Long stupid debate on OOXML and the year 1900 (was: Ecma responses to ISO)

[Ben Scott]

On 3/11/07, Jeffry Smith <jsmith at alum.mit.edu> wrote:
> [the backwards compatibility provision in OOXML] ... are all things that
> should be in a conversion program - not a modern data storage format.

  I don't think it's that simple, for reasons I've mentioned elsewhere
[1].  Effecting simultaneous industry-wide change is very difficult.

> Instead, MS has chosen to include all the bugs of every old program ...

  That is very common in the IT world.  Ibid.

> ... in ways that can't be duplicated by anyone else ...

  *THAT* is a big problem.

  Groklaw asserts that this standard is full of requirements for
undefined behavior.  I have a big problem with any such "standard".

  However, for reasons already given [2], I'm no longer willing to put
blind trust in the Groklaw analysis, and I suspect that doing so
actually hurts the cause: We're basing our arguments on shaky
foundations, and ECMA will see that, and dismiss our objections.

> ... in a format that can't be read by the old programs (that don't
> understand XML) ...

  This sounds rather like a double standard.  You're decrying
Microsoft for maintaining compatibility with old documents in one
place, but here decrying them for not maintaining it.

> ... all of which suggests it really is a dump of the MS Office memory.

  What else would it be?  It has to be based on *something*, and if
you're going to store a document, shouldn't it contain what one has
entered into the program?

  Sure, Excel and 1-2-3 and so on are full of warts and misfeatures
and design flaws, but again, just pointing that out does not justify
rewriting the world from scratch.

> [Microsoft] just incorporated the old software into the new.

  You expect them to just dump everything they've ever written?  Many,
ECMA included, will not doubt see that as unreasonable.

  We don't rewrite the Linux kernel from scratch for every major
release, either.

> Oh and that "optional" bit - it's optional unless you want to say you
> can support the whole standard.

  What if I don't want to say that, and just want to say I can read
what you send me?  Sounds like a win.

  And if I *do* want strict bug-for-bug compatibility, I'm presumably
have legacy baggage that needs such, and so it isn't optional for me,
then, either.

Footnotes
---------
[1] http://article.gmane.org/gmane.org.user-groups.linux.gnhlug/9217
[2] http://article.gmane.org/gmane.org.user-groups.linux.gnhlug/9213

-- Ben