May 2003

XSLT: Creeping Out of the Closet?
by Brian Dunning

digg this article | del.icio.us this article

It seems that, after a couple years of FileMaker, Inc. staunchly pushing XSLT as a web development technology, a few sprouts might be taking root.

When FileMaker first announced that their XML support would become the foundation for their newly preferred web technology, I immediately started work on my XML Master Class book. But, after some research and talks with other developers, it was clear that virtually nobody outside of FileMaker, Inc. was even thinking about XSLT as a technology for building data-driven web sites, let alone using it. So I dropped the idea, and there the book has sat ever since, languishing on page 3.

And to this day, XSLT does not appear on the list of the top 100 technologies most searched for by developers (http://sourceforge.net/top/mostactive.php?type=week), as of this writing. I emailed a sagely PERL developer (and former Fortune 1000 CTO) to ask his thoughts about XSLT as a technology for building data-driven web sites, and here was his learned reply:

??????
that's not what xslt is for.

Which was what I had always thought. But, a quick browse on Slashdot.org has turned up a few odds and ends. For the first time, I saw references to XSLT as a technology for building data driven web sites.

New open source offerings such as StrutsCX (http://sourceforge.net/projects/strutscx/) promise greater adoption of XSLT among developers who embrace conventional web architectures and important frameworks such as Struts. This is a positive sign for XSLT's future, but is it something I would thrust in the face of a K-12 teacher trying to put the class FileMaker database on the school's Intranet? Probably not.

O'Reilly even has some books out now (http://www.oreilly.com/catalog/xslt/) directed toward the general developer community, though they still come with such disclaimers in the abstract as "Useful as XSLT is, its complexities can be daunting." Getting the daunt out of those complexities is the purpose of such books, of course, and hopefully time will bring us some software tools that make a book's guidance more practical.

So with this growing validation in the community, is there yet any advantage to adopting XSLT now as a front end and back end technology for our FileMaker powered web sites?

I can answer one aspect of that question by noting that XSLT editors are still in diapers, and still soiling them. Those that exist, such as XML Spy, are optimized for what XSLT was originally intended for: back end server communication and data transformation. There's still no such thing as a "wizard" that might compare to Lasso Studio, as one example, for walking you through development of basic web functionality.

Of course, most advanced web developers never touch such tools, instead preferring text editors to develop the highly customized code that the wizards don't accommodate. That doesn't do much for the novice or the casual FileMaker user. Those people are still out in the cold, as far as easy development environments go.

So what about those experienced professionals? Is it time for them to adopt? Perhaps, but XSLT is a substantially different paradigm from more familiar code-based languages such as PHP, Lasso, Java, and ASP.NET. Familiarity with conventional coding practices doesn't necessarily translate into fluency with XSLT.

You will hear varying opinions on this: but in my estimation, by no means is XSLT something that anyone other than an advanced, professional web developer should try and tackle (and even then, they have other far more appropriate tools at their disposal). A FileMaker developer will realize a better return on the investment of time by buying Lasso Studio, which is easily usable by point-and-click humans, rather than spinning wheels for endless hours in a text editor trying to make the simplest functions work in XSLT.

Until XSLT appears on Sourceforge's top 100 list (at the very least), I will stick with my original contention: that adoption in the general community, not just a few shoehorn wielders in the FileMaker community, is an important component in the viability of a technology. XSLT is certainly not going away, by any stretch. But will it - and should it - ever become the web developer's tool of choice?