Re: FW: News Flash! Informix FREE! on NT!!!!

jna@carmenta.se (Johan Andersson) offerred: +You have to register to get at this and to do so you have to accept among other +things the following: +> user may not disclose benchmark or +> other performance test results without Informix's prior written consent. +I would prefer that this 'NDA-look-a-like' registration be removed. As I +understand it this is not a 'beta-release' to developers motivating a +Non-Disclosure-Agreement. This is production release software, it should be +able to stand up to public criticism and testing. Am I wrong? In my opinion, yes, you are dead wrong in objecting to the above language. Oracle has done this for years - anyone wishing to publish a benchmark with Oracle needed express written consent to do so. On the one hand, this seems bad, as it prevents anyone with less than favorable numbers from publishing. But in fact, any published benchmark, even a fully sanctioned one, simply serves as a target for the competition to beat. So sooner or later your numbers will look bad (see my anecdote at the bottom of this response for an example of this in action). On the other hand what it does prevent is some boneheaded schmoes who know nothing about setting up and tuning the product from publishing a benchmark demonstrating that the competition's product (or even their own, for that matter) is "better" than Informix. With the "NT giveaway" you remove any financial disincentive for someone to do this, so you now need to protect yourself legally. Informix still places no prohibition on anyone who has purchased the s/w from publishing a benchmark (in fact, they SHOULD have such a prohibition, at least so they have the ability to review the benchmark and verify that it was performed honestly). When I worked for Informix, I did a LOT of competitive benchmarks. Way too often there were hidden (and not so hidden) agendas in place that put the integrity of the benchmark in the back seat to the furthuring of that agenda (most common: person or group responsible for the benchmark had already selected who they wanted to win, and did everything possible to slant the benchmark criteria in favor of their choice. I had one customer change the criteria for "winning" 5 FRIGGING TIMES, as we beat the competition under each new criteria. They finally based it on some "feature" that Informix lacked at the time, a feature that was never on their requirements list until that point.). By allowing anyone who wants to publish a benchmark, you have no way of ensuring that the results are fair and representative of the truth. I recall one Oracle benchmark ad where the Informix numbers used were from an early version of DSA on NT, pitted against Oracle running on a multiway Unix box. Of course the explanatory numbers were printed in cryptic language in very small text. Ok, the air up here on this soapbox is getting thing, so I'll get off now. Dave -- Dave Kosenko davek@summitdata.com Director of Training Services (732) 469-4070 Summit Data Group (an Informix Authorized Education Center)