You have a fair point Yojimbo, and I am sorry if I was accusatory in my tone. We all want this thing to go somewhere, and I am happy to concede on any point that I see I was mistaken on. I think one error you might be making however, is not seeing the problem of player retention as multi-faceted. What I am saying is, it's not one thing against another, its just each thing being part of the picture (performance, features and content, bug fixing all play a role). In technical terms you may be correct, perhaps we have exhausted the engine's capabilities to pump out more frames. In which case, you are 100% correct - focus on the other stuff. I think what Bonage was stating is that performance is still an issue for low end systems, and that given agreement that this is something to care about, what could be done to address it.
No need to say sorry mate lol, I am just stating my point that we are approaching a threshold on how much performance we can squeeze out of SPARK, there is no miracle performance fix that will increase everyone's performance by 20% @McGlaspie can probably clarify this, such fixes have already been realised aka load times and a mysterious intel core i** bug that was discovered and fixed IIRC alittle while back @SamusDroid remembers.
I understand perfectly that we all want this game to run well on all systems but that's a very tall order to ask for, we have to compromise at some point and bite the bullet and say "Well it's almost 2016 and ironically of all the days TODAY is the day to upgrade (BLACK FRIDAY) mega cheap electrical components.
SamusDroidColoradoJoin Date: 2013-05-13Member: 185219Members, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Gold, Subnautica Playtester, NS2 Community Developer, Pistachionauts
Yojimbo is partially right. The CDT has significantly improved performance game wide, from load times, to FPS, server stability and performance, and hit reg, with more fixes coming very soon. Certainly there are more possible improvements that can be made, but at this point we can start focusing more on other aspects of the game.
Yojimbo is partially right. The CDT has significantly improved performance game wide, from load times, to FPS, server stability and performance, and hit reg, with more fixes coming very soon. Certainly there are more possible improvements that can be made, but at this point we can start focusing more on other aspects of the game.
The focus can shift, and other issues should definitely be addressed as well, but there are still a large amount of people out there who encounter performance issues. Unfortunately these are the people that the game hemorrhages, which leads to a rise in comments like 'this game is unbalanced because there is a big disparity in skill level between the remaining players in the community'. It's a cyclical problem that feeds into itself.
Is it the only problem that needs addressing? Of course not, but the reality is the game is a multiplayer fps and unfortunately in this genre many people judge a game by performance first and foremost.
I'm just trying to understand what the limitations of the game/Spark are, and just how far performance in the game can be still yet be improved for lower/mid tier systems before hitting a wall.
SamusDroidColoradoJoin Date: 2013-05-13Member: 185219Members, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Gold, Subnautica Playtester, NS2 Community Developer, Pistachionauts
edited November 2015
I can't speak on the entirety of the CDT, but that imaginary wall is likely getting close. With the exclusion of slight performance increases, not many significant performance improvements will likely be added at this point. The game runs quite well now. Yes there's a hired team now, but I don't see that impacting the issue you are raising. The amount of progress has been so significant though that you never know what could happen though. The focus is likely to shift elsewhere.
I have some semi recent experience playing ns2 on low end systems to use a reference.
I think 3 patches back I played ns2 for awhile with a 2.7ghz g1620 celeron and a 750ti. It was extremely cpu bottlenecked. I would consider this a mid low system in terms of performance. Most of the time I was getting 120fps, but if all 20+ players were in a room together I would drop to 40 fps. There have been performance improvements since then, especially on the cpu side.
Many patches back, when the 4770k's came out I had one but did not have a gpu in the system. I was stuck with intel hd 4600 for a week. That was heavily gpu bound. I sat at 40 fps pretty much always. I never dropped fps as I remember it.
IeptBarakatThe most difficult name to speak ingame.Join Date: 2009-07-10Member: 68107Members, Constellation, NS2 Playtester, Squad Five Blue, NS2 Map Tester, Reinforced - Diamond, Reinforced - Shadow
Ironhorse is gonna beat me for bringing this up again, but when or after the hitch fix for the texture streamer happened my textures started taking forever to load, like 5-10 textures loaded per second with 600+ in the queue. Needless to say my game is a blurry mess for about a minute and the game itself performs just fine before and after.
Is there any way the new team can add some options regarding the texture streamer in how smooth or fast the game handles texture loading? I'm fine with dropping a few frames just to make sure my game loads properly.
