stuff we need is there Charlie: that’s true, that’s true. But if we’re wondering if we’re going to put dynamic infestation in, versus some special welding system, and we think that estimates are the same in actuality one is twice the other, we want to know that, before we start. That’s what I really want, so we can trade which features we’re going to play with. Max: right. Charlie: That’s the biggest goal for me. So that’s probably enough about our stuff. We have a few more here. Otherwise we’re pretty much perfect. Max: Pretty much. Charlie: Yeah, I don’t want us to be … Max: It’s a little dry
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> GAME INDUSTRY: CHANGE OF CONTENT (15:10)<!--sizec--></span><!--/sizec-->
Charlie: …it’s a little dry, sorry about that. Yeah. But we wanted to talk about the industry too. Kind of take a step back and see where the industry is going and presume that we know what’s best for the industry, which we don’t of course, but we just have ideas of what we’d like to see the rest of the industry do for their new years resolutions. Probably the top of the list is to change the focus of content: move it away from fighting and I would even say Sci-Fi and Fantasy, away from fighting and conflict or at least violence. Max: Right. Charlie: I’m sure you have something more to add to that. Max: No I agree. I don’t think this is like us being idealistic either. If you look at the big successes this year or the last couple of years, actually what I’m really thinking of is Guitar Hero Charlie: Nintendogs Max: yeah that’s true, a lot of the Wii titles. You know, if the content is different, it’s a bigger audience. And it’s differentiated. It’s not just another World War II game. Charlie: that you have to fight for advertising space and shelf space or people’s minds. Max: So it seems really pragmatic to me Charlie: I agree. Totally agree. I think it makes great business sense Max: It’s totally not what we’re doing Charlie: It is not. And the reason why we’re not doing it is, because we’re still kind of bootstrapping. We don’t have unlimited resources, but we need to prove to someone else that they can take a bet, take a chance on us. We have to show them what we’ve done. We can’t switch genres in the middle of that. They’re not going to buy into it. Max: right. Charlie: so we have to say: “we’re going to make a sequel to NSâ€â€¦ Max: No I mean we never really had a discussion about that Charlie: No we didn’t Max: because it wouldn’t make sense for us to do anything else Charlie: Yeah Max: we want to make Natural selection 2 Charlie: We do. We do. But I think we also have a bunch of other game ideas which are pretty wild... Max: next game may not be Natural Selection 3 Charlie: Yeah we don’t know. But I would love to try something really wild. So that’s one thing, industry. You, industry. Make less shooting games. Excluding us. Max: It’s a crutch, you know, you look at Portal: maybe is technically a shooting game? Charlie: I don’t think so Max: it’s obviously very different. That doesn’t necessarily mean non-violent, kid-friendly or whatever. I mean in some ways Portal is not either of those. It’s actually kind of extra violent in sadistic… - the characters. If you’re not relying on the crutch of “this is my core game mechanic: shooting a gun at somebody,†then you’re doing something more interesting. Like creating portals that bend space. Charlie: right. Or having players create their own content like scribbling on walls and having to interact with other players or something weird. There was another point I was going to make about that - it’s lost. Oh the phi [??] obstructions that we talked about briefly. We could force ourselves or encourage other people to change their core gameplay in a big way by putting really serious constraints on your game idea. Max: that is a constraint, “no shootingâ€. Charlie: Yeah no shooting. Use a first person shooter engine, but make a non-shooter or make a shooter game without violence, which you get all those kind of camera games, where you take pictures of things in the world. Max: Oh like Pokemon? Charlie: Yeah all those games I didn’t play, but looked cool. Max: they were very successful, so.
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> GAME INDUSTRY: SMALLER GAMES (18:50)<!--sizec--></span><!--/sizec-->
Charlie: yeah, that’s great, so… anyways, there’s that and… We also think that industry in general, you industry, should make smaller games. Max: Yeah think the same thing. Like Portal or Guitar Hero. You don’t have to follow the path that has already been trod, because you don’t have the same concerns about creating a mega blockbuster. It’s the same thing with movies. You know that your Arnold Schwarzenegger big action movie is going to sell a whole bunch of money. You don’t know that, say Blair Witch, like smaller thing, you don’t know that it’s gonna be successful. Charlie: Right Max: But that means you probably get some level of success and then something like Blair Witch obviously you have an opportunity for huge amount of success. So you can do something that’s a little more risky, if it’s small, ‘cause you’re not investing a lot in it. Charlie: and that’s a good constraint right there, because that forces you to use unknown talent. Which is one of the reason why Blair Witch was so big is because they felt like kids you knew. If it was Tom Cruise, then the effect would be totally ruined. It’s like “Kate Winslet as crying girl,†you know that wouldn’t work. But what Da Vinci said… Max: yeah this quote, I get, right. Charlie: with constraints come freedom, actually that isn’t quite the quote. But that’s the idea. Constrains give you the freedom Max: paraphrase Charlie: Or “Let the constraints liberate you� or I don’t know. I’m gonna look this up right now. Max: the correct quote Charlie: So, make smaller games Max: But stuff like, good examples would be <a href="http://pc.ign.com/articles/740/740863p1.html" target="_blank">DEFCON </a>or <a href="http://www.youtube.com/watch?v=Taiurn541SE" target="_blank">Little Big Planet</a>. Charlie: though that doesn’t look too small, especially with the physics tech. Max: Uh it’s small, in the scope of like Halo 3 Charlie: True, or World of Warcraft or something like that. Max: These are like, World of Warcraft is like a $50 million dollar game. Little Big Planet is like a $5 million game. That would be a lot of money for that project. Charlie: True. That’s true. I know there are a ton of other examples. Probably the flash MMOs are probably down at that level. What else? We’re starting to see them at any rate. Max: well I think the market is becoming better for Indy developers who are inherently, like us, who are inherently in the budgetary bracket. Charlie: Way less than 5 million. Max: I mean we can’t make a blockbuster in terms of content. Hopefully in terms of sales it’s a blockbuster. Charlie: Hopefully in gameplay we can make a blockbuster. We can rival anything that anyone else can come up with. Max: well that’s kind of the thing, it’s like if you’re making the Arnold Schwarzenegger movie, you’ve got Arnold Schwarzenegger and you have explosions and you have all this stuff so you don’t need to have a good story line. Charlie: Right, it’s not what they’re there for. It’s a good analogy. Max: Although maybe Arnold Schwarzenegger it’s like dating myself or something. Charlie: It may be, seeing how he hasn’t made a movie in quite some time. Max: I don’t think he’s making any more. Well he made Terminator 3, that wasn’t that long ago. Charlie: That’s true Max: That was probably like 15 years ago. Charlie: We will edit that part out as well. Well those are the main topics we wanted to hit. And we’re up to 22 minutes, so Max: wow Charlie: We can really blather on, can’t we? Anything else you want to say?
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> MORE MEDIA FOR NS2 (22:43) <!--sizec--></span><!--/sizec-->
Max: I guess it’s not really a resolution, because it’s really just the plan, but something people are probably hoping for that we’re planning on doing is releasing a lot more media from the game. Oh yeah. Probably more video some screen shots and stuff like that. Charlie: yep Max: So that can be a resolution that we will definitely fulfill when we fail at all these other resolutions. Charlie: or fail to try even. Yeah but we will be releasing a lot more screen shots and … alright well, thanks for listening and we’ll see you next week. Max: bye Charlie: bye (23:20)
<!--quoteo(post=1666816:date=Jan 8 2008, 09:22 PM:name=CanadianWolverine)--><div class='quotetop'>QUOTE(CanadianWolverine @ Jan 8 2008, 09:22 PM) <a href="index.php?act=findpost&pid=1666816"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->OMG DEATHMATCH IS BAD!!!1!one<!--QuoteEnd--></div><!--QuoteEEnd-->
When I said <i>deathmatch</i> I was talking about the historical and engaging arcade-speed and that comes to mind from the word. I wasn't trying to suppress NS's future rts aspects. I like them too.
locallyunsceneFeeder of TrollsJoin Date: 2002-12-25Member: 11528Members, Constellation
edited January 2008
<!--quoteo(post=1666907:date=Jan 9 2008, 01:28 PM:name=Radix)--><div class='quotetop'>QUOTE(Radix @ Jan 9 2008, 01:28 PM) <a href="index.php?act=findpost&pid=1666907"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->When I said <i>deathmatch</i> I was talking about the historical and engaging arcade-speed and that comes to mind from the word. I wasn't trying to suppress NS's future rts aspects. I like them too.<!--QuoteEnd--></div><!--QuoteEEnd--> Somewhere, I forget where, the dev team mentioned they wanted to focus more on the RTS aspects for NS2. You're saying it's important to realize that the core of NS1 is an incredibly diverse and well balanced death match if I'm not mistaken. Commanders could have the most fun RTS experience ever, but if there's no depth at the FPS level, no one will want to play the game competitively, if at all. If my interpretation is correct, I agree with you.
