A plea: reduce the bloat and simplify ET Pro

[Lekdevil.NL]

I'm writing this post as a counterbalance to all the feature requests that keep appearing on this forum. While I understand the pressure the developers are under to keep adding features to ET Pro, it might be wise to stop for a moment and consider the way we're heading.

At the moment I'm working on the ET server configs for a big tournament. I haven't paid close attention to the development of ET Pro for some time, and right now I'm somewhat taken aback by the sheer number of configurable options that have been added over the past two years. Right now, there are a great deal of new cvars that directly influence the gameplay. At the same time, there is no way of knowing which settings are in use on a particular server. In my opinion, this creates a problem where a player has no idea what kind of gameplay to expect when connecting. It's a bit like the problem that has always plagued Shrubmod.

What I'd like to propose is to cut as many non-essential cvars and configurable features from ET Pro as possible, to simplify the configuration and to make the result more predictable to the player. Some examples:

• * Do we really need features like b_noactivatelean and b_flushitems, to name a few? Are those things you'd like to see enforced in a competition mod?
  * Do bug fixes need to be made optional? b_fallingbugfix comes to mind. Also, do we still need b_brokenlogtimestamps? Why would anyone continue to use the old log format?
  * Do we really need a chat text censor feature (b_cursefilter) in a competition mod?
  * Do enhancements to the engine need to be made optional? Is there really a need to ever turn b_antiwarp off, for instance? Do we really need to make b_realhead configurable? Why would we want to introduce two different hitbox models and let the player guess which one is in use? I'd suggest to just be bold, enable those features by default and remove the cvar.
  * Are people actually using the trick jump features? If only a small group is, does it need to be included in the main ET Pro mod, adding to its size and complexity? Wouldn't a small, dedicated mod for that specific purpose suffice?

Of course, these are just some random suggestions and are in no way ment to bash the ET Pro team, who are doing an excellent job so far. I'd just like to introduce the concepts of "less is more" and "if in doubt, simplify", and am curious what other people's thoughts are on this matter.

Cheers.

[gotenks]

* Do bug fixes need to be made optional? b_fallingbugfix comes to mind. Also, do we still need b_brokenlogtimestamps? Why would anyone continue to use the old log format?

yes, thhe broken log time stamps only effects server log parsers, which would break some. The falling bug fix, um don't really remember this one...

* Do we really need a chat text censor feature (b_cursefilter) in a competition mod?

it is a pub mod too... not primarily designed but doesn't really change competition settings

* Do enhancements to the engine need to be made optional? Is there really a need to ever turn b_antiwarp off, for instance? Do we really need to make b_realhead configurable? Why would we want to introduce two different hitbox models and let the player guess which one is in use? I'd suggest to just be bold, enable those features by default and remove the cvar.

i allways use them, but again changing game play too much could be negative... and iirc some competitions use b_realhead 0.

* Are people actually using the trick jump features? If only a small group is, does it need to be included in the main ET Pro mod, adding to its size and complexity? Wouldn't a small, dedicated mod for that specific purpose suffice?
[/list]

i've seen a couple people use this, not really nessicary but some demo people may beable to use it for clan info

Of course, these are just some random suggestions and are in no way ment to bash the ET Pro team, who are doing an excellent job so far. I'd just like to introduce the concepts of "less is more" and "if in doubt, simplify", and am curious what other people's thoughts are on this matter.

Cheers.

[ReyalP]

Do we really need features like b_noactivatelean and b_flushitems, to name a few? Are those things you'd like to see enforced in a competition mod?

b_noactivatelean is a client side option, created in response to specific complaints. (as an aside, I wish etpro used a convention that made it easy to tell client cvars from server ones, ala cg_/g_. Yes a few things, like b_antilag affect both, but it is *set* on the client.)

Of course, introducing such a thing at this point would cause some pain.

* Do bug fixes need to be made optional? b_fallingbugfix comes to mind. Also, do we still need b_brokenlogtimestamps? Why would anyone continue to use the old log format?

In general, these things start out optional for testing, and pending league adoption. In some cases, it certainly would make sense to retire them if they are widely accepted.

* Do we really need a chat text censor feature (b_cursefilter) in a competition mod?

This is a fairly silly feature, since it is easy to bypass and subject to false positives, but it makes some people feel warm and fuzzy.

* Are people actually using the trick jump features? If only a small group is, does it need to be included in the main ET Pro mod, adding to its size and complexity? Wouldn't a small, dedicated mod for that specific purpose suffice?

A small mod that would have to backport the hundreds of bug fixes from etpro ? I don't think people would like to have trickjumping be in a mod with slightly different physics. IMO, the extra work of maintaining a second mod would *far* outweigh the minor bloat in etpro.

Regarding not knowing what rules you are playing, you can always use pb_cvarval to check a setting (assuming PB is on, if it isn't, WTF are you playing on that server for.) Also, most server one or another leagues rules pretty closely.