(Every single time I want to play now since that update I have to wait for the main menu to load, go into customization and view every playermodel to make sure they load properly, then I join a game and wait in the ready room for about a minute to load the common textures, minimap, and scoreboard, then finally I join the game and wait 30 seconds for my hud and viewmodel to load.)
Every single time. My Hardware has not changed and this texture loading bottleneck issue is exclusive to this game only. There is also barely any difference between the load times and performance of low and medium textures either.
So let's assume that NS2 is here to stay for a while and the possibility of an UE4+NS3 combo that the OP describes is off the cards for a while yet.
What needs to be done to improve spark performance across the board? Do we know what/where the current bottlenecks are? If so, do we know how hard they are to fix?
In line with UWE's desire to grow the game, will we see a renewed focus on ironing out performance issues so that players from a wide range of PC hardware are able to play the game with more stable framerates?
I understand that the new team hasn't been hired to address this issue directly, but assuming the other work they do is successful, can we expect to see this core issue for player retention addressed down the line ?
I can only speak for myself, but there are definitely areas of the rendering pipeline I've wanted to look at for quite a while: Shadow Maps, Ambient Occlusion, Specular Maps, and Anti-Aliasing. My personal to-do list is quite long, so I cannot make any promises. Some time ago I looked at each of the rendering layers to determine which ones were the most expensive, AO and Shadow Maps came out on top (in terms of per-frame cost). If I can find the time, I intend on seeing what I can do to reduce the runtime cost of those two items. I'll focus on Shadow Maps first when/if I can get to it.
Please understand, there are limited gains by improving the renderer. In cases where the game dips below 60fps, rendering times sort of don't matter at that point, because the main limiter cycles back to a CPU-bound issue. Performance and optimization for games is a very complicated and tricky affair. You have to go after areas that have the most overarching impact, regardless of scope in comparison with their frequency. The only reason to even focus on the Rendering Pipeline now, is because of the strides made by the CDT over the past year and a half. In the end, any rendering optimizations will basically only serve to increase the maximum fps, and as a result gain more perceived rendering stability (not actual).
So let's assume that NS2 is here to stay for a while and the possibility of an UE4+NS3 combo that the OP describes is off the cards for a while yet.
What needs to be done to improve spark performance across the board? Do we know what/where the current bottlenecks are? If so, do we know how hard they are to fix?
In line with UWE's desire to grow the game, will we see a renewed focus on ironing out performance issues so that players from a wide range of PC hardware are able to play the game with more stable framerates?
I understand that the new team hasn't been hired to address this issue directly, but assuming the other work they do is successful, can we expect to see this core issue for player retention addressed down the line ?
I can only speak for myself, but there are definitely areas of the rendering pipeline I've wanted to look at for quite a while: Shadow Maps, Ambient Occlusion, Specular Maps, and Anti-Aliasing. My personal to-do list is quite long, so I cannot make any promises. Some time ago I looked at each of the rendering layers to determine which ones were the most expensive, AO and Shadow Maps came out on top (in terms of per-frame cost). If I can find the time, I intend on seeing what I can do to reduce the runtime cost of those two items. I'll focus on Shadow Maps first when/if I can get to it.
Please understand, there are limited gains by improving the renderer. In cases where the game dips below 60fps, rendering times sort of don't matter at that point, because the main limiter cycles back to a CPU-bound issue. Performance and optimization for games is a very complicated and tricky affair. You have to go after areas that have the most overarching impact, regardless of scope in comparison with their frequency. The only reason to even focus on the Rendering Pipeline now, is because of the strides made by the CDT over the past year and a half. In the end, any rendering optimizations will basically only serve to increase the maximum fps, and as a result gain more perceived rendering stability (not actual).
@IeptBarakat is not alone with this problem. We from Wooza have the most exposure to 'random joe' and that includes their problems. I know there a handful of players who also suffer from this.
There are certainly still many ways to improve performance, some of them are now old, neglected trello cards.
While things have, without doubt, been improving, if you now start adding stuff and forget about performance, the same thing as last year will happen. Perf will drop because you keep adding stuff without regard for it.
@bonage is correct. Our exposure gives us a perceived 15% of players that have significant performance issues today. (Most of them are rookies, warping along. It is a sad sight.)
I can only speak for myself, but there are definitely areas of the rendering pipeline I've wanted to look at for quite a while: Shadow Maps, Ambient Occlusion, Specular Maps, and Anti-Aliasing.
