We got an unusual number of e-mails regarding the vulnerabilities in yesterday's release of Safari for Windows. I was a bit hesitant to cover it in a diary. After all, its beta software. We all know better then to use beta software in production. So operational impact of these issues should be nil.

Now... on the other hand, I own two Apple computers. So I know the power of the brushed-metal kool-aid. So lets talk about beta software in general. You got a sales guy in jeans and a black turtle neck, or a monkey running across a banner ad, telling you about the latest and greatest version of product "X". "Now with even more of must have 'Y'".

So how do you resist? I found its usually impossible. However, you can minimize the impact. Keep a "beta" machine around. Use it to install all the free trials, latest beta versions and other junk. The machine will soon become too unstable to use, making the desire for even more free-trial-super-feature-enhanced software vane quickly.

In very few cases you may want to use a beta product or a version downloaded and compiled from CVS. But these cases should be limited and strictly controlled. A couple of check points for approving a beta solution:

* Do we actually need the software?
* Is there a workaround that will make the "release" version workable?
* Is there a competing product that will do the job?
* Whats the track record of the vendor (will they always point to the next version thats just about to be released).
* How can we test if this beta software actually does what it promises?

Similar rules should be applied to any version upgraded or new software, even if its a "release". Sometimes, its better to stick with an older version for a while. At least you know how to work its bugs.

oh. and I am still typing this diary on my 3+ year old Linux system. Its the system I use to actually get work done.