Reasonable people can disagree, even about history.

I believe it was a mistake to implement the Aldor compiler in C.

Implementing a new compiler inside Scratchpad would have cleaned
up the Spad language, the Spad compiler, and made the
interpreter/compiler semantic difference disappear. It would
also have been a much smaller effort as most of the machinery
already existed. 

Dick Jenks thought the same thing but Stephen argued that we
could re-implement the algebra in the C version (which was 
eventually my PhD task). Is there any Aldor algebra in the
build process?

I taught a compiler course at Vassar and used some examples
from Aldor. It is a very pretty piece of software written by brilliant
programmers. However, there are more people that "speak boot" 
than "speak Aldor" and that is already a vanishingly small number.

> it was good change for Spad. 

The huge effort would have been better spent developing more algebra
or, as I argued then and now, developing specification/proof technology
inside Scratchpad. Reasonable people disagree with me :-)

The point is moot. Aldor exists.


On Sunday, April 23, 2023 at 5:15:04 PM UTC-4 Waldek Hebisch wrote:

> On Sat, Apr 22, 2023 at 09:00:23PM -0700, Tim Daly wrote:
> > Historically this change was introduced because of the Aldor compiler.
>
> Well, this was desirable change to Spad language. Aldor was a new
> start, so it was easier to make change in Aldor first. But
> it was good change for Spad.
>
> > There was a clash of the use of the symbols. I had to rewrite code to
> > get Scratchpad (Axiom) and Asharp (Aldor) to work together.
>
> Yes. Around 1990 using symbol renaming was right thing to do, otherwise
> change would require too much effort. But we want to move forward...
>
> > 
> > On Saturday, April 22, 2023 at 3:19:32 PM UTC-4 Waldek Hebisch wrote:
> > 
> > > For many years internally Spad compiler and interpreter renamed
> > > '%' to '$' and internally used '$'. I have now commited a patch
> > > that disables this renaming so that '%' from source is internally
> > > represented by itself, without renaming. This should slightly
> > > simplify structure of Spad compiler and interpreter, but main
> > > gain should be in debugging: Lisp data structures should be
> > > more similar to Spad source, so easier to interpret.
> > >
> > > There is also related cost: patch touches a lot of places in
> > > interp subdirectory and potentially may introduce bugs.
> > >
> > > -- 
> > > Waldek Hebisch
> > >
> > 
> > -- 
> > You received this message because you are subscribed to the Google 
> Groups "FriCAS - computer algebra system" group.
> > To unsubscribe from this group and stop receiving emails from it, send 
> an email to [email protected].
> > To view this discussion on the web visit 
> https://groups.google.com/d/msgid/fricas-devel/783e527c-5071-40b8-90f6-020623b4b07fn%40googlegroups.com
> .
>
>
> -- 
> Waldek Hebisch
>

-- 
You received this message because you are subscribed to the Google Groups 
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/fricas-devel/be189d74-999b-429e-8429-8c77942862a3n%40googlegroups.com.

Reply via email to