@McGlaspie Does this mean that when you have all these things disabled there is nothing else you can improve on for FPS? Specular maps are the only thing in that list I don't already have disabled, because well I can't.
McGlaspiewww.team156.comJoin Date: 2010-07-26Member: 73044Members, Super Administrators, Forum Admins, NS2 Developer, NS2 Playtester, Squad Five Blue, Squad Five Silver, Squad Five Gold, Reinforced - Onos, WC 2013 - Gold, Subnautica Playtester
@Cr4zyb4st4rdAny graphical options will always have a run-time cost. Always. Nothing in 3D Rendering is "free". Just drawing a simple triangle on screen has a measurable cost. Now over time improvements in both hardware and software have _significantly_ improved some facets of Rendering, but still to this day some features are expensive. Ambient Occlusion for example (Screen-Space or World-Space centric) is very expensive. Shadows have always been expensive because they require the World be rendered from a Light's "View-Perspective", for each one (that casts shadows). There isn't some secret "ultra-minimal" config you can do with Spark that massively improves _Rendering_ times. It just comes down to what your PC is capable of in comparison with what graphical options are enabled. If you're asking me if there is some "hidden" config setting tweak you can make to gain fps? Nope, I'm afraid not. Also, specular maps cannot be toggled on/off. They are part of the materials (textures) the various art assets (models, decals, etc) use. In fact, Specular Maps are already downsampled by 50% right now. What I want to change with them is allowing a "Low/High" option. So, High-SpecMaps would not be downsample, and "Low" would be what we have today. The run-time cost of Specular Maps is so low, that you would gain nothing perceivable by outright disabling them. Also we will not be adding new options to control Spec-Maps. That would introduce far too many complexities (and potential bugs) for it to be worthwhile.
For the sake of clarity, when and if I get the time to work on the Rendering Pipeline, I'm not going to 100% be focused on pure performance. I also want to make some upgrades for the sake of visual quality as well. Anti-Aliasing is a good example of this. Current AA is the FXAA algorithm. I want to see if I can integrate the SMAA-T2 algorithm as a "High" option, because it visually looks much better. I'd then leave the FXAA as a "Low" option, and of course still allow AA to be disabled. It's a balancing act of implementing changes where they'll yield the most gains (visually or performance) for the greatest number of people. The last thing I want to do is attempt to force someone to have to leave some graphics option enabled.
matsoMaster of PatchesJoin Date: 2002-11-05Member: 7000Members, Forum Moderators, NS2 Developer, Constellation, NS2 Playtester, Squad Five Blue, Squad Five Silver, Squad Five Gold, Reinforced - Shadow, NS2 Community Developer
While performance can still be increased, all of the easy and most of the not-so-easy stuff has been done. Getting more performance out of the engine is certainly "possible" - if you include stuff like re-architecting the engine and rewriting most of the lua code in the definition of "possible".
And while I would certainly enjoy squeezing more performance out of the engine, we are getting to the point the performance is no longer the complete blocker to NS2 success that it used to be.
Which means it is time to start looking at all the ideas that has come up over the years on how to make it a better game.
Soul_RiderMod BeanJoin Date: 2004-06-19Member: 29388Members, Constellation, Squad Five Blue
I would expect not a huge amount of gain left in the engine. It is being redeveloped as a component based engine for FP, and that should have many performance improvements over time, but you won't see NS2 re-written to work on a component based spark.
You could argue that DX12 or Vulcan could be of great benefit to NS2, because of the reduction on CPU usage etc, but is it really worth that kind of programming investment in NS2?
Not to mention DX12 or Vulcan on its own won't offer a reduction in CPU usage. It has to be optimized for it too. NS2 has DX11 and OpenGL already but both are buggy and in a perpetual beta state.
Not to mention DX12 or Vulcan on its own won't offer a reduction in CPU usage. It has to be optimized for it too. NS2 has DX11 and OpenGL already but both are buggy and in a perpetual beta state.
Vulkan offers multiple core support, it utalises more of each of each core of the multicore processors in a more efficent way, actually it's better this way because NS2 doesn't support using a full core even when multicore processors are in use.
Not to mention DX12 or Vulcan on its own won't offer a reduction in CPU usage. It has to be optimized for it too. NS2 has DX11 and OpenGL already but both are buggy and in a perpetual beta state.
