Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.forth > #57613 > unrolled thread

High level language for writing Forth

Back to article view | Back to comp.lang.forth

Contents

  High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-19 18:32 -0500
    Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-20 00:24 +0000
      Re: High level language for writing Forth "Elizabeth D. Rather" <erather@forth.com> - 2017-04-19 14:42 -1000
        Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-19 20:33 -0500
          Re: High level language for writing Forth "Elizabeth D. Rather" <erather@forth.com> - 2017-04-19 16:21 -1000
          Re: High level language for writing Forth Paul Rubin <no.email@nospam.invalid> - 2017-04-19 19:41 -0700
            Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-19 23:28 -0500
        Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-20 17:02 +0000
          Re: High level language for writing Forth "Elizabeth D. Rather" <erather@forth.com> - 2017-04-20 09:48 -1000
            Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-21 09:42 +0000
              Re: High level language for writing Forth Andrew Haley <andrew29@littlepinkcloud.invalid> - 2017-04-21 09:07 -0500
                Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-21 17:04 +0100
                  Re: High level language for writing Forth Andrew Haley <andrew29@littlepinkcloud.invalid> - 2017-04-21 11:17 -0500
                    Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-21 17:36 +0100
                    Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-21 20:32 +0000
                      Re: High level language for writing Forth Andrew Haley <andrew29@littlepinkcloud.invalid> - 2017-04-21 15:35 -0500
                        Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-21 20:39 +0000
                      Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-21 22:47 +0100
                  Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-21 20:29 +0000
                    Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-21 22:45 +0100
                      Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-22 04:49 +0000
                        Re: High level language for writing Forth Jan Coombs <jenfhaomndgfwutc@murmic.plus.com> - 2017-04-22 10:54 +0100
                          Re: High level language for writing Forth Mark Wills <markwills1970@gmail.com> - 2017-04-22 03:04 -0700
                            Re: High level language for writing Forth Andrew Haley <andrew29@littlepinkcloud.invalid> - 2017-04-22 05:12 -0500
                              Re: High level language for writing Forth lehs <skydda.barnen@gmail.com> - 2017-04-22 04:30 -0700
                                Re: High level language for writing Forth Mark Wills <markwills1970@gmail.com> - 2017-04-22 05:00 -0700
                                  Re: High level language for writing Forth lehs <skydda.barnen@gmail.com> - 2017-04-22 05:40 -0700
                                    Re: High level language for writing [unum] Forth Jan Coombs <jenfhaomndgfwutc@murmic.plus.com> - 2017-04-22 16:51 +0100
                                      Re: High level language for writing [unum] Forth Paul Rubin <no.email@nospam.invalid> - 2017-04-22 13:36 -0700
                                        Re: High level language for writing [unum] Forth Jan Coombs <jenfhaomndgfwutc@murmic.plus.com> - 2017-04-22 22:59 +0100
                                          Re: High level language for writing [unum] Forth Paul Rubin <no.email@nospam.invalid> - 2017-04-22 18:43 -0700
                                            Re: High level language for writing [unum] Forth Jan Coombs <jenfhaomndgfwutc@murmic.plus.com> - 2017-04-23 09:33 +0100
                                    Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-22 11:23 -0500
                              Re: High level language for writing Forth Mark Wills <markwills1970@gmail.com> - 2017-04-22 04:59 -0700
                          Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-29 03:05 +0000
                            Re: High level language for writing Forth Jan Coombs <jenfhaomndgfwutc@murmic.plus.com> - 2017-04-29 09:35 +0100
                        Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-22 12:56 +0100
                          Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-29 03:08 +0000
