home .. forth .. colorforth mail list archive ..

Re: [colorforth] New Linux 4word

On Fri, Apr 08, 2005 at 08:48:38PM -0400, maslicke@xxxxxxxxxxx wrote:
>
> ---- Original message ----
> >Date: Fri, 8 Apr 2005 11:57:41 +0200
> >From: albert@xxxxxxxxxxxxxxxxxx (Albert van der Horst)
> >
<SNIP>
> If you can manage to get colorForth to boot from virtually
> any PC, more power to you. For the many complaints of
> colorForth not working, I have seen so few reports on what
> hardware colorForth has failed on, which version was tried,
> and exact the behavior on boot and in response to 'save'.
> Given this we might see a pattern emerge and might be able
> to diagnose the problem, and even have people who are in
> possesion of the hardware fix the problem.

There is no attempt from Chucks side to systematically
develop colorforth such that it is portable.
This is a job left to the reader. We have had little success
to date, but I admit it is hard.
It starts from an analysis what is guaranteed to work on a
PC. A boot floppy from my DOS 3.2 8086 Toshiba still boots
my Celeron laptop.
By the way, I have done a booting Forth too, and the reports from what
hardware it failed on and what not, were not of the slightest help.
(This is a booting Forth that runs in protected mode.)

> As for an "official glossary", I am not sure what you mean.
> What would make it "official"? Quite early on a glossary was
> published[1], which anyone is free to contribute to.

It is official if it is recognized by the majority of color
forth users as having documented the functionality of some
identified Forth version that can be used as a starting point
for further development. It is a fixed point from which
changes are documented. Even if that is Howerd's version and
you have got a newer colorforth from Chuck the recognision
makes it official. You will be asked what the difference are
and you can give a meaningful answer.

> I don't know what is meant by a "test suite", or how that
> would be a signifigant "milestone".

Do you really not know what a test suite is?
A test suite -- translated to our context -- is a comprehensive set of
tests for each word, that allows to decide whether and how it
deviates from a fixed functionality, in this case probably the above
glossary. This greatly helps to check whether after a modification
you have not destroyed functionality, or whether the changes are
like you intended. It can be used as a regression test.

I don't expect anybody to be willing to work on this, but
I think it is silly to renounce it usefulness.
(I will not comment on renouncing the usefulness of something
you don't know.)

<SNIP>
> >An assembler is the proper starting point. There is no
> Metacompiler
> >without assembler. And an assembler may be enough.
> >Unless of course you think you have a nice source if it is
> sprinkled
> >with
> >        cd45 2,      12 1, 3135 2,
> >
>
> What this sugests to me is that you have not seriously
> studied colorForth, or if you have, you reject its
> principles. As Chuck states clearly[2]:

Maybe I should have said: there is no Metacompiler without
assembler (unless you are willing to use machinecode directly.)

If you think `` cd45 2,'' is nice, you are beyond hope.
Do you really think Chuck likes this, or that it is considered the
god-given fundament of colorforth? As Jeff has pointed out,
Chuck has always used an assembler in this circumstance.
Now he deviates for good reason, and it is related to the Pentium.
Chucks analysis is : "machine code, because the alternative
is worse".

But IMO my assembler tips the balance.

<SNIP>

>
> The key word here is 'factor', if you write 13 blocks of
> code and then force these 13 blocks onto colorForth after
> the fact, clearly what you have done has nothing to do with
> the word 'factor' as Forth programmers understand it.

If you don't understand that a full Pentium assembler in 13 screens
must be a marvellous feat of factoring, even without looking at the
source, you are not much of a Forth programmer (or you have no idea
what the Pentium is like).

Should an assembler  of the postit fixup principle
be accepted  for colorforth, that is a far cry from 'forcing'
upon. It would be a combination of reimplementing the postit-fixup
principle in colorforth and copying some of the tables
for the opcodes. Probably not those of the floating point.
Then the names must be revised, because it uses characters
not present in colorforth.

I have the feeling that the whole idea of assembling
in colorforth is new, and could require a new color.
This may be wrong.

>
> Mark
>
Groetjes Albert

--
Albert van der Horst,Oranjestr 8,3511 RA UTRECHT,THE NETHERLANDS
Economic growth -- like all pyramid schemes -- ultimately falters.
albert@xxxxxxxxxxxxxxxxxx http://home.hccnet.nl/a.w.m.van.der.horst

---------------------------------------------------------------------
To unsubscribe, e-mail: colorforth-unsubscribe@xxxxxxxxxxxxxxxxxx
For additional commands, e-mail: colorforth-help@xxxxxxxxxxxxxxxxxx
Main web page - http://www.colorforth.com