Edit: I chuckled at your paraphrasing, if an exaggeration, of Canadian's post.
<!--quoteo(post=1666816:date=Jan 8 2008, 09:22 PM:name=CanadianWolverine)--><div class='quotetop'>QUOTE(CanadianWolverine @ Jan 8 2008, 09:22 PM) <a href="index.php?act=findpost&pid=1666816"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->OMG DEATHMATCH IS BAD!!!1!one<!--QuoteEnd--></div><!--QuoteEEnd--> <!--quoteo(post=1666908:date=Jan 9 2008, 01:55 PM:name=locallyunscene)--><div class='quotetop'>QUOTE(locallyunscene @ Jan 9 2008, 01:55 PM) <a href="index.php?act=findpost&pid=1666908"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Edit: I chuckled at your paraphrasing, if an exaggeration, of Canadian's post.<!--QuoteEnd--></div><!--QuoteEEnd-->
I know I didn't chuckle. That isn't paraphrasing, that is a smear. Being misquoted, having one's words twisted in such a fashion is nothing short of lying. It presents my stance on NS being more than death match like some rant from a cap lock key stuck ranting CS fanatic which is straight up bull ######. I have a few explicit words for troll tactic users.
Consider what I really said: <!--quoteo(post=1666816:date=Jan 8 2008, 09:22 PM:name=CanadianWolverine)--><div class='quotetop'>QUOTE(CanadianWolverine @ Jan 8 2008, 09:22 PM) <a href="index.php?act=findpost&pid=1666816"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Come to think of it, if all I wanted was Death Match, I would rather be playing Unreal Tournament and its extras to that than NS. Yeah, NS is definitely more than just Death Match.<!--QuoteEnd--></div><!--QuoteEEnd-->
<b>I don't think Death Match is bad.</b> In fact, I enjoy death match games a great deal, back to the days of the first widely spread death match games, such as Quake. So don't go mis-representing my opinions, its not appreciated at all.
<!--quoteo(post=1666907:date=Jan 9 2008, 01:28 PM:name=Radix)--><div class='quotetop'>QUOTE(Radix @ Jan 9 2008, 01:28 PM) <a href="index.php?act=findpost&pid=1666907"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->When I said <i>deathmatch</i> I was talking about the historical and engaging arcade-speed and that comes to mind from the word. I wasn't trying to suppress NS's future rts aspects. I like them too.<!--QuoteEnd--></div><!--QuoteEEnd-->
Well, thank you for clarifying that. I also enjoy the engaging arcade speed and would love to see an implementation in NS2 that allows for action within the frame work of the Marine experience, such as acrobatic animations and correlating defensive gains and offensive downplaying from it - such as dives, rolls, wall manuevers, flips, etc (aka Free Running / Parkour). Do you think your earlier mantra still stands then?
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->NS is a deathmatch game. Please maintain that mantra in NS2.<!--QuoteEnd--></div><!--QuoteEEnd-->
<!--quoteo(post=1666813:date=Jan 9 2008, 01:54 AM:name=corpsman)--><div class='quotetop'>QUOTE(corpsman @ Jan 9 2008, 01:54 AM) <a href="index.php?act=findpost&pid=1666813"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Oh ok! That kinda threw me off. I hear y aknow, "HL source," but then someone always pipes up and says something else like NSTR2, or encoda, or something, so I get confused. You rsaying the game will be made using the Source engine? I fugred it would maybe be old looking because they were using some other older, game engine. I am confused again.... <b>Arg!</b><!--QuoteEnd--></div><!--QuoteEEnd-->
Yes NS2 will be made on the Source engine, the latest version of the source engine, which is ep2 (officially HL2 ep1) standards not Half-Life 1 as NS was.
The NS2TR is the pack that they will release for people wanting to make maps.
Decoda, not encoda, is as far as I remember a thing to help you debug Lua code.
Anything else that you are confused about?
Most of that stuff is here <a href="http://www.unknownworlds.com/ns2/faq/" target="_blank">http://www.unknownworlds.com/ns2/faq/</a>
EDIT: I also suggest that instead of working on random mods StixNStonz, where you may never see your map being released, you go out and buy TF2 or CS:S and make a map for them. That way at least you know people will be play testing them at the end of it and you will at least get some kind of feedback as to what people like and don't like. It also means that if the NS2TR comes out when your in mid map, no one will care if you just drop it or release anything more than a beta version. Who knows you might make the next map that’s played as much as Dust <img src="style_emoticons/<#EMO_DIR#>/tounge.gif" style="vertical-align:middle" emoid=":p" border="0" alt="tounge.gif" />
<!--quoteo(post=1666919:date=Jan 10 2008, 09:08 AM:name=CanadianWolverine)--><div class='quotetop'>QUOTE(CanadianWolverine @ Jan 10 2008, 09:08 AM) <a href="index.php?act=findpost&pid=1666919"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I know I didn't chuckle. That isn't paraphrasing, that is a smear. Being misquoted, having one's words twisted in such a fashion is nothing short of lying. It presents my stance on NS being more than death match like some rant from a cap lock key stuck ranting CS fanatic which is straight up bull ######. That's right Radix, I called you a liar and I have a few other explicit words for troll tactic users like you.<!--QuoteEnd--></div><!--QuoteEEnd--> I wouldn't go so far as to call him a liar. I agree that his quote of you had nothing to do with what you actually said, but it seems to me it was just that he didn't bother reading your post before jumping to conclusions, as opposed to intentionally misrepresenting your opinion.
<!--quoteo(post=1666940:date=Jan 9 2008, 05:47 PM:name=Revenge)--><div class='quotetop'>QUOTE(Revenge @ Jan 9 2008, 05:47 PM) <a href="index.php?act=findpost&pid=1666940"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I wouldn't go so far as to call him a liar. I agree that his quote of you had nothing to do with what you actually said, but it seems to me it was just that he didn't bother reading your post before jumping to conclusions, as opposed to intentionally misrepresenting your opinion.<!--QuoteEnd--></div><!--QuoteEEnd-->
Ahh, alright, I can swing with that. *Breaks out his edit scalpel* I'll give the shadow of a doubt that it was unintentional ######baggery.
Thanks afratnikov for the text of the podcast. I think that is pretty cool about this community that we are getting involved like this, I hope you guys get some sort of mention in the credits or a cameo of sorts like your name in small print on a computer screen in one of the maps.
<!--quoteo(post=1666932:date=Jan 9 2008, 04:58 PM:name=INKEDOUT)--><div class='quotetop'>QUOTE(INKEDOUT @ Jan 9 2008, 04:58 PM) <a href="index.php?act=findpost&pid=1666932"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I also suggest that instead of working on random mods StixNStonz, where you may never see your map being released, you go out and buy TF2 or CS:S and make a map for them.<!--QuoteEnd--></div><!--QuoteEEnd-->
I actually began my mapping career in CS:S. Bought HL2, got back into CS for a month or two, started mapping for it, then went back to playing NS. So i was mapping for CS:S for about a year while playing NS.
