MY Eyes have been opened

[Fairlight]

This public service announcement was brought to you by Laura Brody:
> On Thu, 09 Feb 2006 11:49:21 -0500, Kenneth Brody <kenbrody at bestweb.net>  
> wrote:
[fPXML]
> > It depends on what you want it to do.  If your needs are simple,
> > you could do it with filePro 1.0 using import/export ascii.
> >
> > I have to disagree with your "1 week project" comment, but you're
> > allowed to think that if you wish.
> 
> 	If you wanted some XML code to do a given job, a week
> would be enough. The thing is that filePro XML has to be a
> generalized tool which helps you to quickly create whatever
> you need no matter what the XML project looks like. This takes
> time to design, code, test, test some more, document and finally
> ship it to the paying customers.

Concur.  OneGate's XML parser was done in 31 hours--granted, I was coming
up to speed from 4.1 to 5.0 features during that whole project, so it took
me longer.  Probably should have been doable in under 10, really.  It's
only about 80 lines of code.  Not that LoC has ever been a truly meaningful
metric--there are instances when one line of usable code a day is good.
But if I hadn't been so rusty and had to learn the new 5.0-isms, it would
have gone a lot faster

That said, I considered writing a generic XML import module over the last
year--several times.  I estimated the hours at about 180 to get it "all
there".  And that's in processing, not C--the latter being far more complex
and demanding.

I agree it would take a lot longer to do a proper, full XML import engine.

Unfortunately, I approached two people that do a lot of XML when I last
considered this project.  One said it would be great for the few people
that specifically need it, and might be worth $1-2K for them.  The other
said they wouldn't pay $500 for such a creature.  In the end, I looked at
it, figured volume would be a problem, just as it has been with every other
add-on, and dropped it when I didn't want to make $2.77/hr, figuring I'd
sell like -maybe- one in the first six months at the $500 price.  It just
seemed pointless.  So, *poof* went the project.  Especially since it'd be
on spec, and I've decided not to do that for this community anymore.

I don't know how many fPXML's fP-Tech will sell.  Nobody does.  But if it's
going to be a year-long project, even assuming only 1/5 of that year is
used to develop it around other priorities, I really think it won't even
pay for its development time within the first year of release.  Probably
not even close.  I hope I'm really wrong, but I have one of those feelings
about this.

That said--it stands a hell of a lot better chance of bolstering filePro
than Biometrics does, IMHO.  It at least makes solid sense.

mark->