Vulkan offers multiple core support, it utalises more of each of each core of the multicore processors in a more efficent way, actually it's better this way because NS2 doesn't support using a full core even when multicore processors are in use.
"split workload equally over multiple threads"
"scales effectively"
"increases gpu power to do more processing"
The problem is when Max implemented multithredding, he didn't have enough time to spend on optimization and fully implementing it.
Yes, but that doesn't just happen. If it was as simple as flipping a switch, we'd have done it by now. The game engine has to be built with that in mind.
Not to mention DX12 or Vulcan on its own won't offer a reduction in CPU usage. It has to be optimized for it too. NS2 has DX11 and OpenGL already but both are buggy and in a perpetual beta state.
Vulkan offers multiple core support, it utalises more of each of each core of the multicore processors in a more efficent way, actually it's better this way because NS2 doesn't support using a full core even when multicore processors are in use.
"split workload equally over multiple threads"
"scales effectively"
"increases gpu power to do more processing"
The problem is when Max implemented multithredding, he didn't have enough time to spend on optimization and fully implementing it.
Yes, but that doesn't just happen. If it was as simple as flipping a switch, we'd have done it by now. The game engine has to be built with that in mind.
I never said it "just happen" i am just saying that Norid can't say it doesn't give any reduction in CPU usage, it might not give the gains you want, but as far as most of you have said there's not much left you can do to fix that yet till you get more under the hood of Spark engine and fiddle around, besides you'll be employed for a wee bit and you should be able to have your hands on Vulkan pretty soon, it wouldn't hurt to give it a try with some prototyping, maybe you could get a hand from Max to implement a small element to test in something like futureperfect or the upcoming version of Spark that is being worked on?
I never said it "just happen" i am just saying that Norid can't say it doesn't give any reduction in CPU usage,
Maybe I said it odd, but the basic meaning of what I said is the same as beiges comment. I am not saying it doesn't give any reduction, but that it has to be optimized for it. This is to me just like beige saying the engine needs to be built with that in mind. For example, DX11 and OpenGl are both buggy and in a perpetual beta state.
ArmymongerThe spacebearing Country of AmericaJoin Date: 2014-04-04Member: 195135Members
Personally, with all this talk of improving upon graphics and processing power and such for a NS3 pipe dream...(which has as much imaginative strength has Half life 3...)
I'd like make an interjection that improving upon the graphics for ns2 is...not really that appealing to me, gameplay has priority over graphics in my book.
To be fair, Ns2 does sport some already solid gameplay and experiences so I can understand why some would be pushing for a graphical upgrade..but...
Ns2 already looks pretty damn good...why is it necessary to push the limits even more?
I'd rather have new content, like new maps, a single player campaign separate for each side to ease newbies into the Multiplayer fires, new game modes to add some salt and sugar to my meals of gorge bacon, perhaps even some contradicting new weapons to ruin the status quo of balance.
Stuff that would entertain me for a long time to come rather than graphical power that would bore me after 1 or two matches and force me to reevaluate the inner workings of my computer.
Metaphorically speaking it would be like dressing a woman with copious amounts of makeup, frilly dresses, and shiny jewelry to be the angel of the party, only to realize that her personality, and inner self be something that came from the very depths of hell. Something very contradicting that what she was dressed up to be.
Just some of my two cents right there.
With some dating advice I guess.
While performance can still be increased, all of the easy and most of the not-so-easy stuff has been done. Getting more performance out of the engine is certainly "possible" - if you include stuff like re-architecting the engine and rewriting most of the lua code in the definition of "possible".
And while I would certainly enjoy squeezing more performance out of the engine, we are getting to the point the performance is no longer the complete blocker to NS2 success that it used to be.
Which means it is time to start looking at all the ideas that has come up over the years on how to make it a better game.
matsoMaster of PatchesJoin Date: 2002-11-05Member: 7000Members, Forum Moderators, NS2 Developer, Constellation, NS2 Playtester, Squad Five Blue, Squad Five Silver, Squad Five Gold, Reinforced - Shadow, NS2 Community Developer
I've also noticed that on big rooms and great distances have made players disapear....
Mm... relevancy distance is 40m, so the engine offloads entities beyond that range. There is a reason why there are usually large crates breaking long sight distances - mappers have to hide that.