I made two fun maps:
<u><b>De_Cover</u></b>. Three buildings facing each other, totalling over 150 windows, with the bombsite in the middle. CTs mostly spawn in the buildings with snipers/paras, with a few on the ground with subs. T's spawn mostly under the tarp with rifles and smoke grenades, with a few spawning in the buildings for counter-sniping. There's a ton of cover in the middle of all this. People said it felt like a modern-day Normandy.
This was my first released map, and it got a good amount of playtime on a few servers. Many people complained about the visuals, so I optimized the map significantly and adding in MUCH more realistic looking buildings, but never got around to releasing the new version. Here are some screens from the old version.
<u><b>1v1_scoutflash</u></b>. This was actually for a 'what-can-you-fit-in-a-small-box' competition, the box being about 20 feet in all 3 directions. I created a surprisingly cool little 1v1/2v2 map, set atop a highrise building.
Using a highly optimized 3d skybox (which was under the floor), it simulated a much larger rooftop, and showed the surrounding city. Players could fall off the rooftop on two sides, fall to their death screaming, and you'd see a blood splatter on the street afterwards (go go trigger system). The players spawned with a Scout and two flashbangs (not in sight of each other regardless of spawns), and I somehow fit a decent amount of cover and height variation in this area. There are two semi-hidden knives, a well-hidden grenade, and a secret that teleports the player into the skybox, creating the Godzilla effect.
Then there was <u><b>cs_abscond</u></b>. This was my first serious map. De_Cover taught me the basics, the tricks (skyboxes, displacements, etc) and about optimization, and then 1v1_Scoutflash taught me about real visuals and 'keeping it tight', as i'd say. I had big, big plans for Abscond.
Some highlights: an Embassy with a full suite of security systems (laser trip beams, one-day doors, video cameras, all done), Source's first Shotgun-Breachable door (built and working; only a shotgun shell to the handle could open the door from the outside, thoroughly tested), JFK-style sniping from the sewers in the sidewalks, and a huge scripted Hostage rescue scene involving a helicopter extraction and C4 under the helipad (with a very long fuse, for that final 'scene'), all of this already built and good to go.
The problem was with the map itself. Although i knew all the tricks (especially the triggers) and how it all worked, my fundamental mapping skill was still too noob.
About halfway through this map, I realized how silly it was to be mapping for a game I never played (CS:S), so I decided to try and learn HL1 mapping for NS. Thats when the 16 month process of creating <u><b>Ns_Nexus</b></u> began, during which I had multiple 'mapper enlightenments', often going back and fixing up everything I'd built so far.
HL1 mapping has given me one hell of a solid mapping foundation, so now I'm ready to go back into Source and absolutely tear it up with that Sexy McSexy engine. But I dont want to do CS:S again. I want NS2.
**********************************
So, after giving you my entire mapping history, I will say this. If i can't map for NS2 (since even the TR seems nowhere in the near future), I won't map for CS:S. That game is already flooded with maps, the majority being horrible but still with a TON of really amazing custom maps. Even with a huge effort, whatever map I'd make would get lost in the clutter.
I'm looking for a mod with promising gameplay, for whom I can deliver a stellar map. That way, its almost assured to become one of their 'official' maps, which is what I really need for my portfolio right now. Despite Nexus being roughly on par in terms of scope, depth and visuals as Machina, there's been no peep about its official inclusion, and the Official Map Pack seems to have died out months ago.
"Charlie: right. It’s like in military, they say a plan is essential even though it is irrelevant. What is the saying? “Planning is essential but the plan is useless†or something? I forgot how it’s worded. Max: I think it’s a Patton quote. "
"No battle plan survives contact with the enemy." -<a href="http://en.wikipedia.org/wiki/Helmuth_von_Moltke_the_Elder" target="_blank">Helmuth von Moltke</a>
We are waiting for Finalized Models, and maps to practice <img src="style_emoticons/<#EMO_DIR#>/biggrin-fix.gif" style="vertical-align:middle" emoid=":D" border="0" alt="biggrin-fix.gif" />
Some great posts here...of course it's boring nonsense to some, that's ok...here's my two cents...
<b>Estimating</b> There's are some nice techniques for improving estimation. One useful concept is the 'cone of uncertainty' :)
Briefly you can think of a simple chart with two lines which start wide apart and converge as they move right. The large gap on the left is a reflection of a high level of uncertainty in an estimate. For example your about to develop a feature and there are lots of unknowns, potential problems, questions etc, so you're estimate is really a range: 2 days - 4 weeks. It's important to change you're thinking away from choosing a single number to a possible range. As you spend time on a task/project you learn the gotchas, find rabbit holes you never imagined, develop ways to work faster, so as time moves forward your uncertainty drops and the range narrows.
Using a model like this you can easily think of tasks which have great ranges (lots of unknowns) and use some simple formulas to calculate schedules based on task ranges. Once you use a technique like this the typical approach of trying to come up with a single number and throw on a high simplistic modifier (i.e it usually takes me twice as long) seems ridiculous. Simply put - some tasks can be accurately estimated others can't be - a model like this takes that into account. And you can easily see how you can revisit your estimates as you reduce the uncertainty.
At the end of the day, a big uncertainty is a major change in game direction and that could really through off the high level schedule! There's only so much uncertainty you can remove. <a href='index.php?act=findpost&pid=0'>Cone of Uncertainty</a>
<b>Unit Testing</b> As for as unit testing goes it definitely has a value in typical software development, I can't see how that wouldn't apply in game development. Ultimately as you mentioned it's no silver bullet and if you don't have the skills to write good unit tests it's simply a waste of time. There's a cultural aspect to it, if you believe it's a waste time it more than likely will end up that way, it has to suit the teams mentality and skills. The ability to have repeatable testing to ensure more rigorous testing and allow the impact of changes to be more quickly explored can only be good especially when there will be areas of code with very iteration levels. All end user driven software is going to have a significant user acceptance component for the user experience component of testing (i.e. is it fun and easy to use) but it shouldn't be the only approach.
You don't have to choose either or, you can have a hybrid approach with good algorythmic planning and some level of simple debug time. However some components will be subject to a high level of iteration which would make the overhead of re-writing unit tests a burden, but for more stable core components why not? Ultimately I think the time saved with test drive development has a place in any serious project.
German version finally has arrived: <a href="http://www.lerk.de/2008/02/neujahrsvorsatze-komponententest-und-die-zukunft-der-spieleindustrie/" target="_blank">http://www.lerk.de/2008/02/neujahrsvorsatz...pieleindustrie/</a>
Comments
That wouldn't be very exciting, unless you're a programmer I'm sure.<!--QuoteEnd--></div><!--QuoteEEnd-->
no one is forcing you to listen to them......
hotd0g: Guess what? You'll have no idea whatsoever how development is going if they don't do this.
Domining: I'm not sure what constitutes a professional podcast, but I don't see what's wrong with the current one. This isn't a radio show, it's an audio commentary for people who don't want to read paragraphs and paragraphs of jotted down development notes.
Perhaps small news digests should be released with or around the time of the podcasts for people without the time to listen to them. Perhaps people feel obligated to listen if they want to know what's going on, when some of it can be condensed down to bullet points. I personally enjoy listening to the entire audio commentary clip.
Remember, it's called a <b>development</b> podcast/blog. You don't make games by synthesizing how they should work, then magically produce them.
-Dwight D. Eisenhower
Skimmed the replies, don't think anyone said it.
hotd0g: Guess what? You'll have no idea whatsoever how development is going if they don't do this.
Domining: I'm not sure what constitutes a professional podcast, but I don't see what's wrong with the current one. This isn't a radio show, it's an audio commentary for people who don't want to read paragraphs and paragraphs of jotted down development notes.
Perhaps small news digests should be released with or around the time of the podcasts for people without the time to listen to them. Perhaps people feel obligated to listen if they want to know what's going on, when some of it can be condensed down to bullet points. I personally enjoy listening to the entire audio commentary clip.
