From: Ronald I. Nutter [mailto:[EMAIL PROTECTED] > > I am at 2.6.4. After seeing all the horror stories of upgrades and > my lesser knowledge of Linux as compare to most on this forum, I was > reluctant to go forward with the upgrade.
The thing with all the upgrade horror stories is that there are many times more people who upgraded with no problems and just didn't post about it. Human nature says that people complain when things go wrong, but take it for granted when things work properly. I have kept my systems up to date since I originally installed SA 2.55. I'm currently running 3.0.4 and I'll move up to 3.1 in a week or so once everyone else has ironed out any problems. :) I use CPAN to do the upgrades and I have never had any problems with them. The last upgrade required me to change my Bayes learning scripts to use the new option names, but that's it. CPAN works for me, but you should stick with your original install method (rpm, yum, build from tarball, etc) to avoid problems. Different install methods place files in different locations, so if you mix them, it can cause problems. Go ahead and upgrade. The benefits usually outweigh the risks. Just make sure to read the README, INSTALL, and UPGRADE documentation so that you are aware of any changes in structure, options, requirements, or features. It's also a good idea to run through the list of required Perl modules and make sure you have the latest version of all of them. As far as a lesser knowledge of Linux, that's not really an issue here. If I remember correctly, most of the upgrade problems were related to Bayes database issues. So the worst case is that you delete the database and start Bayes from scratch. With all of the other enhancements in SA 3.1, that's not a bad deal even if it were guaranteed to happen. Bowie
