So,... what's the situation on the Dark engine source code? - by Cliftor
MoroseTroll on 17/4/2009 at 05:29
Volca: Thanks :). What about a quality of English in my petition?
Volca on 17/4/2009 at 05:56
Quote Posted by MaxDZ8
No... you could pretend to say so, but design choices made 10 years ago will be heavyweigts to the updating process... eventually, restarting from scratch
could possibly be easier.
I am not sure you understood how advanced the Q3 engine was for the time. This is not a good example. A better example would be taking less-than-Quake1 to Doom3. And not to forget the engine isn't a magic box: you'll have somebody to redo the artwork and even if somebody takes up the task, odds are it will just look like rusted shit.
You're taking solely about graphics. Let's say some arguments to keep it from being too generic. Q3 is more hardware optimized, that is true (for the hardware of that time).
Comparing those:
Q3:
* has PVS
* uses BSP
* Uses lightmaps
* Lightmaps are readily stored in atlas
* Geometry is exterior based
Dark's Portal renderer:
* computes dynamic PVS
* uses BSP
* Uses lightmaps
* Does not use atlas
* Geometry is interior based
PVS does not matter these days. The geometry chunks in Q3 are too small to be rendered effectively anyway. I'm not sure how people get around this but I'd say it's not easy. Puting lightmaps into atlas is a problem of a few seconds maximally while loading the level.
You see the situation seems to be more or less the same. We could have offline processing tool that would group the geometry cells in WR so the rendering would be more effective, for example. That would mean we could have a pretty modern rewrite of the rendering part. Some of the things in Dark are bound to the way renderer is written for sure. I say it's not a problem if you have the access to the source code.
Quote Posted by MaxDZ8
I've also seen an old game I liked alot getting literally raped by the "recreators": higher CPU usage (even higher than running the original in dosbox!), decreased compatibility and a ton, a ton of useless dependancies...
You can always fork and show them how to do it right, right?
Quote Posted by MaxDZ8
It looks realistic to me that, as already said, the code has probably been destroyed years ago, or it is in some back-back-backup so deep in the basement it is effectively lost.
Pretending somebody invests man-hours in the resurrecting is a pipe dream.
I agree here, sadly. But, I still think it's worth trying.
Quote Posted by MoroseTroll
Volca: Thanks :). What about a quality of English in my petition?
I'm not the right guy to judge this...
MaxDZ8 on 17/4/2009 at 09:37
Quote Posted by Volca
PVS does not matter these days. The geometry chunks in Q3 are too small to be rendered effectively anyway. I'm not sure how people get around this but I'd say it's not easy. Puting lightmaps into atlas is a problem of a few seconds maximally while loading the level.
But then add the fact that the Q3 shading language is essentially the same as the Doom3 one. The Doom3 language has a few extra keywords and admittedly better management of 'waves'... but it has been built incrementally and maintained over the years by multiple licenses, it used relatively stable internal formats, APIs, conventions. Some Doom3 tech was already internally used by ID when Q3 was released (I'm referring in particular to combiners NV_register_combiners and dependant lookups NV_texture_shader), there was an entry on JC's blog (at the time, .plan) back when he had one, Q3 already had some admittedly buggy shadow volume support.
The jump is really shorter than most project announcements want us to believe.The original DE source in the meanwhile just fossilized.
I think ODE team is doing the right thing in restarting from scratch. After all, everything that matters is to load the assets correctly and manage them sort of coherently. I have alot of hope in their efforts.
Rewriting is something, but considering that was essentially the middle ages of 3D graphics I think at least refactoring, and possibly redesigning to be more probable situations. If the original design still makes sense, rewriting will be easy and give advantages. If not, it's going to bite where it hurts.
It is always possible to workaround the design with more efforts but starting from suboptimal state isn't something getting me excited.
What you say is exactly that: yes, we could reprocess everything. Then we don't have a modern rewrite. We have a significant content pipeline workflow modernly redesigned (which is surely a good thing but a different beast).
Quote Posted by Volca
You can always fork and show them how to do it right, right?
Yes, I already have a prototype (but I didn't fork it to avoid the GPL, I redone everything from scratch). Considering my spare time, I suppose it could be done in a couple of years... :erg:
Do you think it would be really useful to have the original? I assumed ODE had completely reverse-engineered everything, but I may be wrong. Surely there would be an improvement when it comes to file layouts and entity behaviours... If you put this in those terms then I agree 100%.
Dront on 17/4/2009 at 18:35
Found a simple mistake.
Quote:
The games Thief: The Dark Project and Thief 2: The Metal Age, which
was released in 1998 and 2000,
Of course,
were released is correct.
However, is there an English-speaking volunteer with a bit of will to help?
Thief13x on 17/4/2009 at 19:33
I hope you don't mind if I step in as the a-hole:). I'm generally pretty picky about grammar/ writing so please take my corrections with a grain of salt and feel free to ditch them because I do think you wrote that well enough. These, however, are the changes I would make...
