Brian Sweeting

In his July 14 Alertbox, Jakob Nielsen comments on the usability of PDF points out that PDF should be used only for documents that need to be printed.

This is good advice to follow. I hate it when I click on a link that should be in html and it turns out to be a PDF. This is especially bad since Adobe Reader 6.0 was released with its unbearable splash screen and load time.

I attribute the overuse of PDF to laziness. Most of these documents are usually created in a program like Microsoft Word. From Word it is much cleaner and easier to just export it as a PDF document rather than html. Word is notoriously bad at the junk code that it spits out to generate an html document. I have my fingers crossed that Office 2003′s XML support will make crappy html code a thing of the past.

If you really have a document on the web that also needs to be available as a PDF, take a look at PDFlib, a library for processing PDF on the fly that can be interfaced with COM, C, C++, Java, Perl, PHP, Python, Cobol, RPG, or Tcl. I have used it before with PHP and it works great.