| |
In lugnet.cad.dat.parts, Larry Pieniazek wrote:
> One good thing about a reference implementation (which is controlled by the
> standards body, which Ldraw 0.27 is NOT since it's not a buildable thing) is
> that it can shift arguments about standards a bit.
>
> That said I'm not arguing FOR, just saying "good question". I think I'd rather
> try to get by without a reference implementation.
Having an official reference implemention would put us in a better position in
several ways. For the parts library, we could resolve a number of nagging
issues, like the dithered-colors-have-to-be-subfiled restriction. For the
rendering language, we'd be able to define new statements, and to more strongly
include our standards in the LDraw definition. People would have a workbench
from which to try new directions for the language (for example, would commands
for curved objects be a good thing?). People wanting to find out more about
graphics programming would have a simple code base to learn from.
Can we "get along" without a reference implementation? Sure. We've been doing
it all along. LDRAW.EXE isn't really an RI. It's been taken as a standard, but
it's restricting us as much as helping us. We don't control it, can't change
it.
My ideal reference implementation would meet the following criteria:
1. Self-contained.
2. Published source code.
3. Portable.
4. Readable.
What it wouldn't (necessarily) be:
1. Full-featured
2. Fast
Steve
|
|
Message is in Reply To:
16 Messages in This Thread:
  
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