That said, I've long be a proponent of a /modrules command, which tells you what gameplay significant options are in effect. Idealy, these would be in server info, but there isn't enough space to include them.

[Valiz]

The ETPro 3.1.9.pk3's size is 1.75 M, are you really worried about it? My network traffic is ~30 Gb in a month, so files of that size are really nothing to me. I haven't yet seen any completely unnecessary cvars added by ETPro, in fact I have at least tested the effect of all the new cvars added since 2.0.7. ETPro is also used on many public servers (including mine), so the additional cvars are usually needed; the latest and imo greatest change being b_realhead. You can always inform the players on your server about the server's settings with banners or in MOTD.

Feel free to post your views about the growing number of b_ cvars, I see this as a good topic for discussion. Once again thanks to DG for your dissecting analysis on the first post's comments on some cvars.

[Valiz]

Oops I mean gotenks's posting of course, too much vodka for this nice Saturday evening I guess

[Lekdevil.NL]

As a clarification, I'd like to add that the examples I mentioned in my post are just that: examples. Obviously they're not all bad, nor should everything I referred to be removed. They were just added to illustrate my point: quality is not only attained by adding stuff, but sometimes also by cleaning stuff up, culling unneeded features and generally streamlining a piece of software.

No, I'm not worried about the size in bytes of the mod, but about the growing number of variable options that are being introduced. Of course they are introduced for a reason, and there's nothing wrong with that. I question, however, whether:

a) every new feature is really needed,
b) new features that are made optional (by adding yet another cvar switch) should really be made optional, instead of just being enabled by default, and:
c) old or rarely-used features should be kept.

I think a point may have been reached where we're getting diminished returns for every new feature that's being added. So, I'm hoping that people might start to contemplate whether a new feature is worth the complexity it adds to the game and to the server configuration process.

[The Necromancer]

as an aside, I wish etpro used a convention that made it easy to tell client cvars from server ones, ala cg_/g_. Yes a few things, like b_antilag affect both, but it is *set* on the client.

I wish it too , since I am holiday ETPro player ( I play ETPro only a little)
it is really chaotic to me , thanks god for UI menus...
However the menus are not best , they are a bit fuzzy ...

[Blackout]

tjl/tjg are excellent features and I don't think they will be going anywhere in the near future (thankfully). /modrules would rule.

Also, here is a version of the Player Guide updated to match the example.cfg in 3.1.9, save the b_tjg/b_tjl/b_demo features.

[DG]

Broadly speaking I heavily agree with Lekdevil.NL, imo some things just dont need to be options, and infact make scene overall worse by giving options (compared to not giving options). fixedphysics, antiwarp and realhead, for example. ok it's prudent and democratic for a while, but with the feature creep it's getting harder to keep track, i think for these 3 at least its fair to say "fucking excellent" and time to fix them in place - at least i think that'd be the consensus euro view?

on the other hand i dont see the problem with anything like b_noactivatelean, it's an option for players to be able to stop activate+moveleft=leanleft. Anyone that needs this can activate it, everyone else can forget about it. Ditto for the trickjump stuff.

b_cursefilter, OK so yeah ETPro is primarily a competition mod. actual use is pubs:comp 5000:1 or something, in terms of hours played. If something is not going to be any kind of problem for comp then lets not treat it like its a problem for comp.

I'd rather have what we have than a "small, dedicated mod for that specific purpose", certainly, but the complexity thing is getting to be a pain, with each major server-optional but key feature there's another divide.

[fretn]

the biggest problem is:

player A wants this
player two doesnt want it

league 3 wants this and that
league IV wants the other this and the other that

but I do agree with you lekdevil.nl there are a lot of options and it can be a pain in the ass to find the right balance

[[MoB]Seany]

choices are good..you wouldnt be complaining if your wern't a config maker. The more control over my game I have the better...imho

[Locked_On]

I odn't find the server config that hard to set. I think all the cvars are pretty well documented. I also don't find it very difficult to figure out what settings are being used on another server. It only takes a minute or two of play to figure it out (no need to even check the cvars). So some servers have different settings, big deal. Find a server with settings you like and play there. And I also don't think using another mod for pub is a good option. ET Pro is better than the n00b mods IMO.

[squadjot]

..hitsounds

[gotenks]

the only ones i have problems with are the ones that forcecvar restrictions they want on you

[Dersaidin]

i allways use them, but again changing game play too much could be negative... and iirc some competitions use b_realhead 0.

gotenks knows the score, in Australia we finished our third ET season with ETPro 3.1.9. Because the gameops didnt want to make any significant changes to gameplay for the remainder of the season they left realhead off. (good choice)

ETPros features are great, having options is great.

Not having options for settings wouldnt stop ETPro from being "Bloated", all the improvements/additions would still be there, It just makes it more configureable.

And as... someone.. said, I believe it would be used in testing them too.