                            Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-29 09:46 +0100
                              Re: High level language for writing Forth lehs <skydda.barnen@gmail.com> - 2017-04-29 02:45 -0700
                                Re: High level language for writing Forth anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2017-04-29 13:44 +0000
                                  Re: High level language for writing Forth lehs <skydda.barnen@gmail.com> - 2017-04-29 07:00 -0700
                                  Re: High level language for writing Forth hughaguilar96@gmail.com - 2017-04-29 07:31 -0700
                                    Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-29 16:38 +0100
                            Re: High level language for writing Forth Joel Rees <joel.rees@gmail.com> - 2017-04-29 04:10 -0700
                              Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-29 11:35 +0000
                                Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-29 16:33 +0100
                                  Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-05-05 15:24 +0000
                                    Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-05-05 20:16 +0100
                                      Re: High level language for writing Forth Lars Brinkhoff <lars.spam@nocrew.org> - 2017-05-05 22:06 +0200
                                        Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-05-05 21:45 +0100
                                          Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-05-05 21:54 +0100
                                      Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-05-06 10:06 +0000
                                        Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-05-06 13:09 +0100
                                      Re: High level language for writing Forth anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2017-05-06 11:04 +0000
                Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-21 20:28 +0000
                  Re: High level language for writing Forth "Elizabeth D. Rather" <erather@forth.com> - 2017-04-21 12:27 -1000
                    Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-22 04:55 +0000
                  Re: High level language for writing Forth Coos Haak <htrofhc@gmail.com> - 2017-04-22 02:01 +0200
                    Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-22 04:58 +0000
                      Re: High level language for writing Forth Paul Rubin <no.email@nospam.invalid> - 2017-04-21 23:41 -0700
                        Re: High level language for writing Forth Coos Haak <htrofhc@gmail.com> - 2017-04-22 22:15 +0200
                          Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-05-06 13:28 +0000
                      Re: High level language for writing Forth Joel Rees <joel.rees@gmail.com> - 2017-04-27 02:51 -0700
                        Re: High level language for writing Forth Lars Brinkhoff <lars.spam@nocrew.org> - 2017-04-27 12:09 +0200
                          Re: High level language for writing Forth Joel Rees <joel.rees@gmail.com> - 2017-04-29 02:48 -0700
                        Re: High level language for writing Forth "Rudy Velthuis" <newsgroups@rvelthuis.de> - 2017-04-29 03:09 +0000
      Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-19 20:11 -0500
    Re: High level language for writing Forth Paul Rubin <no.email@nospam.invalid> - 2017-04-19 19:34 -0700
      Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-19 23:22 -0500
        Re: High level language for writing Forth Paul Rubin <no.email@nospam.invalid> - 2017-04-19 21:35 -0700
          Re: High level language for writing Forth Melzzzzz <Melzzzzz@zzzzz.com> - 2017-04-20 05:12 +0000
          Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-20 01:15 -0500
            Re: High level language for writing Forth Paul Rubin <no.email@nospam.invalid> - 2017-04-19 23:25 -0700
            Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-21 23:12 -0500
              Re: High level language for writing Forth hughaguilar96@gmail.com - 2017-04-21 21:53 -0700
                Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-22 00:42 -0500
                  Re: High level language for writing Forth Paul Rubin <no.email@nospam.invalid> - 2017-04-21 23:44 -0700
                    Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-22 03:08 -0500
