[om] a developer's perspective on CDs
Manfred Riem
mriem at win.tue.nl
Sun Oct 29 16:11:58 CET 2000
Hi,
> I suggest that two different RFCs can make two different
> definitions of the same CD. Then you can distinguish which definition
> to take for your symbol by referring to the number of the RFC. This
> way, eventually, people will be able to agree on meanings of symbols,
> issue revised CDs and make their software adhere to the new RFCs
> without changing all the symbol names.
I agree with this option as far as the administrative part of the job
is concerned. However, if you are using an system that relies on a
specific CD I would suggest the more common approach of giving it a
number (as is also done for the HTTP 1.1 protocol which is described
in a RFC ;-) ).
> It seems to me a far smoother evolutionary path. I think things can
> only get messier if, when I want to slightly change the meaning of
> (arith, times) to say that it only applies in cartesian closed
> categories, I have to call it (arith_andrew, times), and then when
> the community agrees to the change, change all my code back to using
> (arith, times). It would be better if I could just change the RFC
> number in the comment at the top of my source file.
Yes, because it is evolutionary the well established scheme used for
the Linux kernel springs to mind. This still makes it possible to have
just one OM-RFC document, but multiple versions of the CD still can
exist (otherwise a lot of OM-RFC documents will be generated and then
administration is once again .... a big issue).
For the people that don't know how Linux kernel numbering goes:
u.x.yy : x even -> it is a stable release,
: x odd -> it is a development release.
: yy -> a subrelease which doesn't change something major,
but still a number because something was changed
(eg Documentation of the CD is clarified).
: u -> the major release number.
A stable release could be described by a newer RFC document.
Greetz,
Manfred.
--
om at openmath.org - general discussion on OpenMath
Post public announcements to om-announce at openmath.org
Automatic list maintenance software at majordomo at openmath.org
Mail om-owner at openmath.org for assistance with any problems
More information about the Om
mailing list