Messages

Right now weapon convergence is always on, and I could see why you'd want the option to turn it off in that particular case.

That was my position, but I figured it's better in this case to be overly-strict than overly-permissive.

I really do wish I had learned how DR2 organized their data files before BZ2 shipped.  They used file system streams in clever ways, and generally handled mission-specific assets and game mods better than BZ2.  Fundamentally reorganizing BZ2's asset structure would be a very bad idea at this point.  Sure, I could "just write a Perl script" (inside joke) and break the monolithic data.pak into smaller pieces, but that's something I don't want to inflict on anyone before 1.3 is done (and probably not afterwards, either).

Well, I don't know about fixing absolutely anything.  There can be some problems so fundamental that fixing them involves major changes to critical systems, spawning numerous secondary bugs.  Remember what happened when I fixed the terrain collision system? 

Bear in mind that 1.3 was never meant to take anywhere near this long or affect so much--rather like the Gilligan's Island "three hour tour".  It all started with me fixing a problem with the game crashing when people typed a particular character sequence in chat, along with a few other niggling problems.  We started fixing more and more problems, then I started making large-scale changes, and the whole process ran out of control in truly spectacular fashion.  It's been going on for over five years now.  There aren't too many game patches that take this long...

Nathan is already concerned that I've raised people's hopes unnecessarily.  It's amusing to contemplate, but I'm not going to go tearing into anything that before 1.3 is final, and probably not even afterwards.  There's too much of a "Nathan fixes things, Ken breaks things" pattern for me to start ripping up large-scale systems like the renderer.  Little tweaks are fine, but I need to coordinate with Nathan to bring 1.3 to a close.  He has other things to do with his life than BZ2, and frankly I'm surprised he's stuck with it as long as he has...

The main problem with BZ2 is that it has become fairly brittle.  Changing anything tends to lead to unexpected (and usually bad) consequences.  There are whole systems in BZ2 that need replacing (AI, physics/collision, rendering), but too much depends on them to change them much.  Nathan also makes the legitimate point that open-sourcing BZ2 would lead to a free-for-all of incompatible versions, breaking the already-small community into tiny splinters.  You know there would be a split between the "we hate the 1.3 changes" crowd and everyone else, and there would likely be more.  That's why I respect Nathan's decision to avoid releasing source, even if I don't agree with it 100%.

Before everybody starts assuming I'm going to DX8- or DX9-ify the game, it would be in a hypothetical "version 1.4", if even then.  As I mentioned before, I've never written a D3D application before, so it would pretty much be a disaster if I tried with BZ2. 

If it was ported, would it look and feel the same?

I'd say yes, as CW was effectively a port of the BZ2 code base to a new engine, and then modified extensively. 

If you think things are flaky now, just wait until I break everything by replacing the renderer. 

That's one good reason to stay away from it, no matter how much I'd like to.  Still, hardware T&L. 
(I do have a partial DX8 conversion that was a prototype for what eventually became Star Wars: Clone Wars, but it was not full-featured.)

[read the changelog BEFORE complaining]

Why don't ppl read the changelog BEFORE complaining about something. Just do a keyword search or two to see if what you are about to complain about has been worked on. If it has and it doesn't seem to work like you think it should (according to what has been mentioned in the CL) THEN you can come here and tell us that it's broken.

Because it would be hard? 

The changelog is gigantic; I spent several hours reading what Nathan's been doing in my absence, and there's a lot there.  It's also fairly technical in presentation, so it takes some doing to find things.

How does it compare to the FS2 source?

Or even worse, the Serious Engine 1 mod source... *shudder*

Since I haven't looked at either of those, I can't make that comparison.

So far, here's the games that I've worked on:
Battlezone 1 (based on I76)
Battlezone 2 (remove I76, add DR2)
Star Wars: Clone Wars (remove DR2, add I76-like "Ground Zero" engine)
Star Wars Battlefront 1 (entirely new)
Star Wars Battlefront 2 (extension of SWBF1)

CW has the architecture that I like the best, and I've contemplated "porting" BZ2 to it, as I already halfway did that for the CW prototype.

Wow, its been... a very long time.

Indeed.  I looked back through the changelog, and my last change was a third of the way from the bottom.  Wow.

What about buildings that are not terrain owning... and are spawned by dll... say 2k in the air?

That counts for anything that attaches itself to the terrain cluster ("ownsTerrain"), even if it doesn't hide the terrain itself ("isTerrain").  I'd need to change MapClusterClass::clusterMesh into a linked list (not hard), but making the editor handle it correctly is another issue.

(I wish I could port the Clone Wars or Battlefront editor improvements back to BZ2, but there may be too many game-specific changes in them for that to be practical.)

I suppose converting BZ2 over to a nice clean implementation of DX9 would be good for modders, but I'd have to tear out pretty much the whole rendering system.  The current one is pretty heavily fux0red. 

(Getting rid of software T&L would certainly be nice.)

I definitely got tangled up in a lot of the physics debates during the early development of the 1.3 patch.  I busted the hover physics pretty badly before getting it back to something better than it started, and that cost me a lot of goodwill from the community and wasted several weeks of my time.  The current version seems to be holding up well, and Nathan's been exporting a lot of formerly-constant values and fixing things left, right, and center.  The last 1/3 of the changelog is all Nathan, all the time.