From: Caolan McNamara (cmc at domain stardivision.de)
Date: Tue 09 May 2000 - 14:14:43 IST
>>>>>>>>>>>>>>>>>> Original Message <<<<<<<<<<<<<<<<<<
On 09.05.00, 14:00:54, "John P. Looney" <jplooney-ilug at domain online.ie> wrote
regarding Re: [ILUG] Windows to go opensource:
> You don't think that they would just release a new version, citing
> "performance reasons" for not keeping backward compatibility ?
When I say interoperability I really have little interest in backward
compatibility issues, more that when they bring out a new version they
must simultaneously bring out the new spec (which must be correct of
course) so that everyone else can upgrade their source trees to take the
new interface into account. Why force any company to retain backwards
compatibility, from a anticompetitive stance I don't know if this
matters, from a consumer rights angle you could make a case that MS is
forcibly devalueing your software, but car manufacturers do pretty much
that anyway.
To the person talking about the header files being released, well yes
there really are two closely related issues I suppose.
1) the api into windows itself, thats often well specified, though you'll
find that even the well known wmf format is not fully documented.
(http://www.csn.ul.ie/~caolan/docs/libwmf.html), and there are many many
undocumented/badly documented functions, a raw header file is not a full
description of use and doesn't cut it for me.
2) But to be fair I couldn't give a rats ass what happens under windows,
its what happens when windows technologies inpact on the rest of the
world, ms networking protocols and document and data formats. Of more
importance is the necessity of true and binding specifications of various
technologies, rtf is a complete nightmare (Im told) with many little
varients thought up of at the drop of a hat inside ms, easy to write rtf,
hard to have a robust reader. So I suppose I should be more precise and
demand that full and complete documentation for Microsoft's API's,
networking protocols and data formats must be released in parallel as the
products themselves and must be verifiably correct.
You could make a case that that such a "big three release" should be
enforcable on all companies regardless. Its your data that being moved
and saved so you should have the capacity to convert it whatever way you
want. Though there are enough issues with that approach which leads me to
just recommend rather than enforce that data be kept and stored in such
formats.
C.
This archive was generated by hypermail 2.1.6 : Thu 06 Feb 2003 - 13:06:04 GMT