So when should one upgrade?

In my last entry, I quoted Alien Bob stating that Slackware assumes you're smart. Part of what that means to me is having a sense a how to tailor the system to one's own uses, and having the freedom to do so without the operating system insisting it be done its way. But what, really, does running an OS entail? Beyond installation, most of it comes down to certain configuration options, installing additional software, and maintaining the whole mess. Everything else involves actually using the software. All of this is so users can do things like surf the web, check e-mail, write documents, work toward world peace, and launch the Space Shuttle.

Maintaining software is a tricky task. Sometimes one needs to upgrade software to apply a security patch. It is seldom a major version upgrade (although the recent jump from Seamonkey 1.1.18 to 2.0 is an exception), therefore many people call applying security patches an update rather than an upgrade. My policy is to apply security fixes within a few days of when they become available. The Slackware change logs are nearly always good at explaining why a fix is needed, and patches are made available promptly.

What Slackware does not do is update major pieces of the distribution between releases unless absolutely necessary. For example, KDE has had several point releases since Slackware 13 came out. None of them are available as patches for the -stable (released) branch. On the other hand, if a user really wants to get into the latest KDE and Slackware, everything is available in the -current (development) branch. I'm okay with that. That's one of the things that makes Slackware so stable.

But what about non-included software like Open Office? Personally, I take the same approach to those applications that I did (and still sometimes do) for Windows applications. Why would you ever want to upgrade MS Word or Adobe Photoshop? Why abandon Adobe PageMaker for InDesign? There is a tricky balancing act knowing when to upgrade to the latest software and when to hang on to old versions.

Taking MS Word as an example, Word 2000 is perfectly suitable for a vast majority of users. If that's too old, or one has compatibility issues with newer file formats, then grab Office 2003. More than half of the people I work with have Word 2003, and they're not suffering.

The same applies to Open Office. I don't really care if OOo bumps up a version. The old one works fine. When it doesn't work fine, then I look at upgrading. Generally speaking, the only compelling reasons I've found to upgrade are when (1) significant new features are added,(2) a security patch is available, or (3) there are very serious compatibility issues with file formats.

Otherwise, there's often no good reason to upgrade. On the flip side, eventually it's a good idea to upgrade and use the latest stuff. I generally do that after I upgrade Slackware for at least some of the packages I use.

In short, it's better to be a late adopter than an early one.