projects update
5 messages · 2010-07-03 → 2010-09-07 · Yahoo Group era · View archive on archive.org
Participants: cheveron, jdiffendaffer
Preserved from the Timex/Sinclair 2068 Yahoo Group (2001–2019), which is no longer online. Text reproduced from the archive.org archive; email addresses masked.
Messages
1. projects update
cheveron · Sat, 03 Jul 2010 08:16
Not sure if people have been following any of the projects I've been working on or when I last posted an update so here's the current status:
SE Basic - finished.
http://bit.ly/sebasic
16K replacement for the normal Speccy ROM. Burn to EPROM and use in the cartridge port with a Timex. Includes the DELETE, ON ERR and SOUND commands that are similar to the ones found in TS2000 BASIC and a whole lot more besides. Probably the fastest version of Sinclair BASIC.
ULAplus - design finalized. Hardware to follow, perhaps early next year.
http://bit.ly/ulaplus
ULAplus is a replacement ULA (SCLD) that provides up to 64 colours on screen from a palette of 256 by using the BRIGHT and FLASH bits of the attribute byte as a CLUT (color look up table).
The ZX Spectrum ULA Book - editing in progress. Pre-order now.
http://bit.ly/ulabook
"For the first time, the inner working of the Sinclair ZX Spectrum's custom chip and heart of the computer, the Ferranti ULA, is exposed in minute detail."
ZXI - in RFC phase.
http://bit.ly/zxinterface
A new standard for plug in expansions for the ZX Spectrum designed to improve hardware compatibility. Unlike the Timex, the Spectrum does not fully decode the I/O ports. This causes all manner of headaches. ZXI compliant devices may also work on Timex machines with a twister board.
ZX Spectrum 128Ke - spec finished. Haven't built it yet.
http://bit.ly/ZX128Ke
A standard for an 'ideal' Spectrum 128. The early machines had incorrectly documented RAM contention which leads to some games either not working as intended or in extreme cases, crashing. This also resolves some compatiblity issues with earlier games.
ZX Spectrum SE killer app - still working on the 64K demo
(http://scratchpad.wikia.com/wiki/ZX_Spectrum_SE)
A project to demonstrate that the Speccy SE is not a chocolate teapot. I'm not publishing any details about it, as it may never see the light of day. However, I am interested in hearing from C programmers who are willing to use z88dk and are interested in contributing to it.
Secret Project - on target for the end of August
No hints about this one.
Timex Archive - started
http://bit.ly/timexarchive
I have the most comprehensive collection of Timex software including a lot of stuff that isn't available elsewhere (because I made it). Hope to get this finished by the end of the year, but time constraints may prevent that from happening.
2. Re: projects update
jdiffendaffer · Fri, 20 Aug 2010 22:37
--- In [email], "cheveron@..." <cheveron@...> wrote:
>
> Not sure if people have been following any of the projects I've been working on or when I last posted an update so here's the current status:
>
>
> SE Basic - finished.
> http://bit.ly/sebasic
>
> 16K replacement for the normal Speccy ROM. Burn to EPROM and use in the cartridge port with a Timex. Includes the DELETE, ON ERR and SOUND commands that are similar to the ones found in TS2000 BASIC and a whole lot more besides. Probably the fastest version of Sinclair BASIC.
Looks cool but are you sure it's a good idea to use FORMAT as a non-disk oriented command?
3. Re: projects update
cheveron · Mon, 23 Aug 2010 18:01
--- In [email], "jdiffendaffer" <jdiffendaffer@...> wrote:
>
> --- In [email], "cheveron@" <cheveron@> wrote:
> >
> > Not sure if people have been following any of the projects I've been working on or when I last posted an update so here's the current status:
> >
> >
> > SE Basic - finished.
> > http://bit.ly/sebasic
> >
> > 16K replacement for the normal Speccy ROM. Burn to EPROM and use in the cartridge port with a Timex. Includes the DELETE, ON ERR and SOUND commands that are similar to the ones found in TS2000 BASIC and a whole lot more besides. Probably the fastest version of Sinclair BASIC.
>
> Looks cool but are you sure it's a good idea to use FORMAT as a non-disk oriented command?
I couldn't add any more tokens. It only functions that way if you use it in isolation. Any string after the command will cause the normal error at which point whatever disk system you are using will pick it up.
4. Re: projects update
jdiffendaffer · Tue, 24 Aug 2010 20:53
--- In [email], "cheveron@..." <cheveron@...> wrote:
>
> > Looks cool but are you sure it's a good idea to use FORMAT as a non-disk oriented command?
>
> I couldn't add any more tokens. It only functions that way if you use it in isolation. Any string after the command will cause the normal error at which point whatever disk system you are using will pick it up.
>
The TANDY CoCo and Dragon use two token tables to hold Extended Color Basic. If the value of the token is 255 (I think) then it calls a routine to decode the next byte from another token table. Basically new tokens are two bytes. Tokenizing takes longer but it's not much of an impact at runtime, especially if you implement rarely used keywords in the 2nd table.
5. Re: projects update
cheveron · Tue, 07 Sep 2010 06:41
--- In [email], "jdiffendaffer" <jdiffendaffer@...> wrote:
>
> --- In [email], "cheveron@" <cheveron@> wrote:
> >
> > > Looks cool but are you sure it's a good idea to use FORMAT as a non-disk oriented command?
> >
> > I couldn't add any more tokens. It only functions that way if you use it in isolation. Any string after the command will cause the normal error at which point whatever disk system you are using will pick it up.
> >
>
> The TANDY CoCo and Dragon use two token tables to hold Extended Color Basic. If the value of the token is 255 (I think) then it calls a routine to decode the next byte from another token table. Basically new tokens are two bytes. Tokenizing takes longer but it's not much of an impact at runtime, especially if you implement rarely used keywords in the 2nd table.
>
I'd already considered that method, but the problem is that there's only so much you can cram into 16K. There just isn't enough room to extend the parser, tokenizer and detokenizer to handle double-byte tokens. The benefit of using characters 0-5 is that you can rotate the value of the character when a token is detected and the new tokens then occur logically after the originals, which makes it possible to use a single table for tokenization and detokenization.