Re: The strengths of FORTH

On Mar 28, 1:21 pm, m...@[EMAIL PROTECTED]
 (Marcel Hendrix) wrote:
> Bruce McFarling <agil...@[EMAIL PROTECTED]
> writes Re: The strengths of FORTH
> [..]
>
> > Given Forth's extreme range of implementations, where you can
> > find sets of implementations that are mutually exclusive in
> > terms of what they make of a given source, there is no such
> > thing as a source text that is generically Forth.

> What are these Forths that are so extremely different? Of course
> there are some minority Forths that deliberately try to be off
> course, but they needn't be considered (by their own wish) when
> writing generic Forth.

So when you say, "write generic Forth" it implies a position in
the hierarchy below the absolute extreme most general.

Which is neither unusual or confusing. Only devoted quibblers
would object that, "what you are calling generic Forth does not
run on {determinedly-different-implementation-X}".

> > For a few limited source texts, its possible to say that they
> > are members of the Forth-94 source text society.

> That's not a very useful description. "This is a standard
> Forth that will process standard sources correctly in 5%
> of the cases." I would formulate that differently :-)

What you put in quotes is just about the inverse of what the
description says. "This is a standard Forth that will process
each and everyone one of the miniscule number of available
purely standard sources, plus the much larger number of sources
that require the standard and ...". I don't see anything wrong
with that.

In your formulation, 'a standard Forth' is talking about an
implementation, not about a source text. And "5% of widely
available sources can be used on any standard Forth system
with the available codespace to hold the words" would
represent a big increase in the availability of entirely
generic Forth-94 text.

And its also an increase that there is no particular need to
move heaven and earth to aim for. "This source can be used on
94%+ of Forth-94 installations, including ..." is fine.