Re: Forth and Co - The Return of the Jedi

foxchip wrote:
On May 4, 5:55 pm, Jerry Avins <j...@xxxxxxxx> wrote:
If you want to drive on the wrong side of the road, you should build a
private road. You are then free to invite and encourage others to join
you on it. Commuters on ordinary roads need to know what to expect.

Yeah. If you go to driving school, invest a lot of time and money
in a track car you can drive with expert drivers on a private road.
If anyone doesn't drive like an expert they get black flagged
immediately because everyone's goal being to be as good as
possible a certain level of expertise close to expert is required.

Commuters need to realize that they are in a sea of drivers
from the best to the worst.

Yes. And the ones who think they are best are often the worst. A lot like programmers.

Commutes need to realize that letting anyone on the road
with only one in every few thousand cars police cars that
are suppose to enforce good (above the absolute minimal
level of bad) driving that half the drivers will be below
average and quite a few will be really really bad and a
menace to society.

Around here, it's the ones who think they are serious drivers who can handle anything that tend to wind up going over the edge into the canyon. That doesn't happen to the little old ladies who simply go slow.

You better take that up with John Doty. He wants to make a Forth that
everyone will use (More power to him!) while you want to keep Forth for
the elite. It would be interesting to read you two slugging it out.

My position is Forth script for the masses and Forth for experts.
Forth script for the masses requires a few Forth experts. Everyone
is happy unless we try to replace Forth with Forth script and
turn the wheel over the worst drivers.

Like those who think they are serious drivers, you have "good" and "bad" interchanged. In my field, programming specialists generally can't cope until somebody gives them working code. Then they can polish (a valuable capability, to be sure, but not the one that opens the door). But you have to use a language that the folks who spend all their time mastering the problem domain can find the time to use also.


I say that if you are driving that means you are responsible for
not crashing. In Forth script we need sandboxes and rubber
lane guides and colision avoidance systems. Those will have
to be written in Forth before we turn the Forth script
steering wheel to the drivers who have no license.

Not a big issue for me. Physicists cope just fine with C, which is pretty unsafe. Besides, programmers' notions of program safety are pretty primitive: name a language that checks dimensional analysis (a common source of bugs, including one that crashed a spacecraft into Mars).


We are bringing it
closer to something that more mediocre programmers can use instead of
making those same programmers understand everything that is needed to
so that they can be fully empowered in the Forth concept.
I stand with John in reaching for that goal.

Most people agree with that goal. Who is going to do it?
The scripting audience? No. It is work for Forth programmers.

Go ahead and try. I'll cheer. But understand that as a programmer you have only part of the skill set required. Also try to understand previous attempts:

STOIC was developed for laboratory scientists by a programmer with a math background. It was moderately successful, although his next attempt at empowering the masses, Lotus 1-2-3, was far more successful.

LSE was developed by a systems engineer. Despite being unpublished and without any commercial distribution (hardly anybody outside MIT building 37 knew of it) nor any commercial hardware that could run it (to use it you had to build a computer from ~50 chips minimum). Nevertheless, it helped do world class science from the late 70's to the late 90's. It represent a systems engineer's notion of simplicity (very different from that of a programmer) and that to me is much of its power.

MAGIC/L was a language developed by a couple of Harvard/MIT physicists. It superficially resembled Pascal, but had a STOIC derived Forth under the hood. It enjoyed moderate success in the 80's.

There are probably more.

If you think of the customers as mediocre, you will surely fail. If you appreciate that they have strengths you can't match, but that you have strengths that might help them, you might succeed.


--
John Doty, Noqsi Aerospace, Ltd.
http://www.noqsi.com/
--
Specialization is for robots.
.

Relevant Pages

  • Re: why add script-tag in header section ?
    ... Of course the event DOM is ready must be considered by the ... Programmers mainly try to serve users. ... It is also worth noting that on the Internet it is common for HTML ... with the placement of SCRIPT elements. ...
    (comp.lang.javascript)
  • Re: Writing my 1st VBS "Script Component"
    ... In script, though, MS has tried to make it very ... they want more money than other programmers. ... but it's supported in Vista. ... software products) are now converging. ...
    (microsoft.public.scripting.vbscript)
  • Re: Standard Ada Preprocessor
    ... cannot count on it being automatically portable to another platform, ... As soon as you introduce new kinds of system dependencies, ... you have to revice the configure script. ... > programmers and mathematicians will be instrested in solving it. ...
    (comp.lang.ada)
  • Re: What will many cores mean to future Windows releases?
    ... most gpf's occur in end user programs, not in drivers. ... Moreover, you don't even need managed code and VM for that, even in Delphi you can turn on range checking so you won't violate ranges when using arrays. ... I think inexperienced programmers should be kept as far away from C++ as possible. ...
    (borland.public.delphi.non-technical)
  • Scripting tools under Windows
    ... Like a lot of you I develop devices that need to be programmed, configured and tested in a production environment, often in conjunction with firmware test functions on the embedded device. ... Years ago, I used QuickBasic to quickly knock up this sort of script, usually 1/2 an hour at most, a nice set of serial and parallel port manipulation tools and easy to run little external apps. ... It should be capable of being modified by Windows programmers, embedded programmers, hardware engineers and production techs without too much of a learning curve - kind of leaves out huge APIs and frameworks. ...
    (comp.arch.embedded)