                  Re: High level language for writing Forth Andrew Haley <andrew29@littlepinkcloud.invalid> - 2017-04-22 04:39 -0500
                    Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-22 12:37 +0100
                    Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-22 11:33 -0500
                    Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-22 14:10 -0500
                  Re: High level language for writing Forth hughaguilar96@gmail.com - 2017-04-22 15:01 -0700
                    Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-22 23:38 +0100
                      Re: High level language for writing Forth Ilya Tarasov <ilya74.tarasov@gmail.com> - 2017-04-22 16:30 -0700
                        Re: High level language for writing Forth hughaguilar96@gmail.com - 2017-04-22 18:01 -0700
                        Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-23 09:25 +0100
                          Re: High level language for writing Forth Ilya Tarasov <ilya74.tarasov@gmail.com> - 2017-04-23 06:32 -0700
                            Re: High level language for writing Forth Julian Fondren <julian.fondren@gmail.com> - 2017-04-23 07:36 -0700
                            Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-24 14:32 +0100
                    Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-22 18:19 -0500
                      Re: High level language for writing Forth hughaguilar96@gmail.com - 2017-04-22 18:05 -0700
                        Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-22 21:51 -0500
                    Re: High level language for writing Forth Ron Aaron <rambamist@gmail.com> - 2017-04-23 06:54 +0300
                      Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-23 09:16 +0100
                        Re: High level language for writing Forth Ron Aaron <rambamist@gmail.com> - 2017-04-23 18:41 +0300
                          Re: High level language for writing Forth Alex <alex@rivadpm.com> - 2017-04-24 14:35 +0100
                      Re: High level language for writing Forth Ilya Tarasov <ilya74.tarasov@gmail.com> - 2017-04-23 06:15 -0700
                Re: High level language for writing Forth Ilya Tarasov <ilya74.tarasov@gmail.com> - 2017-04-22 08:04 -0700
                  Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-22 11:44 -0500
                    Re: High level language for writing Forth Ilya Tarasov <ilya74.tarasov@gmail.com> - 2017-04-22 12:16 -0700
                Re: High level language for writing Forth Rod Pemberton <NeedNotReplyHere@xrsevnneqk.cem> - 2017-04-22 22:52 -0400
        Re: High level language for writing Forth Ron Aaron <rambamist@gmail.com> - 2017-04-20 07:49 +0300
      Re: High level language for writing Forth albert@cherry.spenarnc.xs4all.nl (Albert van der Horst) - 2017-04-20 19:47 +0200
        Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-20 15:19 -0500
    Re: High level language for writing Forth Lars Brinkhoff <lars.spam@nocrew.org> - 2017-04-20 08:06 +0200
    Re: High level language for writing Forth Ilya Tarasov <ilya74.tarasov@gmail.com> - 2017-04-20 04:35 -0700
    Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-04-25 17:03 -0500
      Re: High level language for writing Forth minforth@arcor.de - 2017-04-26 01:58 -0700
    Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-05-13 17:36 -0500
      Re: High level language for writing Forth hughaguilar96@gmail.com - 2017-05-13 18:37 -0700
        Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-05-13 20:52 -0500
          Re: High level language for writing Forth hughaguilar96@gmail.com - 2017-05-13 19:25 -0700
            Re: High level language for writing Forth hughaguilar96@gmail.com - 2017-05-15 12:44 -0700
              Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-05-15 17:27 -0500
                Re: High level language for writing Forth hughaguilar96@gmail.com - 2017-05-15 19:43 -0700
                  Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-05-15 23:55 -0500
                    Re: High level language for writing Forth Mark Wills <markwills1970@gmail.com> - 2017-05-16 01:59 -0700
                  Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-05-15 23:58 -0500
      Re: High level language for writing Forth Joel Rees <joel.rees@gmail.com> - 2017-05-15 07:00 -0700
        Re: High level language for writing Forth Cecil Bayona <cbayona@cbayona.com> - 2017-05-15 12:05 -0500

