I like interesting tech. If it looks fun and useful, I’m apt to make a point of learning it. Of course, this “fun and useful” is a personal judgment, and I’m geeky, so what I choose to look at probably isn’t that interesting to other people. Still, here is some insight into two bits of technology that I’ve learned in the past couple of years.
LDAP: I first became interested in LDAP because I thought it would be nice to have all user information stored in one place. However, if you don’t have a good grasp of directory service to begin with, LDAP is likely to be trouble. In the end, LDAP enables some nice things (like Netscape Navigator 4’s Roaming Profiles, which, by the way, is the entire reason that Navigator 4.8 exists — because Mozilla doesn’t support Roaming Profiles), but it turns out that while LDAP is a very nice technology in theory, there are very few tools to support it on the administrative side of things. That means that once the data is in the directory and you have your applications using the directory for user-lookup, it works wonderfully. But, if you want to change information in the directory or load information into the directory, lookout!
What’s the problem here? You want your users to be able to modify their own information in the directory. However, if you have a diverse population of users, their common mode of access is likely to be the web browser. Great. I have been able to find virtually no well-written tools that allow end users access to their own data on a directory server. The concept is great, but the implementation sucks.
XSLT: XSLT is an XML transformation language. Like all things XML, it is extremely verbose. However, like many XML technologies, there is a lot of support for it. The primary benefit of XML has been that it is easy to build tools for it and have those tools be useful in a variety of applications.
It has proved worthwhile to learn XSLT because I can use the transformations I write in a variety of applications. Right now, I’m using AxKit, but tomorrow I could switch the XSLT transformations to Cocoon and they would still work as expected (modulo bugs). (Though why I’d want to switch from the the relatively fast AxKit to the dog-slow Cocoon is beyond me.)
In the end, I think the difference between these two technologies is who they are aimed at. LDAP’s primary target is the System Administrator. XSLT’s primary target is the developer. Another reason for the difference would be the relative popularity of the standards creator. The University of Michigan (not known as a technology hotbed) created LDAP, but the relatively prominant W3 standardized and promotes XSLT.