[r6rs-discuss] (no subject)

[Chris Hanson]

On Sun, Feb 22, 2009 at 7:02 PM, Aaron W. Hsu <arcfide at sacrideo.us> wrote:

> I can think of at least one (maybe two?) candidates in this election who
> are involved in academia and who have indeed maintained and continue to
> maintain very large systems with what I would consider to be resounding
> success, in Scheme.

It's not surprising that I feel out of the mainstream, when no one
seems to even know who I am.

MIT/GNU Scheme dates back to 1982 and I've been maintaining it
continuously since then; at least two people posting to this list,
Bill Rozas and Arthur Gleckler, have done significant work on this
implementation -- particularly Bill.  In addition to being one of the
largest implementations, it's also been used as a base for building
moderately large applications, and I'm pretty sure it's the oldest
implementation still in reasonably widespread use.  Building this
implementation certainly provided many lessons in software
development, but I have found that building applications in the
implementation have been more educational.

Although I consider myself as someone who understands large-scale
programming, my recent move to Google showed me that academia really
has NO CLUE what "large scale" means.  Google's code base is several
orders of magnitude larger than anything I'd previously seen.  I've
spent the last year slowly learning my way around a tiny corner of the
code.  A Scheme implementation, no matter how large and complete,
can't really compare to this kind of scale.  So although it may be
true that there are academics who understand programming at scale,
clearly there are lessons to be learned from the people who build
commercial software.