Page 2 of 7 — ← Prev page 1 [2] 3 4 5 6 7  Next page →

#57774

Alex wrote:

> > So you can inline the assembly. That's fine. It is still (x-86, in
> > this case) assembly. <shrug>
> > 
> 
> It's the other way round. The assembler is inlining Forth.

Not sure if you are serious, but that is nonsense anyway. Something
that uses assembler mnemonics to generate machine code is an assembler,
no matter if it is written in Forth or not. Oh, it is a built-in (or
loadable) assembler, but these are not unique to Forth either.

-- 
Rudy Velthuis        http://www.rvelthuis.de

"I was playing poker the other night... with Tarot cards.
 I got a full house and 4 people died." -- Steven Wright

[toc] | [prev] | [next] | [standalone]

#57803

On Sat, 22 Apr 2017 04:49:19 +0000 (UTC)
"Rudy Velthuis" <newsgroups@rvelthuis.de> wrote:

> Alex wrote:
> 
> > It's the other way round. The assembler is inlining Forth.
> 
> Not sure if you are serious, but that is nonsense anyway.
> Something that uses assembler mnemonics to generate machine
> code is an assembler, no matter if it is written in Forth or
> not. Oh, it is a built-in (or loadable) assembler, but these
> are not unique to Forth either.

I think that if you asked a number of assembler programmers to
use the simplest forth assemblers that I have used and written,
then the average response would be:

  "That is _not_ [what I call] an assembler"

The rejection is because to a first approximation does not
support the working environment that they quite reasonably
expect. 

However, the Forth assembler extension can be used to
interactively develop fragments of machine code by using the rest
of a Forth system as support tools. 

Jan Coombs
-- 

[toc] | [prev] | [next] | [standalone]

#57804

On Saturday, 22 April 2017 10:54:25 UTC+1, Jan Coombs  wrote:
> On Sat, 22 Apr 2017 04:49:19 +0000 (UTC)
> "Rudy Velthuis" <newsgroups@rvelthuis.de> wrote:
> 
> > Alex wrote:
> > 
> > > It's the other way round. The assembler is inlining Forth.
> > 
> > Not sure if you are serious, but that is nonsense anyway.
> > Something that uses assembler mnemonics to generate machine
> > code is an assembler, no matter if it is written in Forth or
> > not. Oh, it is a built-in (or loadable) assembler, but these
> > are not unique to Forth either.
> 
> I think that if you asked a number of assembler programmers to
> use the simplest forth assemblers that I have used and written,
> then the average response would be:
> 
>   "That is _not_ [what I call] an assembler"
> 
> The rejection is because to a first approximation does not
> support the working environment that they quite reasonably
> expect. 
> 
> However, the Forth assembler extension can be used to
> interactively develop fragments of machine code by using the rest
> of a Forth system as support tools. 
> 
> Jan Coombs
> --

They may object, but after an hour of coding without having to use a
single label anywhere in their code, they normally come around, just
as I did :-)

[toc] | [prev] | [next] | [standalone]

#57805

Mark Wills <markwills1970@gmail.com> wrote:

> They may object, but after an hour of coding without having to use a
> single label anywhere in their code, they normally come around, just
> as I did :-)

Indeed so.  It's still baffling to me that assembly programs usually
don't use structured programming constructs such as loops and if
statements.  It seems to me that they're just as useful in assembly-
language programs as in high-level languages.

Andrew.

[toc] | [prev] | [next] | [standalone]

#57809

Den lördag 22 april 2017 kl. 12:12:55 UTC+2 skrev Andrew Haley:
> Mark Wills <markwills1970@gmail.com> wrote:
> 
> > They may object, but after an hour of coding without having to use a
> > single label anywhere in their code, they normally come around, just
> > as I did :-)
> 
> Indeed so.  It's still baffling to me that assembly programs usually
> don't use structured programming constructs such as loops and if
> statements.  It seems to me that they're just as useful in assembly-
> language programs as in high-level languages.
> 
> Andrew.

I which there were some pseudo machine code language, not optimized to build up Forth, but optimized to be easy to be built up itselves from most machine codes in use. Of course such an intermediate stage would be less than optimal, but maybe more efficient than c and go.

Then even I would try to learn that language and try to built up a Lehs-Forth.

