You best bet would be HTML::Parser or one of its subclasses. I'm betting that HTML::TreeBuilder would be your best bet in this case.

"Perl makes the fun jobs fun
and the boring jobs bearable" - me

Remember HTML tags that 'serve a purpose' but aren't links or text. If you have a link, and it has an IMG instead of text, you're going to need to put something so there is a link for your viewers. Perhaps for all images use the ALT tag if it exists (or the name if it doesn't) so that the context of the picture can be kept. Look at various HTML for the blind sites on the web for other ideas about translating multi-media pages into more simple forms. Most sites are no longer lynx compatible.

Also, much infomation can be coded into tags. One example is a table, especially one with some blank boxes. Just displaying the text can loose much context.

You also need to figure out what to do with frames - perhaps a top-of page marker that there are frames and the ability to flip through them.

Since you're talking about stripping HTML, I'm assuming that this means that you are not going to sites written just for your Palm program. Look hard at what tags can be dropped safely and which need to have some analog.

Good luck.

=Blue
...you might be eaten by a grue...

Anton