[Aldor-l] Using Aldor or SPAD

[Ralf Hemmecke]

On 08/18/2006 08:25 PM, Bill Page wrote:
> On August 18, 2006 1:49 PM Martin Rubey wrote:
>> Gabriel Dos Reis writes:
>>
>>> I've been writing more and more SPAD, less and less Aldor. 
>>> And I suspect very soon my students will more SPAD than 
>>> Aldor, however imperfect SPAD...
>> why is that? I really can't understand, there are so many 
>> things missing in spad!

On which list did that appear. I don't remember that I have seen these 
mails... :-(

> Ya, like error message a human can understand, for one! :)
> I think SPAD is nearly impossible to learn because the results
> of compiling incorrect programs are so difficult to interpret.
> I usually write something in Aldor first to get the syntax and
> types correct and then (if necessary) back port it to SPAD.

That sounds reasonable and it probably results in better readable SPAD 
code. But anyway, SPAD should die out, either by using an sufficiently 
open source Aldor compiler or by modifying the current SPAD compiler to 
accept the Aldor language (and giving proper error messages).

> On the other hand (with a few notable exceptions like fully
> implemented dependent types), the SPAD *language* is not
> fundamentally very different than Aldor. SPAD is a productive
> language - after all 1,300 packages domains and categories
> are currently implemented in this language.

If you really look at the code then you see that most of it must be 
considered as black boxes. Even if you can read the spad sources (since 
they are open), it remains our task to make pamphlets out of them. And I 
would rather like to do this by replacing a non-strictly defined 
language like SPAD by a more strictly defined language Aldor.

> Then, of course there is the long unresolved issue of the lack
> of a clear open source license for Aldor. I fear that if this
> not resolved in the next few **weeks**, then the consequences
> for both Aldor and Axiom could be very negative.

What can you do? The only option we have if the Aldor compiler is not 
going to open soon, is to write another compiler. But I guess that is 
going to take quite a while. First we have to find people who are 
willing an able to start such a project.

Gaby, have you any idea how long it could take to produce another 
reasonably efficient Aldor compiler?

Ralf