[toc] | [prev] | [next] | [standalone]

#57815

On Saturday, 22 April 2017 12:30:41 UTC+1, lehs  wrote:
> Den lördag 22 april 2017 kl. 12:12:55 UTC+2 skrev Andrew Haley:
> > Mark Wills <markwills1970@gmail.com> wrote:
> > 
> > > They may object, but after an hour of coding without having to use a
> > > single label anywhere in their code, they normally come around, just
> > > as I did :-)
> > 
> > Indeed so.  It's still baffling to me that assembly programs usually
> > don't use structured programming constructs such as loops and if
> > statements.  It seems to me that they're just as useful in assembly-
> > language programs as in high-level languages.
> > 
> > Andrew.
> 
> I which there were some pseudo machine code language, not optimized to build up Forth, but optimized to be easy to be built up itselves from most machine codes in use. Of course such an intermediate stage would be less than optimal, but maybe more efficient than c and go.
> 
> Then even I would try to learn that language and try to built up a Lehs-Forth.

I think the thing you just wished for is... Forth. That's what it is!

[toc] | [prev] | [next] | [standalone]

#57816

Den lördag 22 april 2017 kl. 14:00:13 UTC+2 skrev Mark Wills:
> On Saturday, 22 April 2017 12:30:41 UTC+1, lehs  wrote:
> > Den lördag 22 april 2017 kl. 12:12:55 UTC+2 skrev Andrew Haley:
> > > Mark Wills <markwills1970@gmail.com> wrote:
> > > 
> > > > They may object, but after an hour of coding without having to use a
> > > > single label anywhere in their code, they normally come around, just
> > > > as I did :-)
> > > 
> > > Indeed so.  It's still baffling to me that assembly programs usually
> > > don't use structured programming constructs such as loops and if
> > > statements.  It seems to me that they're just as useful in assembly-
> > > language programs as in high-level languages.
> > > 
> > > Andrew.
> > 
> > I which there were some pseudo machine code language, not optimized to build up Forth, but optimized to be easy to be built up itselves from most machine codes in use. Of course such an intermediate stage would be less than optimal, but maybe more efficient than c and go.
> > 
> > Then even I would try to learn that language and try to built up a Lehs-Forth.
> 
> I think the thing you just wished for is... Forth. That's what it is!

If so, then Forth would be the perfect platform for building an operating system.

[toc] | [prev] | [next] | [standalone]

#57824 — Re: High level language for writing [unum] Forth

On Sat, 22 Apr 2017 05:40:43 -0700 (PDT)
lehs <skydda.barnen@gmail.com> wrote:
> Den lördag 22 april 2017 kl. 14:00:13 UTC+2 skrev Mark Wills:
> > On Saturday, 22 April 2017 12:30:41 UTC+1, lehs  wrote:
> > > 
> > > I [wish] there were some pseudo machine code language, not
> > > optimized to build up Forth, but optimized to be easy to
> > > be built up itselves from most machine codes in use. Of
> > > course such an intermediate stage would be less than
> > > optimal, but maybe more efficient than c and go.
> > > 
> > > Then even I would try to learn that language and try to
> > > built up a Lehs-Forth.
> > 
> > I think the thing you just wished for is... Forth. That's
> > what it is!
> 
> If so, then Forth would be the perfect platform for building
> an operating system. 

This would require starting from scratch, so would only happen
in a big way if something else forced this, for example: 


The Unum[1] data type is a superset of integer, float, and [IMO] 
BigNum which uses a compact variable length representation. It
is incompatible with current float hardware, and using it may
suggest a change to software algorithms which currently use
arrays, and suggests a different approach to handling complex
mathematical analysis. 

Using Unums results in absolutely correct results from float
computation. A hardware implementation should reduce power
consumption by 50% due to reduced memory traffic[2]. 

Handling data items of variable length is only efficient while
they are stored without padding. In stack frames or other
compiler allocated spaces this becomes complex and/or
inefficient.  

Because a Forth engine keeps data local to computation on the
stack, this can be stored efficiently with no gaps or need
for pre-allocated memory. 

So, ISTM that the simplest implementation of Unum processor is
as a dual stack processor, and the obvious first choice for
low level language is fORTH. 


Jan Coombs
-- 

[1] https://en.wikipedia.org/wiki/Unum_(number_format)
[2] http://insidehpc.com/2015/03/slidecast-john-gustafson-explains-energy-efficient-unum-computing/

Much more in book:
http://www.johngustafson.net/unums.html

Not to be confused with Unum v2:
http://ubiquity.acm.org/article.cfm?id=3001758

[toc] | [prev] | [next] | [standalone]

#57846 — Re: High level language for writing [unum] Forth

Jan Coombs <jenfhaomndgfwutc@murmic.plus.com> writes:
> The Unum[1] data type is a superset of integer, float, and [IMO] 
> BigNum which uses a compact variable length representation.

You could also include pointers in the Unum type, that would themselves
point to these arbitrary-sized objects, that could then be nested to
any depth.  Congratulions, you have invented Lisp :).

[toc] | [prev] | [next] | [standalone]

#57865 — Re: High level language for writing [unum] Forth

On Sat, 22 Apr 2017 13:36:54 -0700
Paul Rubin <no.email@nospam.invalid> wrote:

> Jan Coombs <jenfhaomndgfwutc@murmic.plus.com> writes:
> > The Unum[1] data type is a superset of integer, float, and
> > [IMO] BigNum which uses a compact variable length
> > representation.
> 
> You could also include pointers in the Unum type, that would
> themselves point to these arbitrary-sized objects, that could
> then be nested to any depth.  Congratulions, you have invented
> Lisp :).

Yes, I would reserve a Unum representation for use as an
object pointer, perhaps an invalid exponent with the
fraction field containing the object address. 

I'm uncertain of the value of indefinitely successive in-
direction, and showers of parentheses aggravate my dyslexia. 

Python could run on a Unum stack engine with improved
performance.  Three of the four numeric data types would merge:
plain integers, long integers, and floating point. The data
would be in the stack, and not require the two indirections as
used in the CPython virtual stack engine. 

Who is there bold enough, willing to step outside the cage, and
be worthy of congratulions?

Jan Coombs
-- 

[toc] | [prev] | [next] | [standalone]

#57874 — Re: High level language for writing [unum] Forth

Jan Coombs <jenfhaomndgfwutc@murmic.plus.com> writes:
> Who is there bold enough, willing to step outside the cage, and
> be worthy of congratulions?

John McCarthy?  Again, you're reinventing Lisp from 60 years ago.

[toc] | [prev] | [next] | [standalone]

#57891 — Re: High level language for writing [unum] Forth

On Sat, 22 Apr 2017 18:43:26 -0700
Paul Rubin <no.email@nospam.invalid> wrote:

> Jan Coombs <jenfhaomndgfwutc@murmic.plus.com> writes:
> > Who is there bold enough, willing to step outside the cage,
> > and be worthy of congratulions?
> 
> John McCarthy?  Again, you're reinventing Lisp from 60 years
> ago.

Much as I'd like to agree with you, this is unlikely. My focus
is on hardware development. This is because the Unum data type
is incompatible with current float hardware, (see below) so
efficient modelling of Unums is not possible, and adding a
Unum library to any language nearly pointless. 

As for Forth, it would be simplified:  No need for a separate
float stack, automatic promotion of integers to exact floats[1]
or BigNum when arithmetic demands it. Double numbers redundant. 
Significant compression of Flags and small integers. 


The IEE754 float has the leftmost bit of the mantissa hidden,
and the rest in a fixed length field. In this field the bits
toward the left contain data, and the rest junk. 

In a Unum[2] mantissa field, all bits contain valid data. If
padding is needed to match field allocation granularity it must
be a sign bit extension, at the left end. 

So, the mantissa of an IEE754 float is left aligned, and that of
a Unum is right aligned. This is why the Unum data type cannot
be efficiently modelled in current processor hardware. 

