I use HTML::Template, and while it took a little mind-bending to get comfortable, I now swear by it. The functionality it provides is fairly basic, not hard for non-programming designers to master. Essentially:

• Named variable interpolation (Tag: TMPL_VAR)
• Simple IF / UNLESS <condition> variable display (TMPL_IF)
• Repeating sections (TMPL_LOOP)
• Sub-template includes (TMPL_INCLUDE)

One can build apps of amazing complexity using just these basic tools. The key is that you totally seperate the data from the presentation.

The drawback is that you have to decide where you draw the line between what is in the template's domain, and what is in the perl. For instance, I don't think a perl script should ever print table tags/elements. However, I don't put INPUT tags in my templates, preferring to use CGI.pm's methods for generating form widgets, and pass the output from those to the template variables.

Basically, I never put HTML directly in my perl. What little HTML is generated by my perl is done in an abstract way by using modules and objects, so that repurposing would not be too difficult, and would hopefully not have to break the 'API' that my scripts use.

To get a full and total separation, I agree with Russ, XML would be the direction to squint in. I'm not there yet myself.

For a more complex templating system, look at Template::Toolkit. I have friends that swear by it, though I've not used it myself.

Cheers!