Forum OpenACS Q&A: History of OpenACS

Posted by Maciej Dakowicz on
Hey guys,

Some time ago I posted message about my OpenACS master thesis. I've got new questions as I have problems with finding exact history of OpenACS.

As I said before, I'm not a hacker but rather person looking at OpenACS from KM perspective. However to finalize my work I need to get also some information from technical point of view..

Can you please tell me how the story looked like exactly? I found description of the whole case with ArsDigita written by Philips Greenspun ( However I'm interested rather in changes implemented in ACS in time.

1) When company decided to use Java? This change was caused because ArsDigita wanted to get rid of open-source products and start providing close-source ones?
2) When exactly OpenACS was created? How it happened? After closing ArsDigita? Which version was that?
3) How was created? By whom? When?

Thanks in advance for help,

2: Re: History of OpenACS (response to 1)
Posted by Lamar Owen on
I can give you a little history of the OpenACS piece of it. OpenACS was originally known as ACS/pg, that is, ACS for PostgreSQL. It was started by Ben Adida and Don Baccus in late 1999 (not sure of exact month; my e-mail archive of the acspg-devel list starts in December, but I know it was in the works prior to that). Roberto Mello, Jamie Ross, and I also were there at the beginning; Roberto did lots of testing and doc work, and there were others involved, too. I did the initial AOLserver 3 port of the postgres driver (I had just started maintaining the PostgreSQL RPMs that summer) and fixed some issues that Ben really appreciated. Don further improved the driver. I ported what are now known as the FOR_ACS_USE extensions from the Oracle driver used by ACS in early 2000. Lessee, the biggies were ns_table and ns_column, which AOLserver 3 had dropped from the nsdb API.

Red Hat has the archive of the old ArsDigita forums. You can find one of the initial threads at (some of the names should look mighty familiar!). The whole archive is available, and will give you a good idea of the history from early 1999 through late 2001/early 2002. See for Red Hat's home page for the archives.

Ok, I found the initial message thread. OpenACS, 'nee ACS/pg, started in the thread when Ben and Don began collaborating. This was December 1, 1999.

Also search the forum archives here; I found and (this last one shows the name change to OpenACS).

I found an old bboard post from August 2000 in my ACS/pg e-mail archives, but am not sure which bboard it came from, or I'd post a link. This message discusses a visit by someone related to OpenACS to ArsDigita and a meeting there, where ArsDigita finally realized the open source potential. I can forward it by e-mail (it's probably too lengthy to post here).

Wow, lots have changed since those days. Joel's release manager work, the coming and going of Peter Marklund, wow. Lots of things have happened, and I am glad I have been a small part of it.

3: Re: History of OpenACS (response to 1)
Posted by Torben Brosten on
4: Re: History of OpenACS (response to 1)
Posted by Andrew Piskorski on
ACS Java (aka Red Hat CCM, aka Byline) isn't really all that relevent to the history of OpenACS.

OpenACS started as a PostgreSQL-only port of the Oracle-only ACS. I'm not sure what version of ACS OpenACS started with, but both ACS and OpenACS existed in parallel for some time, through ACS 3.2. ArsDigita continued the ACS line of development through one or two more 3.x versions, then 4.0 and 4.2, and then abandoned the codebase.

ACS 4.0 and 4.2 included a bunch of fairly radical changes from ACS 3.x, and unfortunatley as-released were never feature-complete relative to ACS 3.x. But, the OpenACS folks liked it enough to adopt it. So OpenACS picked up ACS 4.2, enhanced it to support both Oracle and PostgreSQL, made lots of other fixes and improvements, etc., and released the result as OpenACS 4.5. Then the OpenACS toolkit continued to evolve from there, to what you have today.

When ArsDigita stopped all further development of ACS 4.x, they shifted the "core team" to starting work on ACS Java. I don't remember when that was exactly, either late 2000 (Nov. or Dec.) or sometime in the first half of 2001. Also sometime in there, ArsDigita officially end-of-lifed "ACS Tcl" and pointed existing users to OpenACS.

I believe early versions of ACS Java would have had the data model in common with ACS 4.2, but probably nothing else, and then started diverging further from there. I don't really know though. I think Jun Yamog is familiar with it, ask either him, or the actual ACS Java / Byline developers.

So, the OpenACS codebase of today is the lineal descendent of ACS all the way back to ACS version 1.0 and earlier, and ACS Java was essentially a radical fork of, which the OpenACS maintainers knew about but had no interest in following.

As for why, back at the tail end of the dot-com boom, ArsDigita twice deprecated their old version of ACS before releasing a feature-complete new version, and tried to jump onto the Java bandwagon... Well, different people tell different stories. There's never been - and probably never will be - a truly definitive account, but plenty has been written on that, search around if you're curious. But again, if what you're interested in is the history of OpenACS, that's pretty peripheral.

5: Re: Re: History of OpenACS (response to 2)
Posted by Roberto Mello on
Don't forget that *I* started the thread where Don and Ben began collaborating :-) Really all I wanted was to run ACS for my Linux user's group web site. Back then we didn't have the gazillion CMSs that exist today.

The ACS/pg project was for me a great opportunity to work with and learn from more experienced people, including Ben, Don, and Dan Wickstrom, who did a lot of the porting work of ACS 4 -> OpenACS 4.

I also managed the last release the 3.x series, 3.2.5.

Maybe we should have a History page somewhere.