Thanks for chiming in.

I’ll note that the install wasn’t all that offensive to me–I installed it several times (laptop, desktop, and two instances on the production server) and it was never all that difficult. But I’m more UNIX nerd than web developer, and the majority of my job for the past 8 years has been packaging software. So that’s not proof of an easy install. As I mentioned in my comment to Dossy, two of the other CMS I’m considering switching to literally install in a couple of minutes (or 30 seconds, if you have Virtualmin Professional running…which, by golly!…I happen to have).

I do strongly suggest, as a parting word to the OACS community, that you lose all mention of Daemontools and the djb stuff from the docs. I ran OACS for a year and a half without them, and I certainly did not feel any pain from it–all of my problems were wholly unrelated to how I fired up the servers, served and received mail, or answered name service requests. It’s senseless complexity in an already complex process. (I’ve said it before, though, in the OACS forums…so I won’t go on more about that.)

I do believe the right things are being said and done in the OACS community, and the people are exceedingly smart. I miss the general aura of smart found in the OACS forums when reading the Joomla forums; some of those folks just aint firing on all cylinders, or perhaps they’ve never used a computer before that very moment when they posted. Those are the only two possible explanations I can think of for some of the questions and answers I see there.

I suppose the problems with OACS are just a matter of the fallout from the AD buyout and subsequent closure, Open Sourcing and porting to PostgreSQL, the coming and going of a lot of core developers, along with their individual vision of how things should be done, over a span of several years, etc. And, the fact that everyone has to first scratch their own itch, and that scratching has been focused on .LRN for a lot of the most serious developers. I don’t begrudge anyone scratching their own itch. After all, I’m leaving to go scratch my various itchy places, and I don’t feel a bit guilty about it. I just need to be working with a CMS where the needs of a sizable number of its developers and users more closely matches mine. I need ecommerce for software licenses, bug tracker, forums, in that order. Small set of needs, but they all have to work really well and be extendable with the skills I have handy (I don’t know any PHP but if the PHP developers I’ve know are any indicator, it aint exactly rocket surgery).

I sympathize with the conundrum you and the OACS folks are in, as I’ve worked on several large Open Source projects over the years (Squid, Webmin, SciPy, etc.). But we’re wrapping up our first year of software sales, and so I’ve got to figure out how to fix our software registration system to support renewals, provide nice license usage reports to users, handle expiries, as well as get several major unrelated issues with our current site resolved (sub-sites never worked right with some of the apps we’re using, for example). Every time I looked at doing it myself, or did the math on what it cost to have the initial simple software registration app produced and how much more work is needed to get the functionality our customers have been demanding, I just felt like I was digging myself further into a hole by sticking it out with OACS. And, of course, the bug tracker issue has been a thorn in my side from the first week. One of our very first users started the trend, and we’ve had regular complaints ever since. Seems small, but it seems bigger every time I look at it.

Not that I’m complaining. OACS has served me well (with a few caveats), the community has been grand, and I enjoyed tinkering with it, when things were going well. It’s sold a few hundred copies of our software for us, and helped us make our products better and our customers happier, and for that I’m grateful. If it could get us to the next level without cleaning out our bank account, I’d stick with it.

I can’t argue with you, you are right (except openacs.org is using OpenACS 5.2.3, including the workflow bugtracker, but that’s not really a big deal.)

The one thing you should notice is that OpenACS 5.3 is being released after we pass our automated tests, and there has been a big effort to improve quality. We have a long way to go, but I think the community effort so far has been outstanding.

Organizing a community around such a huge piece of software is a big problem, the documentation is a big problem. There’s alot of it, but you can’t be sure how up to date it is or if its right.

Install is another issue, and we have folks working on Debian packges, the rest of the OpenACS stack can run on official debian packages already (AOLserver, etc, without daemontools, but that is not necessarily required of course).

Thanks for taking the time to share this.

Thanks for your thoughts.

Actually I’ve enjoyed my time with AOLServer. You’ll note that I didn’t pick on it even a little bit. It really is a nice webserver, and all of the documentation, such as for building SSL support and the TCL binding stage, has been reasonably accurate. It’s not terribly flexible, compared to Apache, not fashionable (the word AOL almost makes it a fasionable anti-fashion statement), and a I had to install a few extra components from third parties to make it all work (SSL, and some other stuff I don’t remember), but it works well and hasn’t caused me any trouble in over a year of operation. There are some in the OACS community that view AOLServer as a serious liability, but I’m not one of them. While I concur that the way the OACS docs push oddball tools for every single purpose (e.g., djb tools, when pre-installed OS-standard tools are perfectly acceptable for the same tasks) is problematic from an adoption standpoint, having it’s own webserver isn’t even in the top ten reasons why I’m leaving OACS. If it were the only oddball tool involved, I’d consider it entirely a non-issue. The fact that oddball tools make up the whole recommended stack, except for the database, is an issue, especially when some of the competition (Joomla and Drupal) literally takes about two minutes to install. But now I’m back to bitching about OACS.

I’ll switch back to the subject of AOLServer, and say, “nice job”. And to answer your query on a thread you posted last month, a webserver with embedded JavaScript would be cool. A little oddball, but cool.