Remember, it's called a <b>development</b> podcast/blog. You don't make games by synthesizing how they should work, then magically produce them.<!--QuoteEnd--></div><!--QuoteEEnd-->
I'd rather see an IRC discussion than a podcast, I really don't feel like having to listen to people talk when I can read it and get it done faster. Wouldnt' be bad for an easy Q&A as well.
I know more now with regards to what they were talking about, but I also feel like I totally missed everything else, like other parts were left out. I still don't know where they are in terms of maps, models, testing, answers to forum topics, and a few others probably. <img src="style_emoticons/<#EMO_DIR#>/sad-fix.gif" style="vertical-align:middle" emoid=":(" border="0" alt="sad-fix.gif" />
I do wish they would still do recaps concerning the rest of the game such as the topics I stated just a second ago [maps, models, and where they are on everything else]. but comments like "the podcast sucks" probably pisses them off a tad. I know I would be a little peeved...
Could you guys maybe do brief recaps or short blurbs on other topics in future podcasts? I was personally curious where you guys were with maps and models progress. Are you making maps as you go, or do you actually have a planned out list or the maps you will be making [i.e. A through Z]? And I don't remember hearing an answer on the customization of models, will you allow it beyond Male/Female? I know I would like to at <i>LEAST</i> be able to change either my face, or something unique! <img src="style_emoticons/<#EMO_DIR#>/tsa.gif" style="vertical-align:middle" emoid="::tsa::" border="0" alt="tsa.gif" />
Also, whoo more media!
they said the initial alpha release would only be sent to mappers that have contributed significantly in the past, mappers they know and love, mappers that will make the official maps for NS2. so unless you made or updated maps like eclipse, bast, machina etc, you'll have to wait until the full NS2TR is released, whenever that is.
<!--quoteo(post=1666721:date=Jan 8 2008, 07:22 AM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Jan 8 2008, 07:22 AM) <a href="index.php?act=findpost&pid=1666721"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Feels like its just a matter of time before I give up on NS mapping and find some source mod to contribute to. At least then I'd be working on a more up-to-date engine...<!--QuoteEnd--></div><!--QuoteEEnd-->
err... they're using the most latest release of source - ep2. you wont find any source mod out there with a more up-to-date engine than that....
The sooner you get the gameplay done, the sooner the -final- maps are going to be done. #$%# unit testing. You have a huge f'ing testing base. Even if the game is buggy to hell (but fixable), I'm sure you can get peeps to help you for free. GOGOGOGO FFS. Work 80 hours a week TESTING and CODING the f'ing SKILLS and movement and stuff. Slap it into a random CS map and go at it. Whatever, it doesn't even matter. Get the gameplay and skills roughly cut out and slap it into your orange-skinned boxy maps. Then 'take it up a notch', and get the maps done. Seriously, GOGOGO.<!--QuoteEnd--></div><!--QuoteEEnd-->
Even though he's making an ass out of himself, MasterP raises a good point - the core reason your game has endured is the gameplay, and while I have no reason to believe you haven't worked more on that than anything else (since we wouldn't see it yet), if, indeed, all the team has done so far is an emulated release of DI and some (very impressive) art assets, or perhaps the port of the desired NS data structures into Source, then you may consider pulling back and rethinking how things are going along.
The gameplay is the infrastructure, the skeleton of the game. While it's necessary to design a complex system from the top-down before it "evolves" through iteration, it's crucial that you build the "polish and frosting" (eg blast radius vars, welding, functions of DI) <i>around</i> the core gameplay that a player spends 70-90% of their time doing, which is combat oriented, and by this I mean the attack and (perhaps even more so) movement systems.
NS is a deathmatch game. Please maintain that mantra in NS2. I would hate to see such a great fantasy simulation become as trivially point-and-click as CS or TF2 in its re-iteration.
<!--QuoteBegin-'Buzzou'+--><div class='quotetop'>QUOTE('Buzzou')</div><div class='quotemain'><!--QuoteEBegin-->err... they're using the most latest release of source - ep2. you wont find any source mod out there with a more up-to-date engine than that....<!--QuoteEnd--></div><!--QuoteEEnd-->
Yes, they're using the newest version of Source, but 99% of NS mappers right now are not. We're using HL1. Because thats all we have. There's no point whatsoever in even pretending to map for NS2 on source yet, since the critical numbers and details have not been given. Why make rooms for gameplay that we don't know, with doors possibly too small for the largest creature or DI, with too many rooms or simply the wrong types of rooms given then new gameplay, and with no textures and no props?
Its a much better idea to find some other interesting Source mod and create maps for those, if you want to learn Source mapping for NS2. I began on source so I already know all the bells and whistles, but i'm pretty sure most NS mappers don't.
The problem with this approach is that NS will lose its mappers. At least some will learn to love those other mods, or simply be midway into their new map and not want to abandon it just because the NS2TR finally came out.
Oh, and you mentioned that
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->they said the initial alpha release would only be sent to mappers [...] that will make the official maps for NS2<!--QuoteEnd--></div><!--QuoteEEnd-->
It sounds like the Alpha went out to like 5 mappers, and I'm sure some of them are simply helping out with the debugging without planning on putting in the huge time requirement for a Source map. The devs have explained that they're letting the community make the maps, and they'll buy the cream of the crop. So no, the majority of the mappers who will be making the official maps are sitting here on the forums (or not) and just watching the months go by. Hell, even the Official Map Pack hasn't gotten a post in about 4 months, despite the fact that it could have been released within a few weeks, 6 months ago.
Oh ok! That kinda threw me off. I hear y aknow, "HL source," but then someone always pipes up and says something else like NSTR2, or encoda, or something, so I get confused. You rsaying the game will be made using the Source engine? I fugred it would maybe be old looking because they were using some other older, game engine. I am confused again.... <b>Arg!</b>
NS is more than DM. Any game where it is possible, however remotely so, to actually win the game without killing another player, is not a Death Match. The elements that make up Natural Selection make it far more than simple Death Match. So, if you are going to maintain a mantra, at least make it accurate. NS is a team objective based game: Kill the last hive or kill the Marine IP + CC + entire team... Actually, why are we maintaining that mantra? They are already changing the resource model so they can have a more flexible game play prototyping process ... and they have their key stones / pillars for the game ...
Actually, the formula they put forward reminds me something of a bride getting ready for a supposedly traditional white wedding: "Something old, something new, something borrowed, something blue." I like the mantra(s) they have much better than your mantra. Come to think of it, if all I wanted was Death Match, I would rather be playing Unreal Tournament and its extras to that than NS. Yeah, NS is definitely more than just Death Match. Even Combat is more than just Death Match.
Having 10,000 testers to find all your bugs doesn't change the fact that it would take too long to fix them all. It's better to not create them in the first place.
Having 10,000 testers to find all your bugs doesn't change the fact that it would take too long to fix them all. It's better to not create them in the first place.<!--QuoteEnd--></div><!--QuoteEEnd-->
Totally agree.
In addition, some bugs may only be fixed by changing the overal design. And 9 out of 10, you wind up creating a big amount of new bugs.
I would say that the usual bugfixing is a process where the coder evaluates his design. He did not think about some structural design problems in his code and tries to clean his code during the bugfixing process.
Where test driven development makes you think about what you are about to build. You design the application before you write your code. Therefore their "should be" less bugs afterwards.
I’ve added the times at which new topics start in the podcast. Thanks <a href="http://www.unknownworlds.com/ns2/forums/index.php?showuser=30793" target="_blank">Crispy </a>for your PM with corrections.
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> INTRODUCTION (0:20)<!--sizec--></span><!--/sizec-->
Charlie: Welcome to the 15th unknown worlds podcast. Today is Friday January 4th 2008. This is Charlie
Max: And Max
Charlie: And welcome to the New Year
Max: yes
Charlie: We did it
Max: I don’t know what we did. We made it.
Charlie: I don’t know, we got through.
Max: That was a close one. We almost didn’t make it
Charlie: actually, yeah, that wasn’t too difficult.
Max: That’s true, well, no I was thinking it actually is almost true.
Charlie: Really?
