GZDoom needs a new programmer

[Rachael]

The creator of GZDoom, Graf Zahl, has pulled the plug on his project. This does not mean the complete death of GZDoom, as it shouldn't, because it is the only source port thus far based of ZDoom that has any sort of decent OpenGL implementation at all.

During its downtime, there will be maintenance releases to keep it up to date with ZDoom code, but from my standpoint I know nothing about OpenGL, and I will not be able to fix too many bugs or help improve the code any. Gez has also stated that he will do the same.

I do not know where this will go, from here. Graf has completely destroyed his website account before I could do anything about it, so at best I can restore a back-up but it will be up to the community to fix the archive.

He requested that I do the same with the forum, but for obvious reasons, I refused.

So here we are. As I post this, I am in the process of restoring my back-up that I had available for the site. The project itself will probably need to be forked or something to avoid issues, though.

For those interested, the thread where the fall-out occurred took place on Doomworld - here is the link.

(thank you to Metroid01, Qwerqsar, and Schneelocke for helping to restore some of the more recent files in the archive)

[metroid014]

I have a copy of the r768 source, if you need it.

[Gez]

I have a copy of the r769 source.

[jbailey8]

Man.... I can't believe this actually happened, I've been involved with
development projects myself and I know what it feels like to have a few
people bad-mouth the work being put into a given project, but I learned
to generally ignore those kinds of people and move on with the work at
hand. Hopefully, everyone involved will cool off over time and things will get
back on track... I have a svn copy of 769 if you need it by the way.

John

[dark-slayer-201]

The only other person I can think of working on this is Torr, but he is already tied up with skulltag atm.

[Dethernal]

Graph Zahl code was pretty dirty. Engine was ineffective. If and to continue to work, the current version of the engine should be inhibited for vanilla, boom and older Zdoom maps. New Engine to be done again, with the default should work the old code - right under the vanilla tricks. New engine should be free from the burden of problems and compatibility, to use modern approaches to rendering (drawing adapted to the geometry of the final optimized buffers to be generated with the BSP, which take into account the possibility of video on the caching vertexes). And changing geometry (doors, floors and other) may be collect in one global buffer that need change.
Unfortunately, I am not an expert on graphics and even more so I can not entertaining programming - I have no free time for this.

[Cutmanmike]

Well at least the port is still usable. I gotta say I've really gotten used to the nicer visuals and full freelook GZDoom gave us (as well as all the other features), and it would be a shame to lose them all over a bit of a squabble.

Now to be fair, I kind of feel for Graf in a sense that everyone criticised the fuck out of him on forums. But unfortunately that's what happens, and you need to either ignore it or come up with a solution to everyone's problems which usually impossible. People would just rant about GZDoom or just generally insult Graf for making it not work for so and so, but Graf would always step in and argue and defend his creation usually while insulting others in the process. I know the feeling when so many people are bringing one of your creations down (especially so when it's uncalled for). You want to go in there and stand up for what you've created, you want to just dash in there and say "Yeah well FUCK YOU" but you just can't do that. Not only does it just ruin your reputation as a developer but nothing comes out of it. You just have to suck it up man, and just enjoy maintaining and creating your work, otherwise you'll just ruin it for the others who have no problem with what you're doing.

[Rachael]

Cutman, being the guy who lead the development of Ghouls vs Humans, I know you know exactly what you're talking about because I've seen GvH come under much criticism, much of which I found unwarranted. I think that people are just upset because it takes away from "the real Skulltag" but the truth is you can only be excited about one single thing for so long (and eventually, you get bored of it, I think that's what happened to old-skool Skulltag). So they blame GvH instead. (And AOW of course, but that's another story)

Anyway, I do agree with what you have said, and my criticism towards Graf was mostly about how he handled people. Yes, his code was inefficient, but what can you expect from someone who did it as a hobby? Mine never would have been better.

That being said, I respect you a lot for coming out and saying that. It did have to be said.

[Gez]

Graph Zahl code was pretty dirty. Engine was ineffective. If and to continue to work, the current version of the engine should be inhibited for vanilla, boom and older Zdoom maps. New Engine to be done again, with the default should work the old code - right under the vanilla tricks. New engine should be free from the burden of problems and compatibility, to use modern approaches to rendering (drawing adapted to the geometry of the final optimized buffers to be generated with the BSP, which take into account the possibility of video on the caching vertexes). And changing geometry (doors, floors and other) may be collect in one global buffer that need change.

You're talking about collecting optimized buffers and video on the caching vertexes, so obviously that means that you actually are very knowledgeable and have passed a lot of time auditing the GZDoom code and comparing its performances with an experimental replacement renderer you are working on. I am very excited to see your work soon!

Unfortunately, I am not an expert on graphics and even more so I can not entertaining programming - I have no free time for this.

Oh wait, then you say you don't know what you're talking about and you don't even know programming and you have no time to do anything anyway. So, what, does it mean that everything you said just above was complete bullshit or just mindless parroting of things you've heard?

The rendering system Graf uses is pretty similar in its concept to the one used in GLBoom+. Which is currently the only OpenGL renderer available that's faster than GZDoom. On the Skulltag development team, the only guy who actually has worked with the renderer likes GZDoom and is the one who is behind the switch from ZDoomGL to GZDoom. So allow me to dismiss as bullshit everything you said.

[Cutmanmike]

That being said, I respect you a lot for coming out and saying that. It did have to be said.

Aw shucks

Has Torr heard the news yet? I'd like to know what he thinks of all this. I'm guessing he doesn't want to take over GZdoom, but I was thinking... Since Skulltag already uses GZDoom code, couldn't Skulltag become the new spiritual successor GZDoom? I believe ST already has some new model features that Torr had to create, that may not even be in GZDoom (don't quote me on that though).

