Forum OpenACS Q&A: Relationship between ACS/pg and ACS/Oracle

I was wondering how you were related to ArsDigita and Philip. I know
that I've seen Ben in several places, but it sure doesn't seem like you
guys work for ArsDigita. Does aD support you guys or are they totally
against it? I know that a pg version will definitely be a big boon to
the ACS, so aD would be just silly to be against it. So what exactly is
their position?

Thanks for the clarification.

Posted by Roberto Mello on

I don't think aD is against us at all... their customer base is probably in a whole different level, at least for now. They aim at big customers with huge bucks.

Ben wrote part of the original ACS, and can probably answer this question better.

ACS is an awesome toolkit and it was sooo said that it was only available to Oracle.

Posted by Ben Adida on
ArsDigita's goal is to spread the religion of the ACS as much as possible. In that respect, Philip has expressed great interest in seeing ACS/pg come about, but he's been quite busy so hasn't been really involved.

ACS/pg is not meant to compete with ACS/Oracle. It is meant to allow anyone to set up a fully open-source, enterprise-quality web application. Open-Source is taking over the world, and there's no reason to stick with the concept only at the Tcl layer :)

In the end, the ACS/pg team and ArsDigita coexist happily. In the true world of open-source collaboration, we hope to be able to push back bug fixes, improvements, and more to ACS/Oracle.

Posted by Aaron (WRONG ACCOUNT) Swartz on
Why of course in theory! But the question is in practice. It seems that aD is not willing to put in an abstraction layer for you and that you seem to be having some trouble keeping things up to date. So this is a straight fork? No relation at all with aD (other than borrowing source)?

It seems Ben is on all sides of the open-source thing. He's open-sourcing the ACS for Postgres, helped open-source AOLserver and I wouldn't be surprised if he was working on Postgres too!

Posted by Ben Adida on
Oh no! I leave the Postgres stuff to Don Baccus who's a much better hacker than I am.

It's true that the ACS could be better abstracted. It's true that the task of keeping ACS/pg up-to-date with ACS/Oracle is daunting when the number of ArsDigita developers is growing rapidly. And it's true that ACS/pg is basically the result of a handful of people borrowing the source from ArsDigita and hacking it out on our own (not counting a quick call to Philip every now and then to complain about various aspects of the ACS).

We accept that ArsDigita has better things to do with their time than help people who can't pay for Oracle licenses... if ACS/pg just ends up being a fork of ACS with the occasional merge, we're also okay with that. After all, it's not like a line-by-line port will ever be necessary. We *will* make one strong effort: to keep the data model as close as possible with that of ACS/Oracle so that it's not impossible for people to migrate one way or the other.

As for my involvement in various open-source projects....yes I admit it, I love open-source :)

Posted by Don Baccus on
I think it's important to understand that while aD is interested in spreading the religion of the ACS, as Ben puts it, aD-the-company is structured around building complex websites for companies with deep pockets.  They've got more business than they know what to do with, which is why they're recruiting people far and wide.  Given Philip's goals and aspirations, it's the only strategy for them to take.

Certainly making the ACS available on Postgres can only help aD in their secondary mission of spreading the ACS religion.  But doing so isn't of much help in the primary mission of establishing aD itself as  a force in the web development world.  For the class of customer aD is chasing, Oracle's still the right choice and will be for some time.  So it is only natural that while Philip might be personally curious about the ACS-PG port, he doesn't see it as important in the short term to aD.  And their focus is very much on the short term at the moment.

This reality is also why parts of the ACS are somewhat disorganized, they're just not in a position to sit back and clean things up.  aD's just moving too quickly.

I'd love to say that Postgres or some other Open Source alternative could unseat Oracle in that world, but we're not there yet.  And, frankly, the cost of an Oracle license is just a fraction of the start-up cost for major sites - the Garnter group claims $1,000,000 is  typical.

Long-term, I think the full Open Source solution has incredible potential.  Short-term, I started porting small portions of it last year because I'm working on a personal project that benefits from  using the ACS and Postgres.  That's enough motivation for me.  What I hope to see short-term are intriguing sites thrown up in short order by ACS-PG afficianados, I can think of nothing better to trigger broader interest in not only the port, but the ACS per se.

So, don't just play around folks, do something cool!  Given the cost of deploying Oracle, there's no reason why the majority of cool ACS-based sites shouldn't be based on the Postgres version.  Make it so!

aD could really benefit from this port in ways that perhaps Philip's not considered.  For instance, someone who's built an intriguing site using the ACS-PG port should be a good aD employee candidate, able to immediately jump in and help with contract site development.  And so forth.  So as time goes on, perhaps aD will work with us to make it easier for us to keep the port aligned.  Perhaps not.  I'm not depending on it, and neither is Ben.  I don't judge aD negatively because of this.  After all the code's available and porting it is a heck of lot easier than recreating it from scratch, with or without active help from aD!