Forum OpenACS Q&A: Looking to hire ACS (TCL) Guru Developer

Someone suggested that many of you would be interested to hear the reason for converting to ACS.
I'll detail that below, but obviously some details and names will have to be left out to protect the innocent (and those not so innocent). ;)
I came on board with the company in September 2000.
A project was already many months under way with a consulting company to build our company a fairly complex combination site.
The site was to provide:
eCommerce
Partner network of content providers and content distributors
Content management system
And more.
I stumbled across ACS for the first time at that October's Oracle Open World 2000 in SF.
I come from a broad background, and one of the companies I worked for specialized in implementing document workflow and content management solutions such as Filenet, Documentum, inConcert, and others. And I used to implement and support these systems.
I kept asking the consulting company why they opted to build everything we needed from the ground up, rather than getting something off the shelf that was at least 80% match to our needs.
They kept insisting there was no such product, and I didn't buy. But alas the momentum was already under way, and I as the new (and youngest executive by at least 10-15 years) CTO, wasn't yet proven in my opnions and instincts.
I saw the source code of this custom product 2 weeks before launch (I tried for months to get my hands on it before them, but was delayed by the consulting company daily).
I could tell once I saw it, that it was going to be a disaster. We launced on it December 13th despite my protests.
It's been hell ever since.
The system doesn't scale to out needs.
The user & group management system is a joke, and has to be hard coded to add anything new.
Most of the upload and download bulk utilities didn't work right until months more of work on our team's part.
We've been unable to add any new functionality because the system is totally monolithic, and in no way modular.
Any changes to any part of the code causes many unforseen failures.
We needed a system that our small team could manage.
Could work on in a modular fashion. Fixing fires in one area of the system, without breaking everything else.
Easily add features.
Has a high caliber user and group management system.
Flexible workflow and content management system.
Decent eCommerce system.
Strong portal/cobrading capabilities.
Scalable for traffic, content, users, searches, data imports and exports.
We work with thousands of pdf's and images at a time.
If it wasn't for the current systems' bottlenecks, our current 50k titles would be 4x greater already.
After months of haranguing the executives, they each one by one finally reached their pain threshold, and were willing to follow my recommendations to move to another product, and more specifically ACS.
I researched, installed, and used ACS Java 4.6, ACS TCL 4.2, OpenACS 3x. I ran each on my own personal websites over a 4 month period at www.merp.com, tolkienhome.org, lotr.ws, manicmechanic.net, and alpentech.net, colorificsigns.com, and other sites.
I now run my busiest and most community centric site (merp.com) on OpenACS (I liked ACS better, by I personally couldn't afford the Oracle license).
It became clear that ACS 4.2 TCL was easily 80%+ match to our current and future needs.
ACS Java was more in line with out development teams skill set, but the Java version would require to much coding from scratch, and doesn't have any proven implementation or enough community adoption(yet).
Maybe in 1-2 years when there are more modules, etc. We'll convert to that, but for now ACS 4.2 appears to have what we needed, just about ready to go.
Now we just need someone who keep show us the most efficient way to do that. :)
I hope this helps others considering these prospects as well.
I'll post to this thread over the coming months as to any issues, and progress made.
The deadline to launch on this new platform, is a VERY solid March 11th 2002. And if it can be moved up 1-2 months sooner, so much the better.
Thanks,
-Hawke
CTO
MightyWords Inc.
www.mightywords.com

Also, the excellent efforts by the testing volunteers within this community are really helping to ensure the OpenACS is of a quality level that matches its implementation and design acheivements...

I'd definately recommend picking up the OpenACS alpha...

Hawke!  Whitespace is your friend!  Well, your readers' friend, anyway!  That's a lot of dense text to wade through...

A couple of comments.  First of all, there are a bunch of very recent ex-aD employees who are available at the moment, so you should have a decent talent pool to select from (and there are other, non-aDers who might make good candidates. too).

I'll second the above comments in regard to OpenACS 4 vs. ACS Classic 4.2.  The OpenACS 4 code base is the ACS 4.2 code base with bug fixes and some enhancements, including a new "CMS lite" that provides an alternative to the "CMS heavy" built into 4.2.  Sloan/MIT is investing in a new education platform based on OpenACS 4 named dotLRN, and as part of that the portals package is being totally revamped and improved (the dotLRN system, much like ACES before it, will be portal-driven).  That will be released in alpha form in a month or so.

And OpenACS 4 supports both Oracle and Postgres, just in case you've been thinking it only supports Postgres as is true with the OpenACS 3.2.

In other words, OpenACS 4 is a superset of ACS 4.2 and new packages are being worked on by a variety of people.

As far as liking the ACS better than OpenACS, keep in mind that OpenACS 3.2 is a port of a relatively ancient version of the ACS.  It's a clone of a point in history.  Rather than continue to incrementally upgrade and track various aD package enhancements we made the decision about a year ago to just wait for ACS 4.x to stabilize, then port that.  With aD's dropping of the Tcl version, we decided to rewrite it to support multiple RDBMS rather than simply port to Postgres and here we all.

Feel free to ask questions regarding ACS 4.2 vs. OpenACS 4 specifics.

There are many of us who never considered ACS 4.2 production-quality software, frankly.  That's one reason for our labelling our current work an "alpha" release.

Also, we're working on an RDBMS-independent search package (that obviously has to interface with RDBMS-dependent things like InterMedia) and until both Oracle and Postgres support that fully I'm hesitant to label our release "beta".  This is my next, personal upcoming task.

However you can still use the ACS 4.2 site-wide-search package with the Oracle version, and it's still included along with the deprecated acs-interface so folks will have a working solution in the short-term.

Your timeframe is very tight, so I do understand your being a bit queasy about OpenACS 4.  The OpenACS 4 framework will support ACS 4.2 packages, however, so you would have that as a fallback if one or more of our versions were to introduce errors.

Hawke,  we (furfly) have built a couple of sites based on ACS 4.2 Tcl and I can confirm the comments here that it was never quite ready for prime time.  We've done a lot of bugfixing along the way, but there is still more to do (if these were exceptionally busy sites I think we would be seeing more problems than we are).  If I were in your shoes I believe I would try OpenACS;  you might still have to do bugfixing but at least there is a community to contribute them back to, which is not really the case any longer at aD.