Max: Well I don’t know. Who knows?
Charlie: Financially?
Max: There is ups and downs.
Charlie: That’s true. Well we made it intact. We got back from our families
Max: Yeah, we survived that as well.
Charlie: Yes. Mom, in case you’re listening, it was awesome.
Max: Since it’s the first podcast of the new year, we figure we would do some resolutions. Our game development resolutions.
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> UNIT TESTING (1:13)<!--sizec--></span><!--/sizec-->
Charlie: Right. So we’re gonna talk about a bunch of things that we want to change about ourselves in respect to game development and maybe also what we think the industry should change for this year. It’s kind of a slow week so. ‘Cause we just got back from vacation so indulge us. So some of the things we’re thinking about: well we always bring up the concept of <a href="http://en.wikipedia.org/wiki/Unit_testing" target="_blank">unit testing</a>, which is a big pain in the butt to implement. But basically what that is, is testing every aspect of your code. Often you write a test before you even write the code itself and then every time you make a change you run a whole sweep of hundreds or thousands of tests. And if you break anything along the way, you see little red lights show up and tells you…
Max: … the siren goes off, people start running around. Lights flashing everywhere
Charlie: That would be nice. That would prevent me from making bugs
Max: That’s contrary to… well traditional testing is not automatic. You have <a href="http://www.smackall.com/viewarticle.php?title=Quality-Assurance-For-Dummies&article=1882" target="_blank">QA [Quality Assurance]</a> people who run the game maybe they have a test plan with a bunch of things they’re going to test out and try to break it. Whereas the unit testing is automatic: you have some small piece of code which gets executed and then the test checks the results to make sure everything works right
Charlie: right
Max: So the nice thing is that you constantly running these checks. You’re not limited by man power. Or women power.
Charlie: That’s right: people power. And with Natural Selection of course we just relied on you guys to test everything, but this time around, seeing how we’re selling the game, I don’t think that’s quite gonna cut it.
Max: Well I would say that…
Charlie: well for balance and stuff the community will do it.
Max: Well, just general QA: my experience on Natural Selection, I thought the QA was just as good as it was.
Charlie: isn’t that weird?
Max: Working on professional products with publisher and their big QA team. Actually that doesn’t seem to matter.
Charlie: in air quotes “big QA team.†It’s like their big marketing budget.
Max: Yeah think we probably did have a bigger QA team in Natural Selection
Charlie: Right. We definitely had a much wider install base probably. Or I’m not saying, not for Titan Quest but… I mean we probably had 30 testers, right? Something like that. So we had a couple hundred thousand.
Max: Oh well I mean in the closed betas as well
Charlie: Right. Oh that’s true. Meaning after [NS] 2.0
Max: Obviously with the mod it’s kind of an ongoing beta you know.
Charlie: right
Max: There is not as much pressure to get all the bugs fixed.
Charlie: that’s true
Max: Oh yeah even just the closed beta - we probably had a comparable sized beta team and you know <a href="http://www.unknownworlds.com/forums/index.php?showuser=8487" target="_blank">grepdashv</a>.
Charlie: Right, he did an awesome job
Max: as the lead QA guy he was great. Well he was kind of professional QA guy isn’t he?
Charlie: I don’t think so. I can’t remember now. It’s been a while. I know he works at Intel but it sounds like we’re talking ourselves out of a resolution is what we’re saying. Cause do we want to automate it or do we want the community do it.
Max: That’s the thing. I mean I’m not sure that unit testing is a good idea.
Charlie: I think if we were writing a web app, something that was “text in text out,†I think unit testing would be fantastic. I think when we have a graphical and interactive product, highly interactive, that’s also a complex system, and plus our key heuristic is fun, or engagement, that’s really difficult to get a set up where you can…
Max: Yeah, but there is the whole other level: just does it work? This is a problem we ran into and this is kind of the reason why I think unit testing is potentially useful is the problem that we would run into was: it’s the eleventh hour you discover some bug, but you cant fix the bug because
Charlie: …the proper fix is risky
Max: you don’t know what it’s going to affect. You don’t now, if I fix this is it going to break something else?
Charlie: Right
Max: If it’s the eleventh hour, you don’t have the opportunity to run through all of your tests to make sure that everything works.
Charlie: eleventh hour happens all the time: publisher demos, investor demos…
Max: it’s pretty much non-stop
Charlie: … internal play test. It’s like every patch, we’re going to have a lot of those.
Max: So that’s why I feel unit testing would potentially be helpful. I imagine a lot of cases even with unit testing people would probably still feel it would be a bit too risky for a lot of things, but it’s even when it’s not the eleventh hour. In general, you don’t know: “Oh, I found some bug, what is it going to break?†Presumably with a unit test you can immediately find out what change in that code would break if anything. But that assumes ideal unit tests that actually find bugs. You could have a bug and you write a test which doesn’t find the bug.
Charlie: True, but it’s never a justification for not putting it in.
Max: Presumably having it is better than not having it, assuming you don’t sacrifice, you don’t scale your other QA efforts because you have it.
Charlie: right. Well without going into too much detail we should…
Max: well I don’t know. I’m not putting that down as a resolution.
Charlie: Ok, I don’t know why I wrote it down. Ok. I think it would be a great goal and I think it should be something we should keep our eyes peeled for. ‘Cause I would love to press a button that says “game is perfect†or “game failed two tests.†But the amount of work involved…
Max: We could make a button that has that, and it has a big green light to just say "This game is perfect", whenever you push it.
Charlie: It would make me happy. We could read up one of those orbs so it’s glowing on our desk [Max inaudible] I think it’s something we should continue to look out for, cause that would make us a better studio…
Max: Ah, the ambient orb.
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> MAKING ESTIMATES (6:50) <!--sizec--></span><!--/sizec-->
Charlie: Let’s see so I think another thing is: we need to get better at estimating. We are really bad at that. Thought it’s not really clear to me if that actually is an attainable goal. ‘Cause if you look at the best developers out there, like Valve, Blizzard, they’re not used to it. They don’t make estimates and then hit them. It kinda seems opposite to making a good game.
Max: Yeah. I don’t know if that’s true or not.
Charlie: I don’t either.
Max: That’s actually the one criticism I have of a lot of these kind of practices. I mean estimating isn’t really in the same class as unit testing, because everyone has to estimate. But some companies really espouse to these techniques and I’m not sure that it actually translates to better products.
Charlie: Well estimating may not, well it could either lead to a better products because you say “I would rather have feature A or a feature B and C because the estimates are both the same†Assuming they’re accurate you can make that choice. Or the bigger thing it just helps your company run cause when you think that you…
Max: That’s what I mean
Charlie: it’s risky to not be able to estimate
Max: the best company practices
Charlie: …may not give you the best game
Max: …not necessarily. It seems that way. I guess I don’t know that much about the practices of Blizzard or I guess we know a little about Valve.
Charlie: I’m guessing they just don’t estimate that much, cause they know that it’s always off so they just assume it’s going to take five times as long as
Max: I don’t think that’s true
Charlie: No?
Max: I think, no, I think that the reason why things take a long time at companies is ‘cause they do stuff and then decide it’s not good and they throw it away.
Charlie: I know Blizzard does that a lot
Max: But that doesn’t mean you don’t estimate
Charlie: Right
Max: That means you need have to have a flexible process, so at the start of the process you have your estimates, that says you have all your tasks broken down and how long they’re going to take. And when you get through the first tenth of that you realize that everything is wrong so you redo it. Estimating is still useful you just need to constantly be revising.
Charlie: right. It’s like in military, they say a plan is essential even though it is irrelevant. <a href="http://www.unknownworlds.com/forums/index.php?showtopic=103563&st=20&p=1666698&#entry1666698" target="_blank">What is the saying?</a> “Planning is essential but the plan is useless†or something? I forgot how it’s worded.
Max: I think it’s a Patton quote.