That, bw, is a performance optimization to keep your client from having to update to many entities. The high client-side per-entitiy cost is one of the more glaring weaknesses in the Spark Engine.
matsoMaster of PatchesJoin Date: 2002-11-05Member: 7000Members, Forum Moderators, NS2 Developer, Constellation, NS2 Playtester, Squad Five Blue, Squad Five Silver, Squad Five Gold, Reinforced - Shadow, NS2 Community Developer
I would once again, as I have in the past, suggest that the developers of NS2 look to their competition for what the game is missing.
Some of the most obvious things are a good in depth tutorial, GOTV, a matchmaking system that allows you to queue as commander or player role etc.
These aren't easily coded features, but if you want the game to come to par with it's competition - it should be considered.
The problem isn't really lack of ideas ... it's more a lack of people skilled and interested enough in implementing them (and priorities; not much use trying to attract people to NS2 before the engine can run the game well enough).
So it takes a lot of time instead.
But yea, single/few player content for learning/training is pretty high up on the wishlist.
Comments
No need to say sorry mate lol, I am just stating my point that we are approaching a threshold on how much performance we can squeeze out of SPARK, there is no miracle performance fix that will increase everyone's performance by 20% @McGlaspie can probably clarify this, such fixes have already been realised aka load times and a mysterious intel core i** bug that was discovered and fixed IIRC alittle while back @SamusDroid remembers.
I understand perfectly that we all want this game to run well on all systems but that's a very tall order to ask for, we have to compromise at some point and bite the bullet and say "Well it's almost 2016 and ironically of all the days TODAY is the day to upgrade (BLACK FRIDAY) mega cheap electrical components.
Disregarding the tiny tweaks to values and map changes
Knew you wouldn't let me down haha
The focus can shift, and other issues should definitely be addressed as well, but there are still a large amount of people out there who encounter performance issues. Unfortunately these are the people that the game hemorrhages, which leads to a rise in comments like 'this game is unbalanced because there is a big disparity in skill level between the remaining players in the community'. It's a cyclical problem that feeds into itself.
Is it the only problem that needs addressing? Of course not, but the reality is the game is a multiplayer fps and unfortunately in this genre many people judge a game by performance first and foremost.
I'm just trying to understand what the limitations of the game/Spark are, and just how far performance in the game can be still yet be improved for lower/mid tier systems before hitting a wall.
I think 3 patches back I played ns2 for awhile with a 2.7ghz g1620 celeron and a 750ti. It was extremely cpu bottlenecked. I would consider this a mid low system in terms of performance. Most of the time I was getting 120fps, but if all 20+ players were in a room together I would drop to 40 fps. There have been performance improvements since then, especially on the cpu side.
Many patches back, when the 4770k's came out I had one but did not have a gpu in the system. I was stuck with intel hd 4600 for a week. That was heavily gpu bound. I sat at 40 fps pretty much always. I never dropped fps as I remember it.
Is there any way the new team can add some options regarding the texture streamer in how smooth or fast the game handles texture loading? I'm fine with dropping a few frames just to make sure my game loads properly.
(Every single time I want to play now since that update I have to wait for the main menu to load, go into customization and view every playermodel to make sure they load properly, then I join a game and wait in the ready room for about a minute to load the common textures, minimap, and scoreboard, then finally I join the game and wait 30 seconds for my hud and viewmodel to load.)
Every single time. My Hardware has not changed and this texture loading bottleneck issue is exclusive to this game only. There is also barely any difference between the load times and performance of low and medium textures either.
I can only speak for myself, but there are definitely areas of the rendering pipeline I've wanted to look at for quite a while: Shadow Maps, Ambient Occlusion, Specular Maps, and Anti-Aliasing. My personal to-do list is quite long, so I cannot make any promises. Some time ago I looked at each of the rendering layers to determine which ones were the most expensive, AO and Shadow Maps came out on top (in terms of per-frame cost). If I can find the time, I intend on seeing what I can do to reduce the runtime cost of those two items. I'll focus on Shadow Maps first when/if I can get to it.
Please understand, there are limited gains by improving the renderer. In cases where the game dips below 60fps, rendering times sort of don't matter at that point, because the main limiter cycles back to a CPU-bound issue. Performance and optimization for games is a very complicated and tricky affair. You have to go after areas that have the most overarching impact, regardless of scope in comparison with their frequency. The only reason to even focus on the Rendering Pipeline now, is because of the strides made by the CDT over the past year and a half. In the end, any rendering optimizations will basically only serve to increase the maximum fps, and as a result gain more perceived rendering stability (not actual).
