Where is the Lilith code?

"Jos Dreesen" <jos.mar@bluewin.ch> wrote in message
news:480b9fc1$1_1@news.bluewin.ch...
>
> The recovered Medos sources are available on
> ftp://jdreesen.dyndns.org/ftp/Medos_4

Fantastic! Thank you very much for all your efforts. I really needed these
25 years ago when I was porting the M2MPC compiler but better late than
never ;-) It has been a surprise to me that some of code is much less
complex than I had imagined it to be at the time -something like the feeling
you get when a magician explains his tricks.

--
Chris Burrows
CFB Software
http://www.cfbsoftware.com/modula2

> never ;-) It has been a surprise to me that some of code is much less
> complex than I had imagined it to be at the time -something like the feeling
> you get when a magician explains his tricks.
>
Which is , of course, the earmark of goof software design !

BTW thanks to Svend Knudsen for letting me share his code.

Jos

>> never ;-) It has been a surprise to me that some of code is much less
>> complex than I had imagined it to be at the time -something like the
>> feeling you get when a magician explains his tricks.
>>
> Which is , of course, the earmark of goof software design !

Make that "good" software design....

Chris Burrows wrote:

> "Jos Dreesen" <jos.mar@bluewin.ch> wrote in message
> news:480b9fc1$1_1@news.bluewin.ch...
>
>>The recovered Medos sources are available on
>>ftp://jdreesen.dyndns.org/ftp/Medos_4
>
>
> Fantastic! Thank you very much for all your efforts. I really needed these
> 25 years ago when I was porting the M2MPC compiler but better late than
> never ;-) It has been a surprise to me that some of code is much less
> complex than I had imagined it to be at the time -something like the feeling
> you get when a magician explains his tricks.

I recall many years ago, downloading the source to one of
Wirth experimental languages, that targeted PLD and Hardware design,
and being rather amazed at the simplicity of the code.

I have just looked for it, but cannot find the exact branch.

It must have been later than PL/0 (1975?), and earlier than Lola
(1995?), probably it led to Lola.

-jg

"Jim Granville" <no.spam@designtools.maps.co.nz> wrote in message
news:481c35ee$1@clear.net.nz...
>
> I recall many years ago, downloading the source to one of
> Wirth experimental languages, that targeted PLD and Hardware design,
> and being rather amazed at the simplicity of the code.
>
> I have just looked for it, but cannot find the exact branch.
>
> It must have been later than PL/0 (1975?), and earlier than Lola
> (1995?), probably it led to Lola.
>

Maybe it is Debora that you are thinking of? This was used in the design of
one of Lilith's descendants, the Ceres-3 computer. A specification of
Ceres-3 in Debora is included in the ETH Technical Report 168 - 'Hardware
Description of the Workstation Ceres-3' by B. Heeb and I Noack. You can
download a copy from the ETH Technical Reports site:

http://www.inf.ethz.ch/research/diss...ndex?range=100

--
Chris Burrows
CFB Software
http://www.cfbsoftware.com/modula2

Chris Burrows wrote:
> "Jim Granville" <no.spam@designtools.maps.co.nz> wrote in message
> news:481c35ee$1@clear.net.nz...
>
>> I recall many years ago, downloading the source to one of
>>Wirth experimental languages, that targeted PLD and Hardware design,
>>and being rather amazed at the simplicity of the code.
>>
>> I have just looked for it, but cannot find the exact branch.
>>
>> It must have been later than PL/0 (1975?), and earlier than Lola
>>(1995?), probably it led to Lola.
>>
>
>
> Maybe it is Debora that you are thinking of? This was used in the design of
> one of Lilith's descendants, the Ceres-3 computer. A specification of
> Ceres-3 in Debora is included in the ETH Technical Report 168 - 'Hardware
> Description of the Workstation Ceres-3' by B. Heeb and I Noack. You can
> download a copy from the ETH Technical Reports site:
>
> http://www.inf.ethz.ch/research/diss...ndex?range=100

Thanks for that link, another one I have not seen before.

Debora looks like a superset, of the one I recal, as I see '=', ".=",
and ":=" assignment options, and have not seen ".=" used before ?.

I have seen '=' for combinatorial and ":=" for 'at the clock' used
before (Abel IIRC).
Others (eg CUPL) use Reg.D =, Reg.T = to infer the clocked transfer.

A separate symbol could make a simulator easier to write, and revise.


The syntax also uses psuedo record constructs, something
that seems to have got lost in modern HDLs.

I think

Register.clk .= ClockNet;
Regoster.D .= Register.Q;

is (to me) quite clear, and reminds the designer they are dealing with
hardware.

Modern HDLs will include this format in their report files, but
do not allow users to write that way ?

-jg