I think the best solution is to have the documents generated in a format that matches the final form necessary for the task. Otherwise, work will need to be done to convert them to HTML or PDF or whatever, and there will be concerns about which document is the latest version, etc. But that means taking the time to convince or train people to use the tool of interest for this task rather than what they're used to.

As I mentioned in my other reply, it seems to me to be less work to modify the documents as little as possible, but make it simple for users to find whatever document they need. You would still need to spend the time to generate indexes or lists of key words if you need them. I don't know of a good way to do that automatically. That might be an advantage of a Wiki approach, but I don't think a Wiki is suitable for casual users (though YMMV).

There are many tools out there to convert .DOC to HTML, but I don't know how good they are.

This is a problem that has been solved, but I imagine that it's in the context of internal documentation (perhaps SGML-like stuff). I would be surprised if there was a free, high-quality, intuitive, and easy to use solution. Perhaps Fu has cracked that nut. :-)

So, I'd vote for making a simple web page with descriptions of the existing documents and pointers to the original files. The client machines would automagically open them in the appropriate application. If your colleagues want something fancier, then I'd branch out from there. If the original documents are modified, perhaps a "simple" script can be run to update the web page based on changes in time stamps, etc.

HTH a bit. Good luck!

Cheers,
Scott.