Eruanna wrote:Which is part of why I put disclaimer notices to that effect very prominently in the announcements as well as the actual download page for GZZScript.
I've noticed the effect of that has made some people afraid to even try it out, even stating that they are afraid converting their mod would be a waste of time because the language would change anyway (even though I did suggest resulting progressive changes would be smaller in comparison to the conversion, itself).
At this point, I think the best thing for ZScript is if you guided the testing. Like made announcements to the effect of "this group of functions" or "this type of syntax" needs tested. Or it may be too early for that right now, but I think something like that does help, if you are able to.
It sucks that people are afraid to try, really. It'd be for the sake of forwarding progress.
But I can think of a few legitimate reasons why:
- If something is broken after a straight up convert, it begs the question: Modder's fault, or zdoom's?
- Mod size is definitely a factor, such as AEons of Death. I sure as hell won't convert that until ZScript is merged, save for a few things that could desperately use the overhaul as-is.
- Internals are ever changing and chaotic at the moment, and syntax is changing.
- Lack of proper documentation makes it painful to even start from scratch, it seems.
- Some features aren't yet available..
- Probably some others that I haven't thought of yet.
Though on that note, yes, converting D4D was a pain in the ass. Took about 3-5 days? I think? That was strictly porting it over, not starting the use of the new functionality. Now I am using whatever I can to eliminate many a bad hack, and now I can't get enough of ZScript. I never want to go back to DECORATE again if I can avoid it.
Some of the planned ideas have to wait until local arrays are available though.
Graf Zahl wrote:I cannot guide the testing and do the programming at the same time. Ideally, Randi should be up here, guiding development, but I think that's a futile hope.
One thing I learned on my job is that it will never work if the programmer of some complex thing also needs to act as producer. A producer needs to have an unbiased outside view on actual development or they'll never catch if something goes wrong.
I have done both - programming and producing, but the projects where I had to do both at the same time were always the ones where something went bad, for the sole reason that nobody was able to give proper technical feedback.
To be honest, what worries me most about the compiler is that nobody with a good grasp on how to stress test such a thing and deliberately look for logical errors is present here.
Sounds like randi just gave you the reigns and said 'have fun'.
Also this brings up an old curiosity -- What is the point of the VM? I know it plays a big part in all this, and that it is likely to be the only thing which allows zscript to exist, but still...