[Prev][Next][Index][Thread]
Philosophy/strategy for CL-HTTP
I have the impression that a number of the improvements and extensions
of CL-HTTP that are developed over time are often rather dependent
on upgrades of the substrate, e.g., OpenTransport upgrades on the Mac.
Often when a person has trouble getting CL-HTTP running, they are told
that they need to have the latest version of this and that in their
system.
To the degree that my suppositions are true, this means that the
system is being built on a base that's in motion. That is, the system
is fragile. At the other extreme, MacPaint, written in '84 or so,
still runs on PPC Sys 7.5.
It would seem to me that there would be more adoption and use of
CL-HTTP if users knew that they could move up to the most recent
version of CL-HTTP with more and better features without at the same
time having to acquire and install upgrades and patches for other
system components.
I fully realize that networking and threads and the like are being pushed
hard in the implementation of CL-HTTP and that there's a great
temptation to take advantage of each improvement in these areas. But though
the wizards in Cambridge may be comfortable with such dependencies,
I think it inhibits spreading use of the system.
I think CL-HTTP is a great system and deserves wider use. I think that
the team there should think carefully about making installation and
upgrading orthogonal to some of these messier issues.
I'm looking forward to hearing what others think.
-- Bob Futrelle
_______________________________________________________________
Robert P. Futrelle | Biological Knowledge Laboratory
Associate Professor | College of Computer Science 161CN
Office: (617)-373-2076 | Northeastern University
Fax: (617)-373-5121 | 360 Huntington Ave.
futrelle@ccs.neu.edu | Boston, MA 02115
__________________________|____________________________________
Follow-Ups: