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

[Nigel Stewart]

Ben,

> Now, it's sounds real nice to say things like "Let's jettison all
> this legacy baggage and start from a fresh, clean design", but reality
> doesn't work that way.  There are huge costs associated with starting
> from scratch.  In this case, we're talking about potentially large
> numbers of "legacy code" in the form of spreadsheets that would need
> to be audited.

There are also huge costs associated with supporting all of this
legacy hackery for the sake of compliance with a standard.  As a
pragmatic issue I have no objection to extensions to the standard
for legacy reasons, but I think the core standard should be
standards-document based, rather than legacy behavior based.

I disagree with the notion of standardisation based on
"the ugly historical reality" instead of "best known practices".

Perhaps it's the difference between being conservative or
being progressive, in the broader sense.

I do see the point you're making, and the reasons for MS to push
OOXML in this manner.  I'm just expecting MS to make claims
about being "standards based" in the same sense that an XHTML-
compliant web browser can claim to be standards based.  I see
this as an effort to undermine the perceived value of standards
in the context of document interchange.

After all, there are legacy file formats for the legacy stuff.
If anyone should have to worry about converting the legacy stuff,
it should be MS, not everyone else trying to interchange OOXML.
(I would like Open Office to have fair chance of dealing with it)

BTW - I think it would be great to have "legacy free" C and C++
build modes, including a length-checked sprintf.

Nigel