Email list about Forth on Sinclair computers

[programandala.net]

An email list has just been created to discuss the Forth programming language on Sinclair computers:

Forth on Sinclair (and Z80) computers

Any QL user with an interest in Forth is welcome. In the Spanish QL users group there were three of us, including me, that programmed the QL in Forth.

By the way, there's also a new email list to discuss Forth in Spanish:

Forth ES

[Silvester]

Great, I've done a few Forth disassemblies myself (Jupiter Ace, ZX81 Husband/Tree Forth, ZX Spectrum F83, White Lightning, and another ZX - I forget which one). One of the things I would love to do is something like ColorForth for QL. Failing that an open QL Forth using a PE front end. Had a quick look at SuperForth F83 many years ago, but for some reason author has encoded it, easier to write new one from scratch. HPR version is OTT (like GForth) - completely misses the point of Forth as designed by Charles Moore.

[programandala.net]

Great, I've done a few Forth disassemblies myself (Jupiter Ace, ZX81 Husband/Tree Forth, ZX Spectrum F83, White Lightning, and another ZX - I forget which one).

Do you keep those disassemblies? Would you like to share them? They would be interesting.

Some months ago I disassembled Abersoft Forth, with some programs written in Abersoft Forth itself:

Abersoft Forth disassembled

Matteo Vitturi, author of vForth for ZX Spectrum, disassembled White Lightning many years ago, by hand! I asked him to publish the manuscript and he kindly accepted to scan the old papers:

https://sites.google.com/site/zxsoffitt ... _lightning

Both Abersoft Forth and White Lighting are ordinary implementations of fig-Forth for Z80, but you always learn new things.

One of the things I would love to do is something like ColorForth for QL. Failing that an open QL Forth using a PE front end. Had a quick look at SuperForth F83 many years ago, but for some reason author has encoded it, easier to write new one from scratch.

I'm not a fan of the ColorForth principles, quite different from "classic Forth", but it's an interesting approach

Years ago, on the QL I used both Computer One Forth and SuperForth. The combination of QL and Forth is powerful.

HPR version is OTT (like GForth) - completely misses the point of Forth as designed by Charles Moore.

I use Gforth a lot. But what is HPR and OTT? I don't follow.

I hope you'll join the Forth on Sinclair (and Z80) computers email list.

[Silvester]

>Do you keep those disassemblies? Would you like to share them? They would be interesting.

Yes, I have kept them. I will tidy them up and post them (forth-sinclair/files?).

>I use Gforth a lot. But what is HPR and OTT? I don't follow.

HPR was author (http://members.tripod.com/phpr/qhpgf6.html) did a much enlarged version of Laurence Reeves 7K QL Forth, but like GForth (et al) added everything and the kitchen sink to it (OTT = over the top). I always thought the idea was you provide a small kernel _and_then_add what you want to it (dictionaries). I think the adage was if you get more than one screen full of words (VLIST/WORDS) from base Forth then you've missed whole the point. Charles Moore commented on this when it took off (F83, ANSI etc). Look at CM's recent multi-processor Forth CPU's for brevity.

[programandala.net]

>Do you keep those disassemblies? Would you like to share them? They would be interesting.

Yes, I have kept them. I will tidy them up and post them (forth-sinclair/files?).

That would be great. Upload them to the forth-sinclar files or mail them to me and I will do. Anyway, I invite you to join the forth-sinclair group.

HPR was author (http://members.tripod.com/phpr/qhpgf6.html) did a much enlarged version of Laurence Reeves 7K QL Forth, but like GForth (et al) added everything and the kitchen sink to it (OTT = over the top). I always thought the idea was you provide a small kernel _and_then_add what you want to it (dictionaries). I think the adage was if you get more than one screen full of words (VLIST/WORDS) from base Forth then you've missed whole the point. Charles Moore commented on this when it took off (F83, ANSI etc). Look at CM's recent multi-processor Forth CPU's for brevity.

I have GF6 in my archive of Forth systems. I had forgotten it.

I agree with you. Gforth is so big, and so complex under the hood, that it's far from the original spirit of Forth. On the other hand, I find it the most comfortable and powerful Forth system I know for GNU/Linux.

Being a rewritable language, it's natural there are many ways to approach, implement, program and use Forth. But I don't think it's a problem. On the contrary, it's one of the strengths and beauties of Forth. You don't have the same needs, requirements, goals and constraints on say an 8-bit bare-bone platform and a modern OS. Forth works great on any field.

I'm writting a modern Forth for ZX Spectrum 128 (and soon also +3), called Solo Forth and by design the kernel contains only the words required by the system: no tools, no extras, no addons. Every optional code is on the library disk. I even added several versions of library code (for example, several implementations of the `case` structure, and other non-standard control structures as well). This way the programmer can choose whatever suits the application (faster or smaller, with compiler security or without it...) or write his own version.

[Silvester]

>I invite you to join...

Thanks, already did. I have just uploaded the files to the group files area. Since they are off topic for QL will continue on group forum Forthwith (pun)...

[programandala.net]

Had a quick look at SuperForth F83 many years ago, but for some reason author has encoded it, easier to write new one from scratch.

Do you mean this?:

The SuperForth words

It has been a surprise. Then I remembered you mention something related here. Maybe you remember some details? I'm discussing this in the email list.

[programandala.net]

The SuperForth words.

Everything is normal. The strange results were caused by a bug in my definition of `>name`, as I explained in the Forth email list. The table has been updated.

[gregtaylor]

One of the best small Forths is Pygmy (http://pygmy.utoh.org/forth.html). It is a tiny application and runs in MS-DOS (or DOSBOX under Windows). It uses blocks for saving (which is both original and, I think, makes you think in a very modular way and has a very basic list of words. It is a great way to learn Forth (since it is so basic).

Greg

[programandala.net]

Had a quick look at SuperForth F83 many years ago, but for some reason author has encoded it, easier to write new one from scratch.

I contacted Gerry Jackson, the author of SuperForth, and he confirmed the code is not encoded or obfuscated. He provided some details about the implementation, for example the 16-bit relative addressing with register A0, used for efficiency. Also, there are many low-level code words which are headerless and whose code is out of the 64 KiB space, in order to save user memory. Perhaps that was what made you think the code was somehow encoded?