miracle help

[Steve Owens <uso01 -at- EAGLE -dot- UNIDATA -dot- COM>]

> OK.  I'm curious.  We have developed a tool that we are using on our own
> stuff.  I want to know if it would be marketable.  In other words, how many
> of you would buy this and how much would you pay for it?

        It probably could be marketable, depending upon details that aren't
in your post...

> The tool would allow you to use the word processing or desktop publishing
> package that you prefer.

        This implies a markup language of some sort... are you thinking in
terms of a PC, MacIntosh, character-based UNIX, or Xwindows-based help
viewer??  I suspect you're not thinking in the broader context here.

> It would allow you to maintain a single source for written and
> on-line doc.

        Always a good idea, but not necessarily automatable.  The raw
information can be single-sourced, but different formatting and
organization work for hard vs. soft.

> It would not require the insertion of millions of hypertext links.
> It would not require mapping on a white board or posty notes.

        This sounds like a very hypish marketing claim... without some
more tangible description of how it does this (i.e. "it provides full
context-sensitive text mapping" or "it uses an AI expert system to
create the links for you") I'd be very skittish of committing money to
it.

> It would produce files that work with any on-line help viewer on any
> platform.

        Uhm... ANY viewer?  ANY platform?  There are a lot of viewers
out there, and a lot of platforms (our company supports our products on
20+ flavors of UNIX, a real fun task).

        Here're the questions I had in mind recently when I was
researching online help solutions for our product:

        1) is the viewer freely distributable?
        2) will it work with our DTP package's native format (frame)?
        3) does it provide a sufficiently robust set of interconnections
                (hyperlinks, whatever you wanna call them) in the text?
        4) is the interface smooth?
        5) is it available on the platforms we need it on?
        6) can we easily tie it into our application?
        7) will it work with a character-based (i.e. vt100) viewer?

        That last one is kind of particular to our market, since a lot
of our customers are still working with old-fashioned vt100 terminals,
but it's probably still a valid concern for a lot of people out there.

        The issue of a distributable viewer is a big one - if each
user has to pay $300-500 a pop then there's no way we (or they) can
afford to distribute it with our package.  Compatibility is also a big
deal - if we have to triple our work to support a second format, we're
not going to be happy.  Platform support is crucial - if we can't put
it on all of the platforms our product goes on, we might as well not
have it - but UNIX platform differences aren't as severe as MAC to DOS
or Windows differences.

        The robustness of the links, the interface, how well it
integrates with our software, they're all considerations, but getting
a product that passes all the other tests is so rare that we really
don't often get a chance to take these into account (as long as
they're basically usable).

Steven J. Owens
uso01 -at- unidata -dot- com