Duke Normandin <dukeofperl@[EMAIL PROTECTED]
> wrote:
> Jonah Thomas wrote:
> > Duke Normandin <dukeofperl@[EMAIL PROTECTED]
> wrote:
> > > Yep! That sucks! Would be hard to make a living writing office
> > > productivity software for the Forth community unless you knew
> > > every bloody Forth dialect that exists.
You can't possibly expect to make a living writing Forth for the Forth
community. If you want to make a professional application that people
will pay significant money for, make it for a larger customer base. So
companies that sell Forth compilers don't make their living by selling
Forth compilers, they write applications in Forth for customers, and
then as a sideline they sell their tools. This has the good result that
their compilers are what's worked well for them, and aren't loaded with
ribbons and whistles that marketing thought up to attract customers.
If you want to sell an application, you mostly don't care about ****ting
it to other Forths. You aren't selling it to people who each have a
Forth that they want to run your application on. You're selling it to
people who want an application. If you need to ****t your application to
a new OS or to new hardware, it's likely to turn out easier to ****t your
Forth to that OS or that hardware, and then the application will go
along for the ride. Start with a foreign Forth and you're likely to have
a harder time ****ting. You can spend a lot of effort trying to make your
application ****table, and you'll probably still run into trouble with
subtle incompatibilities between two different Forths. But you
understand your own Forth. So for example the last time I looked, eforth
had 31 routines written in assembly language and 5 system calls. ****t it
to a new system? Get 31 *simple* assembly routines right, and 5 system
calls, and all your applications will ****t. Why bother with ****ting
applications among a dozen weird Forths?
So why do we spend so much effort thinking about ****ting among Forths?
Because if you have your own personal Forth system that's incompatible
with everything else, you get lonely. Most of the people who might be
interested in using your Forth would rather use their own. A lot of
smart lonely people who can use each other's ideas, but to build things
together they need to deal with the compatibility issues.
If there was one genius that everybody respected, who'd said at the
beginning, "This is the Forth. Everybody use it" then possibly they
would have. But it didn't happen. Chuck Moore kept coming up with new
ideas for what his Forth would be like, continually creating
incompatibility with his own previous systems. Follow him and you get
lots of brilliant yet simple ideas to marvel over, but you don't get a
big body of code. Some ways that's a good thing. The more often you
rewrite an application the more likely you make it really good. Leave
out the things that aren't worth having, streamline the im****tant stuff,
get it simple and small and efficient and reliable. There's a limit to
how any applications one genius can produce that way, but it's too soon
to say where that limit is.
> > If you want your computer to run just one application, it isn't such
> > a big deal. A webserver or other. You write your one application in
> > Forth, the ForthOS supplies whatever low-level things you don't want
> > to bother doing yourself, and it has to be a giant commercial
> > success before anybody makes the effort to hack into it.
>
> Again, we're back to one box, one Forth, and apps for that bundle. In
> the desktop market a proprietary OS for a given set of hardware has
> not been as successful as the MSDOS/Windows Linux/BSD approach. Apple
> has finally realized that - to some extent - as well. Now if ForthOS
> was the only Forth around, or at least the only one worth bothering
> with for use with a one-task box or a general-purpose workstation,
> _then_ we'd have something to build a future on, IMHO. Or is there no
> future in the desktop market?
ForthOS has essentially no applications. It can't run applications
written for Windows or Linux or any other OS. There won't be a market
for a desktop running ForthOS until it can do everything that customers
want. It needs a web browser with java, flash, mp3, that can display
realvideo and wmv and avi at the very least. Will customers use a
desktop that can't download ****? Then it needs an office suite, with a
desktop publi****ng system that puts out great brochures and web pages
that can im****t and ex****t all the incompatible versions of Word files,
and a great spreadsheet that can im****t and ex****t with Excel, and a
database that's completely compatible with SQL. It should upload photos
from every major digital camera and arrange them into tasteful albums.
Have I left anything out? Yes, lots. If you can't keep up with all the
applications customers want, they'll stay with something that can.
On the other hand, if you have one application that everybody wants and
no other OS has, then you can hold onto a niche market until somebody
manages to reverse-engineer your work onto Windows. Call it 6 months to
a year.
No, I don't think there's a future for a Forth OS in the desktop market.
The Forth community isn't big enough to make that future. If you want to
do that, you need to build a much larger Forth community first.
If you want to do that, you should probably start with a big bunch of
potential Forth programmers that you have access to. Show them a great
application that they want to use and modify. As they start using it and
modifying it they'll teach themselves how to use your Forth. As they
share code they'll have an incentive to keep using your Forth instead of
changing it around in incompatible ways. If that works well enough,
pretty soon your collection of Forth converts will dwarf the existing
Forth community and Forth will be what you made it.
I'd tend to think there could be an op****tunity there among the gaming
community. A lot of smart teenagers there with time on their hands. They
might prefer a Forth that uses icons instead of typed names. Each icon
gets its own picture and they can have cool pictures. They mostly line
up and each has colored lines representing inputs and colored lines
representing outputs. You can do some datatyping with the colors. Put
cute pictures on the icons. Nobody worries about RPN when they're
looking at pictures with lines going in and out.
I don't know enough about the games to see what app is needed. There
could be an op****tunity there for somebody who knows how to take it.
|
