Wow! That was a flash to the past.
Anyway, that link is not very helpful. Those pset ports were done two years ago, which is an eternity, even adjusting for the glacial progress of software development. Let's examine some key differences:
- Postgres 7.0.1 was the current version, and many folks were still using 6.5.3. Current version (in case this thread is read in the future) is 7.2.1. Major changes include OUTER JOIN support, breaking of the row-size limit, non-blocking VACUUMs, full-text indexing, etc; all of these substantially affect the pset text.
- ad_page_contract everywhere.
- acs_objects added, unified across OpenACS.
- User model changed almost beyond recognition.
- APM package management system added.
- templating as core, rather than an add-on.
I find it hard to overstate how radically different the psets will be when you take the above into account. All the above should be added, the kludges to get around their absence removed, and we should probably remove the repeated instances of philg telling us how brilliant he is. Just scrap them and start over.
As for me, I think the psets are a dismal failure in the face of oacs4.5. The psets teach the ACS from bottom-up, rather than top-down, which doesn't encourage new people to learn the ACS APIs, it encourages them to reinvent the wheel. Better to teach rapid development within the oacs framework, and from there drill down into the internals and the construction concepts. That would entail a bit more prototyping and development support code, which I have been dabbling in; but the hopeful result would be a system that helps new users along, provides some assistance to the experienced, and the community is happy b/c new members could come into the community with some knowledge of the ACS APIs/style/customs rather than just Tcl/bash/postgres (& looking to reinvent the wheel).
That's my theory, at least.
