On Fri, 15 Aug 2003, Wolfgang Pichler wrote:
found a visualized RD-scheme at http://irref.mine.nu/user/dchud/koha-scheme/ as a starter, but it seems somehow useless without a list of values used for indicator/status-fields or a list of constraints checked obviously programmatically in various places, or more bad, never checked and evocating hard-to-find delayed bugs due to perl's smatrness of operating everything somehow.
Can somebody put that URL somewhere that it's actually accessible? And I seriously doubt that a quest to have Koha work with full referential integrity checked by the databse wouldn't be a worthwhile endeavor unless we're seeing data corruption heisenbugs that need to be tracked down. And honestly those cases have always been handled by looking at the query log and seeing what it's doing much more effectively than "the script crashed because some dippy purist thought referential integrity was as important as some college professor told them and that college professor hasn't written a program that's actually useful to a real human doing real work in thirty years." Feel free to disagree, but if you don't have a signficant amount of nonacademic experience with doing development projects semi-informally then you have no relevant context to what I've seen happening with Koha. That doesn't make Koha bad. That doesn't make you bad. It just means you're going to have to get used to things as they've been done or see if anybody will let you fix them your way.
as stated before a definitive list of calculated (when,how,... nasty questions, i admit,... :-) vs. untouched "bare" legacy data would be great. i ever some "design" was applied, this could be no problem, but koha seems to be some grown beast with different code-quality and maybe no one remembers the old assumptions any more :-)
If you'd like to write that sort of documentation I'm sure someone will provide some web space for it. Honestly, "design after" isn't a bad thing for people that have some clue about implementing practical databses. And since you can obviously browse cvs and read perl you can find all of those code-quality misassumptions and submit patches. Right?
so i do not expect some kind of petri-net, but if koha is a multi-developer-effort, at least some negotiated interfaces could be documented in some descriptive emails ...
Given how few people are actually coding and they seem to be making real progress, how much negotiation do you expect is still required?
Did someone post a link to a cross-referenced source of koha, or did I imagine that?
SO WHERE IS AT LEAST THE LATEST CVS ? sf ? looked at dates : seems, everyone keeps it's own copy and check-ins are rather rare.
That's a pretty normal developer pattern. Push things along in personal cvs until it's worth worrying about merging. It is generally better to put things out as often as possible and people may be pushing cvs commits after every day they've worked - we just don't know. -- </chris> The death of democracy is not likely to be an assassination from ambush. It will be a slow extinction from apathy, indifference, and undernourishment. -Robert Maynard Hutchins, educator (1899-1977)