Project 'brainstorming'
Moderator: General Discussion Moderators
Actually the idea being my question was that maybe OA could be used as a base for the game code, instead of vanilla ioq3 source, to avoid requiring Q3 data and to have a few extra enhancements like unlagged right from the start.Willis wrote:If TFn were to switch engines, it would be based on top of an engine, not a game/mod. It would be based on ioq3, which is the engine that Open Arena uses, or xreal, a fork of ioq3. I understand all mention of OA so far has likely been in terms of testing what the Q3 engine could do with the current maps, I just want to make that point clear.
Now that's very interesting. To be honest, I have never thought that the QC game code could have a visible impact on the online performance. In your opinion, what part of our game code is responsible for that? The animation code?Willis wrote:DPs netcode is nowhere near as bad TFn makes it appear. DP certainly ins't perfect, but the difference between Nexuiz and TFn is night and day because Nexuiz has a better and more mature code base on the QC front.
I disagree here: Python is a good scripting language. From a performance standpoint though, I'm not sure how it would stand against QC. But if it's just a proof-of-concept, I guess it's not really interesting for us anyway.Willis wrote:No, there are no other game code programming methods available for DP aside from a Proof-of-concept python engine that is dangerous, and when it comes down to it, python isn't that much better of a language.
Doh... I forgot those models. Don't we use MD5ANIM files too? Are those 2 formats the only Q3-incompatible ones we use, or I am still missing something?Willis wrote:The most notable of such is on the model front. PSK is a very advanced format, the best available in ioq3 to the best of my knowledge is currently md3, and mixed support for an unpopular skeletal format. This is one of the biggest reasons we stayed with DP.
- Willis
Master of the Mask
Lead Programmer- Posts: 872
- Joined: Tue Aug 10, 2004 09:28 am
- Location: Eau Claire, WI USA
- Contact:
Being that TFn from a DP standpoint doesn't use any Quake1 data, it shouldn't require any Q3 data with a Q3 engine either, so making it a mod as opposed to a standalone would make it unnecessarily more limiting.Elric wrote: Actually the idea being my question was that maybe OA could be used as a base for the game code, instead of vanilla ioq3 source, to avoid requiring Q3 data and to have a few extra enhancements like unlagged right from the start.
I've never had much luck at tracking it down, but at idle, Transfusion sends approximately 129 bytes of game data per server tick. In comparison, Nexuiz is somewhere in the neighborhood of 19 bytes per tick. Granted that is just idle, once movement is configured in, that number needs to go up drastically, but I think you get the picture.Elric wrote:Now that's very interesting. To be honest, I have never thought that the QC game code could have a visible impact on the online performance. In your opinion, what part of our game code is responsible for that? The animation code?
I personally have never cared for Python, and yes, it's performance is horrid. The PoC that exists does work, but the reason it isn't implemented or used anywhere is because of the security risks associated with it.Elric wrote:I disagree here: Python is a good scripting language. From a performance standpoint though, I'm not sure how it would stand against QC. But if it's just a proof-of-concept, I guess it's not really interesting for us anyway.
No, Kazashi was originally requesting MD5, and we ended up with PSK, I'm guessing MD5 was more difficult to impliment into the DP renderer. I'd suspect with some help from Kazashi, all PSK models could likely be converted to MD5/MD5Anim, but am not certain.Elric wrote:Doh... I forgot those models. Don't we use MD5ANIM files too? Are those 2 formats the only Q3-incompatible ones we use, or I am still missing something?
In looking at GPL Q3 engines, XReal seems like the one that has all the features out of the box that would make sense to attempt a change with. Granted, that would still require that model format conversion. My first question however would be performance, as XReal changes many things in the rendering quite drastically, that I wonder how it performs on lower end cards.
http://xreal.sourceforge.net/xrealwiki/Features has a list of features included in the XReaL engine.
- Willis
Master of the Mask
Lead Programmer- Posts: 872
- Joined: Tue Aug 10, 2004 09:28 am
- Location: Eau Claire, WI USA
- Contact:
I wouldn't go so far as to say that XReaL has any more Blood-like features than any other Q3 engine. However, it does come with features that would ease a transition more than any other Q3 engine I've seen so far. In a default compile, it has Ogg/Vorbis enabled, MD5 model format, and has other luxury features similar to DP such as Real-time Lighting. It also has TrueType font support, PNG image support (as well as continued TGA support), tons of other lighting and shadow effects, etc...
While IOQuake3 has code and support for Ogg/Vorbis, it isn't enabled by default, and the model options are much more lacking.
XReaL is also a full game, so like OpenArena, could provide a base code to fork from, and their code certainly includes Unlagged as well.
While IOQuake3 has code and support for Ogg/Vorbis, it isn't enabled by default, and the model options are much more lacking.
XReaL is also a full game, so like OpenArena, could provide a base code to fork from, and their code certainly includes Unlagged as well.
Please read the FAQ: http://www.transfusion-game.com/faq.php
^^DustyStyx^^ I understand the value of open source, but considering the recent bottle neck in team members and progress, new options must be explored. Sure, coder mountain might go to the Transfusion Mohammed. It is more likely however that the Transfusion Mohammed will go to coder mountain if you catch my drift.
^^Slink^^ Kel knew what was going to happen. He new all along.
^^Slink^^ Kel knew what was going to happen. He new all along.
Now Slink, progress in DP does not equate to progress in transfusion.
http://en.wikipedia.org/wiki/Correlatio ... _causation
http://en.wikipedia.org/wiki/Correlatio ... _causation