Here's a detailed description of what I meant, but keep in mind this
is off the top of my head and is just a single example of the kind
of killer app you need for your technology.
What I was suggesting is that perhaps you could extend your code to
read in a Word doc and detect the "forms", i.e. the text that is set up
like forms with lots of underscores for the various blanks, and import
that into your tool in a way that can be published to the web (and indeed
transformed in a host of ways).
It's possible that this is unrealistic, and that the things people do in Word
are just too unstructured for a tool to detect. It would be a bit like OCR or
face-recognition software, only it would be form-recognition and would work
with text rather than pixels. Still, this kind of thing is a good fit for a Lisp
program, if it's doable at all. Supposing it's doable, I think you'd get some
huge benefits from it:
1. People could plug in the docs they already have and get immediate value
without having to learn your tool. That is huge.
2. There are a zillion Word docs floating around the web, so you won't lack
for data to test.
3. It will make for a compelling demo. Instead of demoing your format, you'd
simply make a Word doc and upload it. Better still, upload some Word doc
you've never seen before and show that it captures the info.
With what you've shown us so far, your big hurdle is going to be convincing
people to learn your format. The advantage of the above idea is that they wouldn't have
to learn your format; you (or rather Formlis) would learn theirs.
There would be ugly glue code needed to interoperate with Office. But
other startups (Xobni, Docverse) have shown that this is a pretty good path
for delivering immediate value, and that it gets people's attention.
In short, Word is where the users are. Don't expect them to come to you; go to them.
Dan