Yeah!!...... What he said!
There are certainly still many ways to improve performance, some of them are now old, neglected trello cards.
While things have, without doubt, been improving, if you now start adding stuff and forget about performance, the same thing as last year will happen. Perf will drop because you keep adding stuff without regard for it.
@bonage is correct. Our exposure gives us a perceived 15% of players that have significant performance issues today. (Most of them are rookies, warping along. It is a sad sight.)
@McGlaspie Does this mean that when you have all these things disabled there is nothing else you can improve on for FPS? Specular maps are the only thing in that list I don't already have disabled, because well I can't.
For the sake of clarity, when and if I get the time to work on the Rendering Pipeline, I'm not going to 100% be focused on pure performance. I also want to make some upgrades for the sake of visual quality as well. Anti-Aliasing is a good example of this. Current AA is the FXAA algorithm. I want to see if I can integrate the SMAA-T2 algorithm as a "High" option, because it visually looks much better. I'd then leave the FXAA as a "Low" option, and of course still allow AA to be disabled. It's a balancing act of implementing changes where they'll yield the most gains (visually or performance) for the greatest number of people. The last thing I want to do is attempt to force someone to have to leave some graphics option enabled.
And while I would certainly enjoy squeezing more performance out of the engine, we are getting to the point the performance is no longer the complete blocker to NS2 success that it used to be.
Which means it is time to start looking at all the ideas that has come up over the years on how to make it a better game.
You could argue that DX12 or Vulcan could be of great benefit to NS2, because of the reduction on CPU usage etc, but is it really worth that kind of programming investment in NS2?
One would have to say probably not.
Hopefully the opportunity down the line arises where you guys get the time to look at the engine.
Grow up.
??? who is this in response to??
https://docs.unrealengine.com/latest/INT/Engine/Rendering/Materials/PhysicallyBased/index.html
Vulkan offers multiple core support, it utalises more of each of each core of the multicore processors in a more efficent way, actually it's better this way because NS2 doesn't support using a full core even when multicore processors are in use.
https://www.youtube.com/watch?time_continue=23&v=P_I8an8jXuM
"split workload equally over multiple threads"
"scales effectively"
"increases gpu power to do more processing"
The problem is when Max implemented multithredding, he didn't have enough time to spend on optimization and fully implementing it.
Yes, but that doesn't just happen. If it was as simple as flipping a switch, we'd have done it by now. The game engine has to be built with that in mind.
My point.
I want new exciting content and not just shiny stuff.
I'd like make an interjection that improving upon the graphics for ns2 is...not really that appealing to me, gameplay has priority over graphics in my book.
To be fair, Ns2 does sport some already solid gameplay and experiences so I can understand why some would be pushing for a graphical upgrade..but...
Ns2 already looks pretty damn good...why is it necessary to push the limits even more?
I'd rather have new content, like new maps, a single player campaign separate for each side to ease newbies into the Multiplayer fires, new game modes to add some salt and sugar to my meals of gorge bacon, perhaps even some contradicting new weapons to ruin the status quo of balance.
Stuff that would entertain me for a long time to come rather than graphical power that would bore me after 1 or two matches and force me to reevaluate the inner workings of my computer.
Metaphorically speaking it would be like dressing a woman with copious amounts of makeup, frilly dresses, and shiny jewelry to be the angel of the party, only to realize that her personality, and inner self be something that came from the very depths of hell. Something very contradicting that what she was dressed up to be.
Just some of my two cents right there.
With some dating advice I guess.
@matso
I would once again, as I have in the past, suggest that the developers of NS2 look to their competition for what the game is missing.
Some of the most obvious things are a good in depth tutorial, GOTV, a matchmaking system that allows you to queue as commander or player role etc.
These aren't easily coded features, but if you want the game to come to par with it's competition - it should be considered.
Mm... relevancy distance is 40m, so the engine offloads entities beyond that range. There is a reason why there are usually large crates breaking long sight distances - mappers have to hide that.
That, bw, is a performance optimization to keep your client from having to update to many entities. The high client-side per-entitiy cost is one of the more glaring weaknesses in the Spark Engine.
The problem isn't really lack of ideas ... it's more a lack of people skilled and interested enough in implementing them (and priorities; not much use trying to attract people to NS2 before the engine can run the game well enough).
So it takes a lot of time instead.
But yea, single/few player content for learning/training is pretty high up on the wishlist.