Charlie: So basically it’s the planning that goes into formulating it. Also a horrible reference. I’m sure you can find in Google or something. So we would like to get better at that, ‘cause we’re pretty bad about some things projects, the website and stuff. And <a href="http://www.unknownworlds.com/decoda" target="_blank">Decoda </a>and <a href="http://www.unknownworlds.com/zen" target="_blank">Zen of Sudoku</a> and everything.
Max: Well maybe. I mean I’ve always been bad at estimating the time that something takes. I know because at <a href="http://www.ironlore.com/" target="_blank">Iron Lore</a> we would track all that stuff. We had a schedule it was pretty detailed...
Charlie: Everyone has a multiplier
Max: we had a programming schedule in terms of days. I think art actually was scheduled down to the hour, but programming we would do on days so it’s fairly detailed. And pretty much my estimates were always about one third of what they should have been. But I guess that’s ok, so long as you know there's a common...
Charlie: If it’s consistent, you come up with the factor.
Max: the Max factor of three. That’s good to know. I don’t know if I stand by it. The problem is that it wasn’t like it’s going to take the third of the amount of time. What would actually happen is I would spend two thirds of my time doing something else. Like helping people with problems they have or stuff like that or scheduling or whatever. So there is that scaling factor which is probably based on underestimation of how long a task is going to take and then also how much time you actually spend working versus watching you do videos or whatever.
Charlie: Right.
Max: But there is another piece of it I want to talk about. There is figuring out how long a task is going to take and then figuring out what the tasks actually are. And I’m not as concerned about how long the tasks take, because I guess there is a scaling factor you can usually figure out, but figuring out what exactly those tasks are that’s what I found to be a bigger problem for me personally. Once I list 10 tasks I need to do, I’ve done the first task, it’s like…
Charlie: Now you have fifteen tasks,
Max: … well all the other tasks don’t even make sense anymore.
Charlie: Oh really? Huh
Max: I guess it’s the matter of actually thinking through of what you’re going to do
Charlie: Right
Max: I’m more of a seat-of-the-pants kind of guy I guess.
Charlie: right. I’ve made little to-do lists but I usually double the task in the to-do lists.
Max: You don’t figure all out
Charlie: Yeah. This is going to require something else first so you add that in. So you need to do two things before you get to the first item on the list. So that slows you down right there. So we should take measurements and record them in a spreadsheet and check them every month and just a little bookkeeping to see if we can get better.
Max: Well <a href="http://www.fogcreek.com/" target="_blank">Fog Creek</a> has software that does that, which I guess is an interesting idea
Charlie: it’s expensive.
Max: Well yeah, I don’t know if we necessarily need to get the software, especially because the technique they use to estimate is pretty simple. But yeah, they basically take the historical data and they automatically make predictions about the future, which I think is not the traditional way. You don’t do it automatically. I don’t know how you do it - with your hands or something. I think it’s an interesting idea; maybe we can try and play with it.
Charlie: There is something else there too that I thought is really amazing that’s slipping my mind now, but which I can really bring up right now I guess. But I think they said, what was it? Oh I know what it was: the average amount of time you think something would take you is biased not with an equal fall off on both sides but it actually is slanted to the – it’s an exponential falloff the average amount of time. So if you think a task is going to take you a day, the amount of time it’s actually going to take you is not evenly distributed around the day. It’s biased after the day. prediction of the future.
Max: Oh yeah. I guess that’s built in into the
Charlie: Ok. Which seems like a totally key realization. Like that kinda blew my mind when I saw that. This is in <a href="http://www.fogcreek.com/FogBugz/" target="_blank">fogbugz </a>they have a video you can watch for their QA product. So estimating is something we should get better at. We should come up with some next action for that.
Max: Perhaps. My original approach was just to eliminate the necessity of estimating and scheduling from my life, but…
Charlie: How?
Max: By working at Unknown Worlds
Charlie: oh really? Charlie: I don’t think it works that way. We still have lots of constrains still
Max: No but the smaller your team, the less important that stuff becomes to the point where if you’re just one person, it’s not nearly as important.
Charlie: I don’t know about that. Because when you have X amount of dollars in the bank, you need to get something done before anyone else.
Max: I guess there are different reasons for scheduling. There is that, there is also working with other people. They need something at some date to do their work. So I need to deliver
Charlie: That’s true. But we even have that with out tiny team. That happens
Max: I’d say that’s reduced. It is reduced because it’s reusing an existing engine. 99% of the stuff we need is there
Charlie: that’s true, that’s true. But if we’re wondering if we’re going to put dynamic infestation in, versus some special welding system, and we think that estimates are the same in actuality one is twice the other, we want to know that, before we start. That’s what I really want, so we can trade which features we’re going to play with.
Max: right.
Charlie: That’s the biggest goal for me. So that’s probably enough about our stuff. We have a few more here. Otherwise we’re pretty much perfect.
Max: Pretty much.
Charlie: Yeah, I don’t want us to be …
Max: It’s a little dry
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> GAME INDUSTRY: CHANGE OF CONTENT (15:10)<!--sizec--></span><!--/sizec-->
Charlie: …it’s a little dry, sorry about that. Yeah. But we wanted to talk about the industry too. Kind of take a step back and see where the industry is going and presume that we know what’s best for the industry, which we don’t of course, but we just have ideas of what we’d like to see the rest of the industry do for their new years resolutions. Probably the top of the list is to change the focus of content: move it away from fighting and I would even say Sci-Fi and Fantasy, away from fighting and conflict or at least violence.
Max: Right.
Charlie: I’m sure you have something more to add to that.
Max: No I agree. I don’t think this is like us being idealistic either. If you look at the big successes this year or the last couple of years, actually what I’m really thinking of is Guitar Hero
Charlie: Nintendogs
Max: yeah that’s true, a lot of the Wii titles. You know, if the content is different, it’s a bigger audience. And it’s differentiated. It’s not just another World War II game.
Charlie: that you have to fight for advertising space and shelf space or people’s minds.
Max: So it seems really pragmatic to me
Charlie: I agree. Totally agree. I think it makes great business sense
Max: It’s totally not what we’re doing
Charlie: It is not. And the reason why we’re not doing it is, because we’re still kind of bootstrapping. We don’t have unlimited resources, but we need to prove to someone else that they can take a bet, take a chance on us. We have to show them what we’ve done. We can’t switch genres in the middle of that. They’re not going to buy into it.
Max: right.
Charlie: so we have to say: “we’re going to make a sequel to NSâ€â€¦
Max: No I mean we never really had a discussion about that
Charlie: No we didn’t
Max: because it wouldn’t make sense for us to do anything else
Charlie: Yeah
Max: we want to make Natural selection 2
Charlie: We do. We do. But I think we also have a bunch of other game ideas which are pretty wild...
Max: next game may not be Natural Selection 3
Charlie: Yeah we don’t know. But I would love to try something really wild. So that’s one thing, industry. You, industry. Make less shooting games. Excluding us.
Max: It’s a crutch, you know, you look at Portal: maybe is technically a shooting game?
Charlie: I don’t think so
Max: it’s obviously very different. That doesn’t necessarily mean non-violent, kid-friendly or whatever. I mean in some ways Portal is not either of those. It’s actually kind of extra violent in sadistic… - the characters. If you’re not relying on the crutch of “this is my core game mechanic: shooting a gun at somebody,†then you’re doing something more interesting. Like creating portals that bend space.
Charlie: right. Or having players create their own content like scribbling on walls and having to interact with other players or something weird. There was another point I was going to make about that - it’s lost. Oh the phi [??] obstructions that we talked about briefly. We could force ourselves or encourage other people to change their core gameplay in a big way by putting really serious constraints on your game idea.
Max: that is a constraint, “no shootingâ€.
Charlie: Yeah no shooting. Use a first person shooter engine, but make a non-shooter or make a shooter game without violence, which you get all those kind of camera games, where you take pictures of things in the world.
Max: Oh like Pokemon?
Charlie: Yeah all those games I didn’t play, but looked cool.
Max: they were very successful, so.
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> GAME INDUSTRY: SMALLER GAMES (18:50)<!--sizec--></span><!--/sizec-->
Charlie: yeah, that’s great, so… anyways, there’s that and… We also think that industry in general, you industry, should make smaller games.
