Russell Beattie

Ahhhh... It seems that Rickard was on to the truth, thank goodness. He's put up a detailed rebuttal online pointing out the flaws in the report, and The Server Side has also posted more information on their forum, including this bit:

* Was Microsoft involved in this, did they fund this, where were the tests done?

Yes, Microsoft was certainly involved, as the paper describes. The Middleware Company approached Microsoft regarding performing such an experiment. Microsoft provided the lab, which was located in Seattle, funded the setup costs, and reimbursed us for expenses, including travel expenses.

ARE YOU KIDDING ME? I hadn't read the report before I posted before (err. I still haven't, actually. I won't waste my time now.) People are nitpicking on the report for various problems ("Why use Linux?"). My big problems are the fact that this solution wasn't the best. No one in the real world is going to do an app the way that The Middleware company did theirs. Also BEA and IBM weren't part of the process, but MS had a chance to tweak their system.

That last bit's important. I've seen for myself what having MS on board an do for a project: I was working on Starbuck's original eCommerce website in 1998 - all in Microsoft. We had a problem with SQL Server replication between the staging and production servers (it would work and IIS would go down, or vice versa... ) Finally the only result was a high-level call to MS and we got sent some weird-ass patch directly from the developers that solved the problem. I'm quite confident that the people at MS can get their systems running perfectly and with high performance anytime they want (well, not really. Hotmail's still on Unix, right?) but for the rest of us who aren't an international firm with contacts in Seatle, we have to suffer. There shouldn't have been ANYONE from MS involved in this study, let alone having the freakin' servers in Redmond. What a joke.

Ugo's got some thoughts on this too:

Lies, damned lies and benchmarks

...

IF [Microsoft's involvement] is true, TMC have really shoot themselves in the foot with this one. I have EJB Design Patterns and Mastering Enterprise Java Beans on my Amazon wish list (assuming I will ever find a project that requires the usage of EJBs), but after this I'm not that inclined to trust them anymore.

I do not agree with Rickard, however, when he writes:

It is bad for MS (really), because they are linking to this report, helped create it, and will be using it as a marketing tool against J2EE. They are used to being called FUD-makers, but perhaps not in this way. Gotta hand it to them, usually good MS FUD is a lot more subtle and "kind-of accurate" than this report.

I think MS has already won on this one. 99% of managers are going to read MS's take on the report and maybe one out of ten will read the actual report and one out of 1000 will ever find Rickard's analysis or other similar arguments. The FUD is spread and there's no way bringing it back.

I'm starting to feel sympathetic for Russell's view: Say NO to .Net.

Suddenly my rantings are making sense, hey? (Crap, that's scary even to me.)

-Russ