Forum OpenACS Development: Times are changing

It is interesting to reflect back on the second posting made when the openacs forums were first opened. The topic of discussion then was where does ACS/pg fit in? People were concerned back then of how PG will keep up-to-date with Oracle.

Today, it is a different story. We are now concerned with keeping Oracle up-to-date with PG!

When I came across the OpenACS community 5 years ago, I found a project that wanted to commit to a full open source stack (web server, language, db, os). The main reason (I think) was to distinguish OpenACS from ACS. I think it was always the goal for OpenACS to support PostgreSQL. This was probably why scripts were created to port Oracle queries to PG queries back then, so as to make the transition from Oracle to PG.

Things changed when Arsdigita got swallowed up. Those that were able to snatch up some of Arsdigita's clients ended up continuing with ACS support, as well as transitioning some businesses to OpenACS, whilst keeping data in Oracle. Since a clean break from Oracle could not have been established in OpenACS, Oracle support was continued.

Now we have reached another interesting point in the history of OpenACS. Do we continue with Oracle or not? I personally would like to see OpenACS become a completely open source stack by severing ties with Oracle. However, there are those that rely on Oracle, as they believe it to be a marketable feature. If this helps them get business and spread the OpenACS gospel, then so be it, but they should also commit resources to supporting their preferred DB, instead of just leeching off others.

Within the last 2 years there has been a lot of discontent with supporting oracle. This occurs mostly in irc, but also appears in the forums from time to time, so it isn't something that "just happened". During this time nobody has stepped up to defend Oracle. So most of the people that have been contributing code were under the impression that nobody uses oracle, except for a few legacy ACS systems. I know I was thinking that. I didn't know people were still wanting Oracle support until this thread.

For me, supporting oracle is a problem. It means having to write more code than what a client is willing to pay for. I have not come across a client that is willing to pay for extra time spent on supporting a database that they will not use. In fact I have not come across a client that has demanded Oracle, so I welcome the drop in support for Oracle.

I find this outrage about dropping oracle support to be quite amusing. If nobody has been supporting oracle for the last two years, with the exception of Don, then what makes people think that creating a new openacs/oracle project will suddenly muster community support? I would like to see this happen, I really do, but I'm afraid it is just hot air, just like this thread. It would be easier to just maintain the 5.3 branch.

Moving forward, I would like to see the Active Record pattern employed in openacs, as is done in RoR and Turbo Gears. Perhaps we can look towards using xotcl for this. That way we won't have to bother about strictly supporting a specific type of database. We might get there, but it would probably be a few years away if we ever decide to go down that path. Hopefully it will happen in time for OpenACS 10, aka OpenACS X! ;)