Re: [office-collab] Normalization of ODF XML (Earlier - Re: IRC log from today's meeting - 2013-10-09)

[robert_weir@us.ibm.com]

<office-collab@lists.oasis-open.org> wrote on 10/14/2013
01:11:08 PM:

> From: Svante Schubert <svante.schubert@gmail.com>
> To: "office-collab@lists.oasis-open.org"
<office-collab@lists.oasis-open.org>
> Date: 10/14/2013 01:09 PM
> Subject: [office-collab] Normalization of ODF
XML (Earlier - Re: IRC
> log from today's meeting - 2013-10-09)
> Sent by: <office-collab@lists.oasis-open.org>
> 
> For the description of ODF XML test output and for the comparison
of
> ODF XML of different applications an ODF normalization is useful.
> 
> In our last meeting we listed already several points, where ODF XML
> may differ, but the ODF document not:
> 
> Am 09.10.2013 16:18, schrieb Svante Schubert:
> [15:52] Svante Schubert: Other examples
of ODF variations of the 
> same document lead to in different XML:
> [15:52] Svante Schubert: 1) Order of automatic
styles
> [15:52] Svante Schubert: 2) automatic style
names
> [15:52] Svante Schubert: 3) text:space elements
might be split or a 
> single one (similar to span)
> [15:52] Svante Schubert: 4) span can also
be nested or split
> [15:52] Svante Schubert: 5) Hyperlink can
be split as well

> [15:52] Svante Schubert: 6) Further existence
of optional elements/
> attributes..
> [15:55] Svante Schubert: An implementation
note: Nesting spans 
> should be avoided it just makes things complicated (traversing 
> recursivly content, searching for spans, inheriting.. etc.)
> In addition I would like to add that the text
properties of an 
> automatic paragraph style are equal to the text properties of an 
> automatic text style, which uses a span embracing the full paragraph
content.
> 
> This will be mapped, when two paragraphs will be merged and the text
> properties of the automatic paragraph style of the second paragraph
> will be mapped as span around the remaining content of the second.
> 
> In addition, whenever an ODF application does a "Clear Formatting",
> which removes all the "hard" text attributes. The automatic
text 
> properties of automatic paragraph styles are being mapped to spans
> to the remaining "uncleared" area of the paragraph.
> 
> Does anyone know further examples or categories of examples?
> 

Of course there are all the lower level XML differences
which are covered in the Canonical XML Recommendation:

http://www.w3.org/TR/xml-c14n

Other things to think about:

1) Is a signed document the same document as an unsigned
one?  (In general we might want to think about how signing works with
CT.  For example, can you sign a document and then I sign only my
changes?)

2) Are non-visible RDFa/RDF XML annotations considered
changes?

3) Document level settings, e.g., zoom factor in a
spreadsheet.

4) Removing stuff that is not referenced, e.g., removing
a style that is not used.

5) Some spreadsheet functions are "volatile"
and change values at every recalc, e.g., NOW() and RAND().  If I load
a document and save it, the value of that cell changes.  But should
that be a tracked change?

Regards,

-Rob

> Thanks,
> Svante