Great article Talli. My few comments are below. I must apologise in advance that I am a harsh editor.
I would qualify "but none of them have the track record or the fundamentally sound data model as the OpenACS" with some reasoning. With no information to back up this statement it is simply rhetoric. Same goes for calling other applications data models "idiosyncratic".
In the same section, I would make a bigger deal about the permissioning structre - it is a major asset even to sites whic may use no other features. I also wouldn't use the phrase "cobble together" - no-one wants to think they might get a solution that is cobbled together - especially since you denigrate ad-hoc development in an earlier paragraph.
On a different tack, I would probably mention that Postgresql plays in the Lamp space as well - so that you give the feel that the OpenACS solution is not totally divorced from the lamp world. It might also be mentioning that Tcl is a popular language in its own right, albeit in a different problem space.
All in all a great summary though.
