thanks for the comment. I have heard similar tales from others who have deployed on the new system.

But to me, there was more appalling corporate-speak in the accompanying post: “Though we have had some bumps along the way, this platform has performed exceptionally…” I think the gnashing of teeth and cursing in college newsrooms as sites have gone down, sites have worked very slowly, emails to readers haven’t gone out for days and weeks on end, etc. exceeds the standard for “some bumps”. I’d be willing to bet that a survey of college newspaper editors at the papers running CP5 would NOT describe the past two months on CP5 as “exceptional performance.”

I don’t take away from CMN that they were doing something different on an entirely new platform, and there were bound to be problems to overcome along the way. And I give credit to their small team for working long hours to be responsive and try to make things work for their newspaper users. But it seems clear to me that they launched before the software was really ready because they had previously over-promised and under-delivered and a new school year was starting and papers were clamoring for what CMN had been promising for many months. They ended up handing themselves and their newspaper users more problems than they should have. I think the two month “time out” sounds like a great idea, to give them time to step back, take a breath, address performance issues like optimization of the database and other things they haven’t had time to do while launching new sites every few days, and to hopefully install upgrades from Polopoly/Atex without messing up sites the way some previous updates have done. I hope it leads to a much better 2009 for CMN and their newspaper users.

I wasn’t suggesting that you decried it as ineptitude. At most, I was only referring to a generalization of people who are particularly wary (or anti-) CP/CMN. Unfortunately, Lewis’ comment was bound to be read into by those who don’t understand database systems administration, which is why I shuddered when I first read it.

A single database “file” is infinitely more scalable — I’d be concerned if it wasn’t one database. Most importantly, to the layperson, 100GB seems like a ridiculous amount of data — “up to 25,000 MP3s,” I can hear college students equating it to — and most people would immediately suspect CP has no idea what they’re doing. And while we all question that sometimes, here’s an instance where we probably shouldn’t. In reality, 100GB is child’s play for a database, which can typically number in the many thousands of GB.

When I first read his post, I sent the link in an email to someone with the subject “Amateur hour.” Their response was expected: “Why is it one giant 100GB file? I don’t know anything about tech, but it can’t be good to have 550 sites access a singe 100GB file.” They had no idea (understandably) I was only labeling Lewis’ explanation — not the database setup itself — as “amatreur.”

Is the system optimal? I’m not sure any system ever is, and that’s bound to be the case with CP based on track record. That said, I think Lewis’ explanation was quite naive, and that includes his additional remarks you quoted. The words “100GB database file” was bound to raise eyebrows, even though a sysadmin wouldn’t flinch at the thought.

Where is the “decried as ineptitude” in my comment. I found it interesting that they are using a single database to serve out all these sites when they are looking to scale that up to 600 sites.

As Lewis noted in his original post:

“For context, it is important to understand that while this database configuration is not optimal right now, it was designed intentionally in this manner to make the sites in CP5 more flexible – both in publishing on a 24hr cycle versus an issue based model as well as making content extensible and modular.”

Perhaps it would have been better to write a separate post rather than include it in this one.