Edit: Although this is a long way off I know, as ST has a lot of catching up to do in terms of ZDoom and GZDoom features. But eventually maybe?

Oh wait, then you say you don't know what you're talking about and you don't even know programming and you have no time to do anything anyway. So, what, does it mean that everything you said just above was complete bullshit or just mindless parroting of things you've heard?

Ohh. Not to change the topic but I think everyone who has had a stab at Graf blame his bad/dirty/stupid code, but has anyone of them even touched the source? I bet if they saw what complications were involved they'd keep their trap shut, but eh... That's life

[MartinHowe]

You just have to suck it up man, and just enjoy maintaining and creating your work, otherwise you'll just ruin it for the others who have no problem with what you're doing.

Unfortunately, that isn't obvious and needs to be taught and this is especially true of brilliant clever people who find it hard to cope when criticism isn't soberly stated and when fair criticism and unfair criticism are included in the same sentence and thus difficult to separate. Just look at Yuvraj Singh in the IPL, Sreesanth in international cricket, or David Beckham at Manchester United.

People often fail to give enough respect to truly outstanding work, akin to criticising the Space Shuttle because it can't take off by itself instead of praising it for doing what other aircraft simply can't. Yes GZDoom has problems that will need to be addressed, and I have had my disagreements with Graf in the past, but right now if you want 3D floors and total ZDoom compatibility, it's the only game in town; don't forget that much of DECORATE's exposure of internal engine functionality was by Graf alone and without it, many of today's projects would be well and truly fucked (including my own). Note the deliberate use of that word in plain text and not asterisked; something I rarely do.

At least in the working world, a good line manager or director can show leadership and guidance and help such people to learn to cope; in a public forum there's just a shitstorm and, if one is lucky, the moderators might clean up the worst of the shit. It should be noted that writing unambiguously constructive criticism isn't as easy as one might think.

[Cutmanmike]

Yes GZDoom has problems that will need to be addressed, and I have had my disagreements with Graf in the past, but right now if you want 3D floors and total ZDoom compatibility, it's the only game in town; don't forget that much of DECORATE's exposure of internal engine functionality was by Graf alone and without it, many of today's projects would be well and truly fucked (including my own). Note the deliberate use of that word in plain text and not asterisked; something I rarely do.

Here here! Honestly if it wasn't for DECORATE's expansion I doubt most projects today would even exist, most players would have ditched Doom altogether (probably myself included). Despite his attitude sometimes, we owe Graf a lot and I did try to stick up for him when he's getting shat on over at the ZDoom forums.

[Dethernal]

Gez
Yes, I really see and debug visual code of GZDoom.
In general, I think that there is calculated too much for passage. In modern implementations of graphics systems, on the contrary, on the video card is shifted as much as possible and the processor operates only on logic. Although I do not really see what there is for the "Buffers" Graph Zahl added. In any case, In a current version you can create a map with the 10-20 thousand polygons, which will not change, but the low FPS to 10-20. It's Normal?

[Gez]

It's Normal?

I don't know. I will be completely and brutally honest: I have absolutely no experience in OpenGL or 3D rendering; indeed I barely have any experience in 2D rendering (and even then, it was in QBASIC and when I was less than half my age).

So I am completely clueless about OpenGL, what is good or bad practice in OpenGL and what is faster than what. Okay so far? So I have to base my opinion on what people who are actually experts, or as close to experts as we can find in our niche anyway, say. Let's see the tally:

- Torr Samaho, without whom Skulltag wouldn't even exist anymore: pro-GZDoom
- entryway, coder of the fastest OpenGL Doom renderer available: pro-GZDoom
- Other developers (andrewj from EDGE, DaniJ from Doomsday, etc.): no comments about the renderer that I could find.

Yeah, see where I'm going? VC is the only one who criticized the renderer while claiming to have some actual expertise, but I haven't seen his work so all I have to weight his opinion is his own words...

Also on benchmarks GZDoom was consistently the second-fastest OpenGL Doom renderer, behind GLBoom+. The partial rewrite that GZDoom underwent was also entirely driven by benchmarks, with new code being dropped or kept only on its merit as far as performance improvement went.

So, again, see where I'm going with it? Given no previous bias or preconceptions, I can weight on the one hand the opinion of the leading experts and the cold hard facts of benchmarks, and on the other hand the opinions of people with an axe to grind and one self-styled expert who works on an obsolete version. I'm judging on what I know and what I can test, and my verdict is that GZDoom is actually rather good. Anybody who says it's inefficient crap is perfectly free to try to sway my opinion, all they have to do is release an alternative renderer that is actually faster yet performs the same features (including software hack emulation and all current G/ZDoom features such as cameras, portals, mirrors, etc.). Just do that, which should be easy since apparently everyone knows how it should be done, and come back to me.

[entryway]

Gez
Yes, I really see and debug visual code of GZDoom.
In general, I think that there is calculated too much for passage. In modern implementations of graphics systems, on the contrary, on the video card is shifted as much as possible and the processor operates only on logic. Although I do not really see what there is for the "Buffers" Graph Zahl added. In any case, In a current version you can create a map with the 10-20 thousand polygons, which will not change, but the low FPS to 10-20. It's Normal?

Trying to draw walls/flats with the same texture in one batch just does not help for Doom IMO. Even more, it is slower. At least on my NVidia. Even for such scenes: http://prboom-plus.sourceforge.net/hg_map04.png. It is slower (I think), because you need to fill not only buffer with vertexes indexes, but color buffer also and you should use INT indexes instead of SHORT which is noticeably slower. I am sure it helps for iPhone, but not for my GeForce 8800. Immediate mode is coded very fine at least by NVidia guys and you need 1k+ or better 10k+ primitives per batch to have profit, not 10-20.