[cad-linux] Re: executable geometry data. attribute definitions

  • From: Janek Kozicki <janek_listy@xxxxx>
  • To: cad-linux@xxxxxxxxxxxxx
  • Date: Mon, 8 Sep 2003 06:25:04 +0200

Eric Wilhelm said:     (by the date of Sun, 7 Sep 2003 23:07:38 -0500)
> >Subject: Re: [cad-linux] executable geometry data. attribute definitions

> I had thought of something similar using code-refs which might be fairly tied 
> to Perl, but the thought that put me off of it is that executables and code 
> references do not express the associative relationship in a human-readable 
> way.

hmm, not always. consider a C program used to generate some complicated 3D
geometry of... a dinosaur. It is lauched with some parameters.
What is better then? 
A source code for this program (that very wisely and in a clear form
generates data) or a sequence of digits that describes the actual dinosaur?

I think that sometimes algorithms are easier to read than any other solution.
The same applies if we wnat to generate some complicated furniture.

if someone changes the source file - diff will show it. And it can be patched
in usual way.

more about this here:
[cad-linux] How many files to use? executables? then also makefile and compile 
them!
//www.freelists.org/archives/cad-linux/09-2003/msg00110.html

also I'll point this again (to have all links in one place)
[cad-linux] Re: executable geometry data. attribute definitions
//www.freelists.org/archives/cad-linux/09-2003/msg00105.html
 
> In revisiting this, I think that it might be best to setup a standard syntax 
> such as the input stream format for a well-established computer-algebra 
> language.  This would allow the parsing code for that CAS package to be 
> hacked in such a way that it could show the relationship definition in a 
> reliable way (which could then be displayed in a relationship 
> browser/editor.)

standard syntax for streaming is a good thing. I think of those executables
as to give exactly this formatted output.
 
> The same could be done for any scripting laguage, but using a CAS would give 
> a 
> lot of power that could be leveraged and you would not want to have to tie-in 
> every scripting language in existence (although the Inline Perl module could 
> provide a fairly slick way of doing this.)

I'm not falimiar with CAS, so unfortunatly I don't know how it would work...

-- 
# Janek Kozicki

Other related posts: