GUI Front Ends - the discussion

[Fairlight]

Confusious (Richard Hane) say:
> The question to be answered: What to use to create the GUI front end?
> 
> 1. Do we use another database program like MySQL, Miscrosft SQL Server,
> Access, Oracle, etc.? If so, which one and why?  

Well, there are GUI RAD toolkits galore for the *SQL engines.  Kylix comes
to mind, and Oracle has their own enterprise stuff.  Access is going to
bind you to VB.  Either way, you'd be learning something new if you go with
a desktop app against *SQL.

If you're going SQL and want browser based, evaluate CodeCharge
(www.codecharge.com).  An fP developer clued me to it, and it impressed me
as well as another fP developer I referred to it.

> 2. Or rather do we use a browser on an intranet initially and maybe
> port to the web later?  

What porting?  If you design properly from the start, under OneGate it's as
easy as altering the application firewall rules to allow outside access to
something that was previously restricted.  You can do this by task, so you
can allow as little or much of it to be accessible as you choose.  If you
meant design only so much initially and add public facilities as needed,
then ok, I follow.

If odds are good you'll want a web based interface eventually, it makes
sense to do the project once as such, giving the best ROI and least
redundant development.

> 3. If we use the browser what do we use; FPCGI, fpWeb, One Gate, etc?

Not fPCGI.

As I know it from secondhand reports and a peek at some docs that a client
showed me, fPWeb is more technique than product.  The screen scrape
approach is doable under any gateway; in fact, one of my OneGate users
gave me a screen scraper that he wrote for use with OG, to do what I wish
with.  I should explore offering that, probably.  But what I remember of
fPWeb code on installed sites I audited for security scared the heck out of
me.  The docs at the time had chmod 777 instructions--not a sign of good
security.  A few who used it told me OG could handle the same techniques
with a few tweaks, but more securely with more features.

I'm of course biased in favour of OG in general, but even were I not, it's
the most flexible and secure.  It enforces security, and it's otherwise as
flexible as your imagination or education will let you be.

> 4. Or is fpGUI a possibility?  What ever happened to fpGUI?  I own one of
> the first copies but you never hear anyone discussing it.

fPGI?  Almost everyone I talk with in private about it laughs with derision
and dismay at what is being fobbed off as a GUI in that case.  Nobody
probably discusses it publicly because it's more or less irrelevant
to embarrassing--the bastard child of the franchise, even to many fP
loyalists.  It's a, "Not much to see, move along," sort of thing, IMHO.

You're -far- better off looking at Reface, by Sound Ideas.  I'm surprised
nobody has mentioned this alternative yet. (www.soundideas.com) The
screenshots and feature list look awesome, and I hear nothing but good
things about it from folks I trust.  Ahh...but I think it requires a *nix
server, which might explain the lack of mention with you currently on
Windows.  But you said *nix wasn't necessarily excluded.  Might look into
this--could be worth migrating for--but you'd lose fP ODBC potential.

> We are not concerned with output at this time as we have decided to use
> (and we own) Print Wizard. Nice program Bob.

I should mention that both PW and WePO play nicely with OneGate.  A client
uses WePO, I've personally hooked PW into an app to do PDF printing in the
browser from live data on the fly.  Awesome combos.

> OK, so there is our delema.  Which way to go.  We really don't want to
> start down one path and the later turn around and reinvent the wheel.

You need to look at cost:benefit and ROI.  Designing for the web is not
hard, but it is honestly time consuming because (with only low level tools)
you're handling every handshake along a data flow, rather than dealing with
one cohesive, stateful program instance.  Not hard, just a bit tedious and
it takes some getting used to.

Reface would let you stick with fP and give great results.  Not sure on the
learning curve.  Cost is kinda seemingly high, but you're writing off a lot
of web development and coming out ahead, probably, if all you use is custom
code on low level tools.  If STN can finish fPagePro, I'd evaluate between
the two.

Either way, you need to figure the probability of wanting web-based, and
what percentage would be redundant development if you go that way later.
Then you can assess if you start out that way, or go with something like
Reface and just add minor web bits in parallel later.  They're far from
mutually exclusive, but you want to avoid redundant work as much as
possible to max your ROI.

No hard and fast answer, really.  A lot depends on what looks attractive
aesthetically and financially, your skill sets, and your short and long
term direction and goals.

mark->
-- 
"Moral cowardice will surely be written as the cause on the death
certificate of what used to be Western Civilization." --James P. Hogan