Re: [office-collab] Action Item ODF application vendors

[Svante Schubert <svante.schubert@gmail.com>]

Let me start from your last comment: you thought we decided that runtime
behavior is irrelevant as long as the format doesn't restrict what an
implementation can store.

In change-tracking a user edits a document in "ODF application A" from a
state to another. Let's say from A1 to A2 - and the document can be
opened by any other "ODF application B", where the former changes are
being reviewed and in case of a rejected the former state change from A1
to A2 should be undone, under the assumption that both ODF applications
are able to represent both states A1 and A2.

In addition we know that ODF applications in general does not have a
generic model where arbitrary XML changes can be executed, therefore an
generic approach like GCT required the application always to analyze the
change and map it to an internal state change by API.
ECT with its buckets where tending to use a different approach: to save
the state before and after the change. MCT is going a third way by
identifying reoccurring ODF XML changes and defining them in the
specification, referring to such a specified change by a label or more
precisely by a serialized parametrized operation in the document.

In our use case from last call, we have identified the reoccurring
change of merging of a paragraph/heading to a single one and the related
style changes.

The task for the ODF application vendors is to test for their
applications the overall state change driven by their UI, when merging
two paragraphs/headings and to find potential edge cases (e.g. empty
paragraphs), where state changes might be unexpected.
The expected state change was documented by myself at
https://lists.oasis-open.org/archives/office-collab/201210/msg00048.html
(see the example from the URL with XML comments below)
In the end we want to adapt the change-tracking to the existing behavior
of ODF applications, if the behavior is different we might need to split
the operation to more fine granular and/or applications might consider
to adapt the change behavior to archive better interoperability on user
operations.

Does this clarify your question?

Unfortunately Oliver will not be able to sent/upload a test document to
the list today, please start working on the one given given by Dennis
prior and extend it by adding complexity and/or edge cases as empty
paragraphs/headings.

PS: I have updated below the XML example for the undo by using XML
comments and added a missing quotation mark:
<split s="/2/10">
    <add type="style" name="Normal">
        <add type="style-text" name="hyphenate">false</add>
    </add>
    <!-- style family is redundant, as a Paragraph was be chosen,
"Normal" the name of the parent, as no "name" it is automatic style -->
    <add type="style" s="/3" inherit="Normal">
        <!-- perhaps we use a measure unit parameter -->
        <add type="style-text" name="margin-left">0.5in</add>
    </add>
</split>
<add type="paragraph" s="/2/11">simple demonstration of cross-structure
change-tracked deletion.</add>
<add type="paragraph" s="/3/1">The deletion is a </add>


Best regards,
Svante


On 13.11.2012 20:45, John Haug wrote:
> To be clear, what exactly is requested?
>
> As I recall, have two adjacent headings, one formatted as a heading and the other as text, then see what happens when you backspace or delete (or something else) to join the paragraphs?  I thought we decided that runtime behavior is irrelevant as long as the format doesn't restrict what an implementation can store.
>
> John
>
> -----Original Message-----
> From: office-collab@lists.oasis-open.org [mailto:office-collab@lists.oasis-open.org] On Behalf Of Svante Schubert
> Sent: Tuesday, November 13, 2012 6:26 AM
> To: office-collab@lists.oasis-open.org; Oliver-Rainer Wittmann
> Subject: [office-collab] Action Item ODF application vendors
>
> As discussed on our last call all ODF application vendors have the action item to test the prior discussed merge paragraph (heading) scenario with their application to identify edge cases.
> Did I understand it correctly that Oliver volunteered at the Apache EU Con to come up with some test scenarios/documents?
>
> PS: The next call is next Wednesday -
> http://www.timeanddate.com/worldclock/meetingdetails.html?year=2012&month=11&day=21&hour=14&min=30&sec=0&p1=179&p2=37&p3=136&p4=224&iv=1800
> Postponed an hour to 6:30 PST so even Dennis might have a change to join. ;)
>
> Best regards,
> Svante
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: office-collab-unsubscribe@lists.oasis-open.org
> For additional commands, e-mail: office-collab-help@lists.oasis-open.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: office-collab-unsubscribe@lists.oasis-open.org
> For additional commands, e-mail: office-collab-help@lists.oasis-open.org
>