Max: Yeah think the same thing. Like Portal or Guitar Hero. You don’t have to follow the path that has already been trod, because you don’t have the same concerns about creating a mega blockbuster. It’s the same thing with movies. You know that your Arnold Schwarzenegger big action movie is going to sell a whole bunch of money. You don’t know that, say Blair Witch, like smaller thing, you don’t know that it’s gonna be successful.
Charlie: Right
Max: But that means you probably get some level of success and then something like Blair Witch obviously you have an opportunity for huge amount of success. So you can do something that’s a little more risky, if it’s small, ‘cause you’re not investing a lot in it.
Charlie: and that’s a good constraint right there, because that forces you to use unknown talent. Which is one of the reason why Blair Witch was so big is because they felt like kids you knew. If it was Tom Cruise, then the effect would be totally ruined. It’s like “Kate Winslet as crying girl,†you know that wouldn’t work. But what Da Vinci said…
Max: yeah this quote, I get, right.
Charlie: with constraints come freedom, actually that isn’t quite the quote. But that’s the idea. Constrains give you the freedom
Max: paraphrase
Charlie: Or “Let the constraints liberate you� or I don’t know. I’m gonna look this up right now.
Max: the correct quote
Charlie: So, make smaller games
Max: But stuff like, good examples would be <a href="http://pc.ign.com/articles/740/740863p1.html" target="_blank">DEFCON </a>or <a href="http://www.youtube.com/watch?v=Taiurn541SE" target="_blank">Little Big Planet</a>.
Charlie: though that doesn’t look too small, especially with the physics tech.
Max: Uh it’s small, in the scope of like Halo 3
Charlie: True, or World of Warcraft or something like that.
Max: These are like, World of Warcraft is like a $50 million dollar game. Little Big Planet is like a $5 million game. That would be a lot of money for that project.
Charlie: True. That’s true. I know there are a ton of other examples. Probably the flash MMOs are probably down at that level. What else? We’re starting to see them at any rate.
Max: well I think the market is becoming better for Indy developers who are inherently, like us, who are inherently in the budgetary bracket.
Charlie: Way less than 5 million.
Max: I mean we can’t make a blockbuster in terms of content. Hopefully in terms of sales it’s a blockbuster.
Charlie: Hopefully in gameplay we can make a blockbuster. We can rival anything that anyone else can come up with.
Max: well that’s kind of the thing, it’s like if you’re making the Arnold Schwarzenegger movie, you’ve got Arnold Schwarzenegger and you have explosions and you have all this stuff so you don’t need to have a good story line.
Charlie: Right, it’s not what they’re there for. It’s a good analogy.
Max: Although maybe Arnold Schwarzenegger it’s like dating myself or something.
Charlie: It may be, seeing how he hasn’t made a movie in quite some time.
Max: I don’t think he’s making any more. Well he made Terminator 3, that wasn’t that long ago.
Charlie: That’s true
Max: That was probably like 15 years ago.
Charlie: We will edit that part out as well. Well those are the main topics we wanted to hit. And we’re up to 22 minutes, so
Max: wow
Charlie: We can really blather on, can’t we? Anything else you want to say?
<!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo--> MORE MEDIA FOR NS2 (22:43) <!--sizec--></span><!--/sizec-->
Max: I guess it’s not really a resolution, because it’s really just the plan, but something people are probably hoping for that we’re planning on doing is releasing a lot more media from the game. Oh yeah. Probably more video some screen shots and stuff like that.
Charlie: yep
Max: So that can be a resolution that we will definitely fulfill when we fail at all these other resolutions.
Charlie: or fail to try even. Yeah but we will be releasing a lot more screen shots and … alright well, thanks for listening and we’ll see you next week.
Max: bye
Charlie: bye (23:20)
When I said <i>deathmatch</i> I was talking about the historical and engaging arcade-speed and that comes to mind from the word. I wasn't trying to suppress NS's future rts aspects. I like them too.
Somewhere, I forget where, the dev team mentioned they wanted to focus more on the RTS aspects for NS2. You're saying it's important to realize that the core of NS1 is an incredibly diverse and well balanced death match if I'm not mistaken. Commanders could have the most fun RTS experience ever, but if there's no depth at the FPS level, no one will want to play the game competitively, if at all. If my interpretation is correct, I agree with you.
Edit: I chuckled at your paraphrasing, if an exaggeration, of Canadian's post.
<!--quoteo(post=1666908:date=Jan 9 2008, 01:55 PM:name=locallyunscene)--><div class='quotetop'>QUOTE(locallyunscene @ Jan 9 2008, 01:55 PM) <a href="index.php?act=findpost&pid=1666908"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Edit: I chuckled at your paraphrasing, if an exaggeration, of Canadian's post.<!--QuoteEnd--></div><!--QuoteEEnd-->
I know I didn't chuckle. That isn't paraphrasing, that is a smear. Being misquoted, having one's words twisted in such a fashion is nothing short of lying. It presents my stance on NS being more than death match like some rant from a cap lock key stuck ranting CS fanatic which is straight up bull ######. I have a few explicit words for troll tactic users.
Consider what I really said:
<!--quoteo(post=1666816:date=Jan 8 2008, 09:22 PM:name=CanadianWolverine)--><div class='quotetop'>QUOTE(CanadianWolverine @ Jan 8 2008, 09:22 PM) <a href="index.php?act=findpost&pid=1666816"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Come to think of it, if all I wanted was Death Match, I would rather be playing Unreal Tournament and its extras to that than NS. Yeah, NS is definitely more than just Death Match.<!--QuoteEnd--></div><!--QuoteEEnd-->
<b>I don't think Death Match is bad.</b> In fact, I enjoy death match games a great deal, back to the days of the first widely spread death match games, such as Quake. So don't go mis-representing my opinions, its not appreciated at all.
<!--quoteo(post=1666907:date=Jan 9 2008, 01:28 PM:name=Radix)--><div class='quotetop'>QUOTE(Radix @ Jan 9 2008, 01:28 PM) <a href="index.php?act=findpost&pid=1666907"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->When I said <i>deathmatch</i> I was talking about the historical and engaging arcade-speed and that comes to mind from the word. I wasn't trying to suppress NS's future rts aspects. I like them too.<!--QuoteEnd--></div><!--QuoteEEnd-->
Well, thank you for clarifying that. I also enjoy the engaging arcade speed and would love to see an implementation in NS2 that allows for action within the frame work of the Marine experience, such as acrobatic animations and correlating defensive gains and offensive downplaying from it - such as dives, rolls, wall manuevers, flips, etc (aka Free Running / Parkour). Do you think your earlier mantra still stands then?
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->NS is a deathmatch game. Please maintain that mantra in NS2.<!--QuoteEnd--></div><!--QuoteEEnd-->
Yes NS2 will be made on the Source engine, the latest version of the source engine, which is ep2 (officially HL2 ep1) standards not Half-Life 1 as NS was.
The NS2TR is the pack that they will release for people wanting to make maps.
Decoda, not encoda, is as far as I remember a thing to help you debug Lua code.
Anything else that you are confused about?
Most of that stuff is here <a href="http://www.unknownworlds.com/ns2/faq/" target="_blank">http://www.unknownworlds.com/ns2/faq/</a>
EDIT: I also suggest that instead of working on random mods StixNStonz, where you may never see your map being released, you go out and buy TF2 or CS:S and make a map for them. That way at least you know people will be play testing them at the end of it and you will at least get some kind of feedback as to what people like and don't like. It also means that if the NS2TR comes out when your in mid map, no one will care if you just drop it or release anything more than a beta version. Who knows you might make the next map that’s played as much as Dust <img src="style_emoticons/<#EMO_DIR#>/tounge.gif" style="vertical-align:middle" emoid=":p" border="0" alt="tounge.gif" />
I wouldn't go so far as to call him a liar. I agree that his quote of you had nothing to do with what you actually said, but it seems to me it was just that he didn't bother reading your post before jumping to conclusions, as opposed to intentionally misrepresenting your opinion.
