hence i don't really understand on what basis the comment was made about the lack of scalability of zope's persistence. rereading the don's reponse it appears that the comment was related to general scalability, and on that question i don't understand from where that was inferenced,
The inference comes from a client's experience, one who has chosen OpenACS rather than Zope for their next project.
As I said, though, I don't know if the slowness they experience is due to issues with the persistence layer or other design issues, i.e. with the forum software and other packages/modules/whatever-the-hell-they're-called-by-Zopefolk that make up the site.
I did specifically say I didn't know if the problem was with the persistence layer - you did read that, right?
This is just one data point, admittedly.
As far as scalability goes in this context, remember that old saying, "size doesn't count". It's hits per second that count. Telling me that there are ZODB databases up to 10 GB in size is as meaningless as my telling you that there are PG databases in the several hundred GB range (as is true). The client's experience is that their site's so *slow* that it's nearly unusable on a P500 with ample RAM and hard disks. So *slow* that they're unlikely to ever reach 10GB worth of data, not in this lifetime anyway.
Again, I don't know why it is so slow, nor do I have any idea as to how much work would be involved in speeding it up. This wasn't why they chose OpenACS for their next project, it was shown me as a curiousity, that's all.
On the other hand I've never seen an OpenACS site as slow as the one in question ...
