Re: [RE: emc really needs a copyleft cad/cam package]

[Ray <rehenry-at-up.net>]

On Sat, 18 Nov 2000, Anne wrote:
<S>
> I'm saying that if we start on another application, a CAD/CAM package, we 
> cannot assume that we don't have to do any design.
<s>
> Your comments are a classic example of why designing first works.
> You want it to produce RS274. I want it to produce the peculiar variant of HPGL
> that my Roland Modela takes. Probably there are controllers out there that take
> various variations on each. This is a clear place for a plug-in architecture.
> The Roland uses windoze based software and looks like a printer. emc runs on rtlinux
> and looks like a filter.

You're exactly right -- this CAD/CAM app must have a life of it's own.  I
may have baited you on that one but I didn't mean to.  And I firmly
believe that before we are done with a good version 1.0 we'll need to do a
lot of design, a lot of expected outcomes, and a lot of documentation of
what this part does and how it hooks to that part.

I have been watching the development of the LinuxPLC (puffin) since
Curt's first post to the automation list.  There are a couple of lessons
there.  As you mention design needs to go first, although code can often
serve as the operational definition of specific features.  Don't close up
the design to quickly or good features get left and evolution gets boxed
in.  

Give the brainstorming time and patience.   On puffin there was a big todo
over the insistence by some on ladder as the "only" programming language
for the operation of the thing while the core folk were already writing
code in c++ and didn't see why classes and such wouldn't be more a
more powerful and therefore better op language.  This turned a lot of folk
into lurkers.

And last KYFSIP - keep your flame suit in place.  We are all explorers.

<Rant Off>

Wow!  Thanks!  It sounds like we may be gaining momentum here.

Ray