> There is clearly a deeply felt need for software abstractions that map
> one-to-one to the mathematical abstractions one sees in technical
> writing that uses "extended objects"*. This is reflected in the many
> efforts to produce a successful array language. It's not clear that a
> successful array language has ever been produced. Perhaps it would be
> fruitful to examine why this is so rather than arguing against the
> fact.
APL certainly tried to be a successful array language, and was for arrays,
but left out the rest of programming.
I think that the problem is that the mathematics that appears in technical
writing leaves out a lot of details, relying on the "common sense"
of the readers. Thus the issue becomes one of AI. There's also an
argument (I forget where I saw it) that programming is in fact
so different from mathematics that trying to treat programming as
mathematical writing won't work.
When I visited IBM's research lab a decade ago, there were people
working on a system for converting mathematical scrawl into code,
but I don't know what became of it.
- Arch
--------------------- Object Oriented Numerics List --------------------------
* To subscribe/unsubscribe: use the handy web form at
http://oonumerics.org/oon/
* If this doesn't work, please send a note to owner-oon-list@oonumerics.org
This archive was generated by hypermail 2b29 : Wed Feb 20 2002 - 03:20:11 EST