Ahh, alright, I can swing with that. *Breaks out his edit scalpel* I'll give the shadow of a doubt that it was unintentional ######baggery.
Thanks afratnikov for the text of the podcast. I think that is pretty cool about this community that we are getting involved like this, I hope you guys get some sort of mention in the credits or a cameo of sorts like your name in small print on a computer screen in one of the maps.
I actually began my mapping career in CS:S. Bought HL2, got back into CS for a month or two, started mapping for it, then went back to playing NS. So i was mapping for CS:S for about a year while playing NS.
I made two fun maps:
<u><b>De_Cover</u></b>. Three buildings facing each other, totalling over 150 windows, with the bombsite in the middle. CTs mostly spawn in the buildings with snipers/paras, with a few on the ground with subs. T's spawn mostly under the tarp with rifles and smoke grenades, with a few spawning in the buildings for counter-sniping. There's a ton of cover in the middle of all this. People said it felt like a modern-day Normandy.
This was my first released map, and it got a good amount of playtime on a few servers. Many people complained about the visuals, so I optimized the map significantly and adding in MUCH more realistic looking buildings, but never got around to releasing the new version. Here are some screens from the old version.
<a href="http://img162.imageshack.us/my.php?image=decoverb200045qb.jpg" target="_blank"><img src="http://img162.imageshack.us/img162/6298/decoverb200045qb.th.jpg" border="0" class="linked-image" /></a><a href="http://img162.imageshack.us/my.php?image=decoverb200030kj.jpg" target="_blank"><img src="http://img162.imageshack.us/img162/3207/decoverb200030kj.th.jpg" border="0" class="linked-image" /></a>
<u><b>1v1_scoutflash</u></b>. This was actually for a 'what-can-you-fit-in-a-small-box' competition, the box being about 20 feet in all 3 directions. I created a surprisingly cool little 1v1/2v2 map, set atop a highrise building.
Using a highly optimized 3d skybox (which was under the floor), it simulated a much larger rooftop, and showed the surrounding city. Players could fall off the rooftop on two sides, fall to their death screaming, and you'd see a blood splatter on the street afterwards (go go trigger system). The players spawned with a Scout and two flashbangs (not in sight of each other regardless of spawns), and I somehow fit a decent amount of cover and height variation in this area. There are two semi-hidden knives, a well-hidden grenade, and a secret that teleports the player into the skybox, creating the Godzilla effect.
<a href="http://img386.imageshack.us/my.php?image=1v1scoutflashb100059ej.jpg" target="_blank"><img src="http://img386.imageshack.us/img386/2985/1v1scoutflashb100059ej.th.jpg" border="0" class="linked-image" /></a><a href="http://img386.imageshack.us/my.php?image=1v1scoutflashb100039gr.jpg" target="_blank"><img src="http://img386.imageshack.us/img386/2783/1v1scoutflashb100039gr.th.jpg" border="0" class="linked-image" /></a><a href="http://img393.imageshack.us/my.php?image=1v1scoutflashb2h00031ld.jpg" target="_blank"><img src="http://img393.imageshack.us/img393/2346/1v1scoutflashb2h00031ld.th.jpg" border="0" class="linked-image" /></a><a href="http://img386.imageshack.us/my.php?image=1v1scoutflashb100009co.jpg" target="_blank"><img src="http://img386.imageshack.us/img386/5409/1v1scoutflashb100009co.th.jpg" border="0" class="linked-image" /></a>
Then there was <u><b>cs_abscond</u></b>. This was my first serious map. De_Cover taught me the basics, the tricks (skyboxes, displacements, etc) and about optimization, and then 1v1_Scoutflash taught me about real visuals and 'keeping it tight', as i'd say. I had big, big plans for Abscond.
Some highlights: an Embassy with a full suite of security systems (laser trip beams, one-day doors, video cameras, all done), Source's first Shotgun-Breachable door (built and working; only a shotgun shell to the handle could open the door from the outside, thoroughly tested), JFK-style sniping from the sewers in the sidewalks, and a huge scripted Hostage rescue scene involving a helicopter extraction and C4 under the helipad (with a very long fuse, for that final 'scene'), all of this already built and good to go.
The problem was with the map itself. Although i knew all the tricks (especially the triggers) and how it all worked, my fundamental mapping skill was still too noob.
About halfway through this map, I realized how silly it was to be mapping for a game I never played (CS:S), so I decided to try and learn HL1 mapping for NS. Thats when the 16 month process of creating <u><b>Ns_Nexus</b></u> began, during which I had multiple 'mapper enlightenments', often going back and fixing up everything I'd built so far.
HL1 mapping has given me one hell of a solid mapping foundation, so now I'm ready to go back into Source and absolutely tear it up with that Sexy McSexy engine. But I dont want to do CS:S again. I want NS2.
**********************************
So, after giving you my entire mapping history, I will say this. If i can't map for NS2 (since even the TR seems nowhere in the near future), I won't map for CS:S. That game is already flooded with maps, the majority being horrible but still with a TON of really amazing custom maps. Even with a huge effort, whatever map I'd make would get lost in the clutter.
I'm looking for a mod with promising gameplay, for whom I can deliver a stellar map. That way, its almost assured to become one of their 'official' maps, which is what I really need for my portfolio right now. Despite Nexus being roughly on par in terms of scope, depth and visuals as Machina, there's been no peep about its official inclusion, and the Official Map Pack seems to have died out months ago.
Max: I think it’s a Patton quote. "
"No battle plan survives contact with the enemy." -<a href="http://en.wikipedia.org/wiki/Helmuth_von_Moltke_the_Elder" target="_blank">Helmuth von Moltke</a>
It's misquoted in COD4 to Colin Powell.
<b>Estimating</b>
There's are some nice techniques for improving estimation. One useful concept is the 'cone of uncertainty' :)
Briefly you can think of a simple chart with two lines which start wide apart and converge as they move right. The large gap on the left is a reflection of a high level of uncertainty in an estimate. For example your about to develop a feature and there are lots of unknowns, potential problems, questions etc, so you're estimate is really a range: 2 days - 4 weeks. It's important to change you're thinking away from choosing a single number to a possible range. As you spend time on a task/project you learn the gotchas, find rabbit holes you never imagined, develop ways to work faster, so as time moves forward your uncertainty drops and the range narrows.
Using a model like this you can easily think of tasks which have great ranges (lots of unknowns) and use some simple formulas to calculate schedules based on task ranges. Once you use a technique like this the typical approach of trying to come up with a single number and throw on a high simplistic modifier (i.e it usually takes me twice as long) seems ridiculous. Simply put - some tasks can be accurately estimated others can't be - a model like this takes that into account. And you can easily see how you can revisit your estimates as you reduce the uncertainty.
At the end of the day, a big uncertainty is a major change in game direction and that could really through off the high level schedule! There's only so much uncertainty you can remove. <a href='index.php?act=findpost&pid=0'>Cone of Uncertainty</a>
<b>Unit Testing</b>
As for as unit testing goes it definitely has a value in typical software development, I can't see how that wouldn't apply in game development. Ultimately as you mentioned it's no silver bullet and if you don't have the skills to write good unit tests it's simply a waste of time. There's a cultural aspect to it, if you believe it's a waste time it more than likely will end up that way, it has to suit the teams mentality and skills. The ability to have repeatable testing to ensure more rigorous testing and allow the impact of changes to be more quickly explored can only be good especially when there will be areas of code with very iteration levels. All end user driven software is going to have a significant user acceptance component for the user experience component of testing (i.e. is it fun and easy to use) but it shouldn't be the only approach.
You don't have to choose either or, you can have a hybrid approach with good algorythmic planning and some level of simple debug time. However some components will be subject to a high level of iteration which would make the overhead of re-writing unit tests a burden, but for more stable core components why not? Ultimately I think the time saved with test drive development has a place in any serious project.
My guess would be the transcript of the podcast.