Bruce McFarling <agila61@[EMAIL PROTECTED]
> wrote:
> Jonah Thomas <jethom...@[EMAIL PROTECTED]
> wrote:
> > You might get a foothold if you have an application that could use a
> > dedicated platform similar to a PC. If the PC is cheaper than
> > dedicated hardware, then use it even though it does more than you
> > want. ...
>
> But then, where is the network economy, because if the PC is cheaper
> than dedicated hardware, it can more often be done by stringing
> together existing libraries plus a little glue code. And it may end up
> being overkill in terms of the computing resources devoted to the
> task, but *so what*? If the PC is cheaper than the dedicated hardware
> that can *just* do the task, then the extra resources to do it by
> stringing together libraries written for the PC resources are *free*.
Except you have to adapt them to do what you want. And you get whatever
overhead that comes with these foreign libraries and OS etc.
So for it to work as I suggest, you need something that can be done
adequately with PC hardware, but that works *sufficiently better* with
your own OS and application that there's an advantage to using them.
This makes the reasonable assumption that it will cost you more effort
to do it yourself than to hook into foreign libraries etc on top of an
existing OS. But after you do that once, then it's a sunk cost and you
get it cheap for the second application.
> The foothold for this strategy is if you have an application that
> could use a dedicated platform that is *massively underpowered*
> compared to a PC, and *massively cheaper* than a PC. An order of
> magnitude cheaper. Since the cost of a PC is in $100's, then an order
> of magnitude means in the $10's. And then the "PC" ****t is simply
> emulating that cheap piece of hardware on the PC, typically with
> upwardly compatible expansions of resources (RAM, display space,
> etc.), and running the same apps in both the PC and the cheap piece of
> hardware.
That's where Forth already has a foothold. But I haven't heard that
there's any big economy of scale there, unless you already know you have
a product that will sell in large numbers. If you develop an "OS" for a
particular board with a particular processor etc, aren't there thousands
of other boards that do somewhat similar things or very different things
in the same price range? You can use that one until they stop making it
but there isn't anything to get momentum with other people. For that
sort of thing you'd probably do better with something like SwiftX. Write
SwiftX code that ****ts to whichever boards and processors. Ask Forth,
Inc to write a compiler for the next system you want to use.
> On this:
> > Again going by the utterly unreliable TIOBE numbers, if you write
> > other langauges in Forth and ****t the Forth, you could theoretically
> > ****t a lot of apps easily.
>
> > Java 20%
> > + C 35%
> > + C++ 45%
> > + PHP 55%
> > + Perl 61%
> > + Python 66%
>
> Again, there's no network economy available to any of those languages
> in working with Forth. They get the theoretical op****tunity to more
> quickly ****t their language base to marginal or not-yet-developed
> hardware, assuming the sup****t of an active developer base around the
> "Java-in-Forth" implementation that does not in fact exist, while in
> practice, due to the size of their language community, they already
> get ****ted reasonably quickly to most of the fringe or not-yet-
> developed hardware that is of interest to them for most of their code
> base.
Oh. Too bad. But if you have a Forth OS and you implement the popular
languages for it, then you get to im****t apps. Chances are the other
language communities won't bother. And if you can ****t your Forth and
piggyback languages to other marginal OSes then they get that advantage
too. It could make the difference between a marginal OS having a chance
versus having no chance.
But it doesn't look like something you could write a great business plan
for.
> Better to target a language with a smaller language community, where
> there is more likely to be an actual op****tunity to have an actual
> network economy from having a "Language-in-Forth" implementation. I
> have long used Lua as an example ... another example would be Cobra:
> http://cobra-language.com/
Same problems, though. If the language doesn't have much of a code base,
what good does it do to ****t it? And how much good does it do them to
****t their language to marginal systems? Actually I guess you could get
a little momentum that way. But don't the big wins usually come from
hooking up with winners?
|
