My comments on DR 10-0001

[MURATA Makoto (FAMILY Given)]

> Thinking more in terms of standard error recovery strategies would a
> useful tool in our toolbelt. This is how the WhatWG are treating HTML for
> example.

I think that this approach is ugly.  It is a backdoor for allowing 
more data and definig semantics.

I might be persuaded if the current situation is already  ugly and 
we cannot forget broken data.  However, almost no existing OOXML 
documents contain ISO 8601 cell values.

Do we really want to "allow" documents containing "1900-02-29"?

Cheers,
Makoto