This difference is what allows a Unum processor to produce
correct results[1] and use less power.  Correct because no junk
is stored in the mantissa. An anticipated 50% power saving
because of not having to move this junk to and from main
memory.[3]


Is this really an opportunity to support simpler programming
languages with greater computational efficiency, or a magical
fairy tale from the dark side, or perhaps both?

Jan Coombs
--   

[1] The Unum mantissa is extended one bit past the ULP, this
'Uncertainty bit' declares whether the mantissa is exact, or is
the base of an open interval one ULP wide. Unum computation
results are correct, even where this requires declaring the
bounds of uncertainty to achieve this. (Unlike IEE754)

[2] That is Unum type 1.  In Unum type 2 reduced benefits
are anticipated because the binary format used is of fixed
length.

[3] "The End of Error" pg6: Energy cost of 64b float read from
reg and MAC operation is 70pJ, the cost of fetching 64b from
DRAM is 4200pJ, 60 times more! (on 2010 era hardware)  Unum data
is inherently compressed. (at the cost of on-chip complexity)

[toc] | [prev] | [next] | [standalone]

#57828

On 4/22/2017 7:40 AM, lehs wrote:
> Den lördag 22 april 2017 kl. 14:00:13 UTC+2 skrev Mark Wills:
>> On Saturday, 22 April 2017 12:30:41 UTC+1, lehs  wrote:
>>> Den lördag 22 april 2017 kl. 12:12:55 UTC+2 skrev Andrew Haley:
>>>> Mark Wills <markwills1970@gmail.com> wrote:
>>>>
>>>>> They may object, but after an hour of coding without having to use a
>>>>> single label anywhere in their code, they normally come around, just
>>>>> as I did :-)
>>>>
>>>> Indeed so.  It's still baffling to me that assembly programs usually
>>>> don't use structured programming constructs such as loops and if
>>>> statements.  It seems to me that they're just as useful in assembly-
>>>> language programs as in high-level languages.
>>>>
>>>> Andrew.
>>>
>>> I which there were some pseudo machine code language, not optimized to build up Forth, but optimized to be easy to be built up itselves from most machine codes in use. Of course such an intermediate stage would be less than optimal, but maybe more efficient than c and go.
>>>
>>> Then even I would try to learn that language and try to built up a Lehs-Forth.
>>
>> I think the thing you just wished for is... Forth. That's what it is!
> 
> If so, then Forth would be the perfect platform for building an operating system.
> 
One of the items with Go is that its assembler uses generic instructions 
so they are the same with several 32/64 bit CPUs, I have never tried it. 
That maybe accomplished with an assembler if good macro capabilities.

-- 
Cecil - k5nwa

[toc] | [prev] | [next] | [standalone]

#57814

On Saturday, 22 April 2017 11:12:55 UTC+1, Andrew Haley  wrote:
> Mark Wills <markwills1970@gmail.com> wrote:
> 
> > They may object, but after an hour of coding without having to use a
> > single label anywhere in their code, they normally come around, just
> > as I did :-)
> 
> Indeed so.  It's still baffling to me that assembly programs usually
> don't use structured programming constructs such as loops and if
> statements.  It seems to me that they're just as useful in assembly-
> language programs as in high-level languages.
> 
> Andrew.

Absolutely. In my case I didn't realise just *how* useful there were
in assembly until I started using them in my own Forth TMS9900 assembler.

[toc] | [prev] | [next] | [standalone]

#58323

Jan Coombs wrote:

> I think that if you asked a number of assembler programmers to
> use the simplest forth assemblers that I have used and written,
> then the average response would be:
> 
>   "That is not [what I call] an assembler"

That may be. I wouldn't call my wife's Smart (car) a real car either,
but it officially is. <g>

-- 
Rudy Velthuis        http://www.rvelthuis.de

"The man with a toothache thinks everyone happy whose teeth
 are sound." -- George Bernard Shaw

[toc] | [prev] | [next] | [standalone]

#58338

On Sat, 29 Apr 2017 03:05:16 +0000 (UTC)
"Rudy Velthuis" <newsgroups@rvelthuis.de> wrote:

