Re: [oiic-formation-discuss] Deliverable: odf-diff?

[Sander Marechal <s.marechal@jejik.com>]

Dave Pawson wrote:
> 2008/6/21 Sander Marechal <sander.marechal@tribal-im.com>:
> 
>> I have a better idea: Perhaps the TC can come up with a spec to "normalize"
>> an ODF document? Then the geeks can use a standard XML diff tool (or maybe
>> even plain text diff) to spot differences between the normalized version.
> 
> Give us an idea of what 'normalization' would mean?
> Bearing in mind http://www.w3.org/TR/xml-c14n Canonicalization
> and
> http://www.jclark.com/xml/canonxml.html from James.

Those links you provided are actually very close to what I mean by
"normalizing" an ODF document. Perhaps "Canonicalizing" an ODF document
would be a better term. The W3C spec you point to also coin a nice
phrase: "logically equivalent". That's exactly the term I'm looking for
all this thread :-)

Two ODF documents that differ only in the names of the automatic styles
(but the content of those styles are the same) are logically equivalent.
The ODF spec explicitly says that an application is free to change the
names of the automatic styles (but not the manual styles!)

Two documents that differ only in what order the image references appear
in the manifest are logically equivalent. The spec dictates nothing
about the order. It's not important because it's a set, not a list. As
long as they contain the same references, they're logically equivalent.

I hope this also answers your question from your other message in this
thread:

> User = viewer, the two examples are 'the same'
> User = editor, the stylenames used are different.

> ouch. That is 3 sources of confusion!
> User, editor, now .... the boss!
> Keeping him|her out of it, how can we define this one?

Don't try to view it from the position of a user looking at a document.
View it from the position of a programmer looking at the raw
datastructure. "Do these two datastructures tell you the same thing?"

-- 
Sander Marechal