On 3/22/2013 3:19 PM, Bill Meahan wrote:
Hi All,
Maybe, we could setup a collaborative work group to do the documentation.
That is a group of us are responsible for certain groups of commands. This way the manuals can become more complete. That way some of the more advance stuff that is hardly documented finally gets documented.
What we would need is a specification for: [snip] In 45+ years of programming[1] it has never ceased to amaze me how the wheel has to be reinvented for every new system whether language, macro
On 03/22/2013 03:31 AM, Keith J. Schultz wrote: package or whatever. Why do it again? Why not adopt some documentation system that is already widely-used and for which infrastructure and knowledge of use is already in place?
I have no investment in any particular system. I'm happily generating other types of non-computer-related documents using reStructuredText since I can easily convert that various publication formats as required without separate source files for each format. It seems to me docutils has everything that would be needed to document ConTeXt and is very widely used given the popularity of Python (which makes me cringe). If doxygen or something else would work better, so be it. The point is, **use something that exists instead of expending time and effort reinventing the wheel yet again!**
[1] I was, am and will be a "programmer" and not a "software developer" or "software engineer." The term adequately depicts what I did/do while the others are simply too pretentious. Find the old article "Real Programmers Don't Use Pascal" in an archive somewhere -- I've been a "real programmer" and I suspect Hans is, too. :)
Sorry for the rants but it is so frustrating to have to install so many different language support and documentation systems simply because I use FOSS tools exclusively.
I can only speak for myself, but - I did my share of programming (pascal, modula 2) when I university but at that time documentation was mostly in-source. My background is educational technology and not programming but I always ended up doing that. (I still have a stack of old listings somewhere of a formatter that I wrote for vms that took some kind of tagged ascii and paginated that etc.) - Later on when I ended up in educational consultancy and development of all kind of educational stuff, context was developed simply because we needed a flexible typesetting tool. We also developed tools and workflows around it. Ha, there was no internet, at least not for us, so we didn't even know what else was around. - So, whenever I have to write some documentation, I use context itself, after all, one needs to typeset examples. I normally pay a lot of attention to the document source code and can live with some tagging. If I had to do that in some * ** == -- & based ascii text format or whatever, I'd probably never write manuals (too much hassle to go beyond the obvious and not looking nice either, but that's personal). - When I started with the command specification in xml, it was also because xml is easy to process, and (in mkiv) we can also easily filter based on expresssions. So, for that xml is quite natural for me. Just as nowadays lua is my natural choice and most of my current docs are a mix of mp, lua and tex, also depending on what looks nicest in document source. - I happily leave additional documentation to others and whoever does that should should the tools he/she likes most. In these days one can always convert. - But, to come back to your last comment: tex can typeset its own documentation so that's a rather natural choice for part of it. Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------