Quote:
Dear Mr. Yoichi Wada!
We are the Thief-community and are inspired by the fact that your company, Square Enix, has bought Eidos. Your corporate slogan, «It is the mission of the SQUARE ENIX Group to spread happiness across the globe by providing unforgettable experiences», has brought hope to us, who are pained for ten years by the futile, but at the same time ineradicable, dream. Let us share it with You.
The games Thief: The Dark Project and Thief 2: The Metal Age, which were released in 1998 and 2000, opened for PC players a door to the mysterious and mythical world based on Medieval Europe. The world that is tissued from steam and magic, made by the geniuses of Looking Glass Studios, became so attractive to Thief players that they began to create their own missions by themselves, and these missions enlarge and fill up the original games (now there is more than 800 of them). Not few among them are better in quality than the original games (see, for example, rating at (http://www.keepofmetalandgold.com/fmarchive.htm)).
The world of Thief still lives over a decade after its initial release through the efforts of its fans - isn't it proof of the fact that the game is a work of genius? However, time passes by, and the technology, made for the base of Thief: The Dark Project and Thief 2: The Metal Age, can no longer provide enough freedom for the activities of those who want to enlarge and improve the original games. We, as the Thief-community, dream of receiving in our hands the original source code of Thief: The Dark Project and/or Thief 2: The Metal Age in order to make changes and improvements to it. By releasing the source code to us fans, it would give more scope to those who continue to create the world of Thief.
We are by no means calling on Square Enix to release the games Thief: The Dark Project and Thief 2: The Metal Age as freeware. We are only requesting the original source code for the graphics engine to these games, not anything more. More than ten years has passed since the release of Thief: The Dark Project and more than nine years since the release of Thief 2: The Metal Age. Because of this, we realize that these two titles are likely to be of little technological and/or commercial value for publisher. That's why we charitably ask you, the new owner of the "Thief" brand, to/B] make the Thief-community happy and publish the original graphics source code of the games Thief: The Dark Project and/or Thief 2: The Metal Age, thus confirming Your corporative slogan.
Publishing the original source code of an old game is a widespread demonstration of good will to players - it is done by known of PC game developers: id Software (Wolfenstein 3D, Doom 1&2, Quake 1-3), Raven Software (Heretic 1&2, Hexen 1&2), 3D Realms (Duke Nukem 3D, Shadow Warrior), Volition (Freespace) and others. We cherish hopes that Square Enix will do the same with Thief: The Dark Project and/or Thief 2: The Metal Age as well.
We understand that it has been nine years since the closing of Looking Glass Studios, and that it is possible that the original source code of Thief: The Dark Project and Thief 2: The Metal Age does not exist integrally and ready for compilation. If this is true, we understand this but then, please, publish everything that is left, and we will try, like archeologists, to rise the great past from the ashes! I think I can speak for the reast of the community by saying thanks for writing this. I hope I could be of some help, keep us posted!
MoroseTroll on 18/4/2009 at 07:21
Dront: Sorry! It's my mistake. Thank you for the whole translation :).
Thief13x: Thank you very much :)!
All: Well, so how can we attract Mr. Yoichi Wada's attention? I think that just to write and start the petition in not enough... jtr7's suggestion to invoke journalists seems very good. Maybe somebody from ex-LGS too? Any thoughts?
jtr7 on 18/4/2009 at 07:42
I'm still hoping this is gonna happen:
(
http://www.ttlg.com/forums/showthread.php?p=1827607#post1827607)
An interview with Sean Barrett, the guy who worked on the Renderer and the Dark Engine, and was working on it's next bigger brother, the Siege Engine, sounds like one good source. What we'll need is the person who last saw it AFTER the LGS assets were auctioned and purchased by Eidos.
I think the fans are the ones who can best preserve the code for the original Thief. Sitting in a box somewhere doing nothing is a shame. For preservation and archival reasons alone, we should have it. Add to that the chance that future systems won't be compatible with it, so we need to be able to adapt it ourselves. Add to that the FM legacy, and the need for future generations to see Thief as we see a classic written work, or classic movie, or composition.
MoroseTroll on 18/4/2009 at 08:12
jtr7: I'm totally agree with you! Perhaps, he (Sean Barrett) or his ex-colleague Kevin Wasserman (he did the Direct3D hardware support in Thief 2) is just waiting for the legal permission from the copyright owner - in this case it would not make sense does Eidos have the sources or does not. I wish I could know how exactly the situation is!
eddy on 18/4/2009 at 08:52
Did anyone actually ever emailed Sean Barrett about the whereabouts of the code?
jtr7 on 18/4/2009 at 09:08
All known attempts to contact devs or Eidos came up empty, or hopeless, or too vague and business-like. If Brethren gets his interview, I hope we can get new information, even if it's still unhelpful to our cause.