> Jan Coombs wrote:
> 
> > I think that if you asked a number of assembler programmers
> > to use the simplest forth assemblers that I have used and
> > written, then the average response would be:
> > 
> >   "That is not [what I call] an assembler"
> 
> That may be. I wouldn't call my wife's Smart (car) a real car
> either, but it officially is. <g>

Perhaps there is no official designation for half a car.:)

Jan Coombs

[toc] | [prev] | [next] | [standalone]

#57813

On 4/22/2017 05:49, Rudy Velthuis wrote:
> Alex wrote:
> 
>>> So you can inline the assembly. That's fine. It is still (x-86, in
>>> this case) assembly. <shrug>
>>>
>>
>> It's the other way round. The assembler is inlining Forth.
> 
> Not sure if you are serious, but that is nonsense anyway. Something
> that uses assembler mnemonics to generate machine code is an assembler,
> no matter if it is written in Forth or not. Oh, it is a built-in (or
> loadable) assembler, but these are not unique to Forth either.
> 

It's turtles all the way down. Or up, depending where you are.

Yes, I was serious. Semantic arguments about where the assembler ends 
and where Forth starts are pretty pointless, because everyone has their 
own opinion. What is true is that languages are the end product of 
bootstrapping from some other language. In the case of a lot of Forths, 
bootstrapped from Forths with Forthish assemblers written in Forth.

-- 
Alex

[toc] | [prev] | [next] | [standalone]

#58324

Alex wrote:

> Yes, I was serious. Semantic arguments about where the assembler ends
> and where Forth starts are pretty pointless

Not at all. The Forth *assemblers* (note the name) are still
assemblers, i.e. they turn assembler mnemonics into machine code.
Proper assemblers may be more comfortable, but the Forth ones remain
assemblers.
-- 
Rudy Velthuis        http://www.rvelthuis.de

"Researchers have discovered that chocolate produces some of the
 same reactions in the brain as marijuana. The researchers also 
 discovered other similarities between the two but can't remember
 what they are." -- Matt Lauer on NBC's Today Show.

[toc] | [prev] | [next] | [standalone]

#58340

On 4/29/2017 04:08, Rudy Velthuis wrote:
> Alex wrote:
> 
>> Yes, I was serious. Semantic arguments about where the assembler ends
>> and where Forth starts are pretty pointless
> 
> Not at all. The Forth *assemblers* (note the name) are still
> assemblers, i.e. they turn assembler mnemonics into machine code.
> Proper assemblers may be more comfortable, but the Forth ones remain
> assemblers.
> 

You're proving my point.

Or, try this; 
http://www.complang.tuwien.ac.at/anton/euroforth/ef13/papers/ertl-paf.pdf

-- 
Alex

[toc] | [prev] | [next] | [standalone]

#58341

Den lördag 29 april 2017 kl. 10:46:04 UTC+2 skrev Alex:
> On 4/29/2017 04:08, Rudy Velthuis wrote:
> > Alex wrote:
> > 
> >> Yes, I was serious. Semantic arguments about where the assembler ends
> >> and where Forth starts are pretty pointless
> > 
> > Not at all. The Forth *assemblers* (note the name) are still
> > assemblers, i.e. they turn assembler mnemonics into machine code.
> > Proper assemblers may be more comfortable, but the Forth ones remain
> > assemblers.
> > 
> 
> You're proving my point.
> 
> Or, try this; 
> http://www.complang.tuwien.ac.at/anton/euroforth/ef13/papers/ertl-paf.pdf
> 
> -- 
> Alex

Can Paf be downloaded? Does it works well? Several processors?

[toc] | [prev] | [next] | [standalone]

Page 2 of 7 — ← Prev page 1 [2] 3 4 5 6 7  Next page →

Back to top | Article view | comp.lang.forth

csiph-web