IRC log from today's meeting

Dear SC,

Patrick and I decided to cancel the meeting over the holiday season, therefore our next meeting will be on the 11th of January: http://www.timeanddate.com/worldclock/meetingdetails.html?year=2016&month=01&day=11&hour=15&min=30&sec=0&p1=179&p2=37&p3=136&p4=234&iv=1800

The teleconference login data for next call will be found in the OASIS calendar event:

https://www.oasis-open.org/committees/event.php?event_id=39512

Please find the IRC log of today's meeting below:

Attendees: Patrick, Svante

[16:38] Svante Schubert: We should skip this call - https://www.oasis-open.org/apps/org/workgroup/office-collab/event.php?event_id=39511

[16:38] Svante Schubert: ^^ It is in the holiday season

[16:40] Svante Schubert: Patrick: ODF file (proposed) consists of (think of the present package) of:

content.xml (created with new document, has all the default markup for an instance)

change.xml which consists of an xml document with valid XPath/XQuery statements against content.xml and/or against other statements in change.xml.

[16:41] Patrick: original content.xml

[16:41] Patrick: then, there are changes in change.xml

[16:41] Patrick: Svante says, but changes.xml has undo in case change tracking not supported

[16:42] Patrick: undo implies acceptance of the changes

[16:42] Svante Schubert: Svante: ^^ if the content.xml has the lastest state and the changes.xml has the undo changes, the applications without support of change-tracking can still read the latest proposal.

[16:43] Patrick: Svante says latest state is original content.xml with changes (accepted or not) reflected in content.xml and then undos are in change.xml

[16:44] Svante Schubert: Patrick: What if there are conflicting changes?

[16:44] Svante Schubert: If we use change-tracking it is like exchanging the document on the floppy disc, there is no real-time collaboration nor merge at any time.

[16:45] Svante Schubert: There is only one editor at the time.

[16:46] Patrick: Ah, ok, with one editor, writing the changes into the document and then putting undo statements in change.xml would work, not sure it is the best way

[16:47] Patrick: I say not the best way, because the reader would not know which changes are pending

[16:54] Svante Schubert: Patrick and I discussing pointing to changes.

[16:55] Svante Schubert: If there is are four siblings a table, an image and two paragraphs. The pointer to the fourth component would be "/4"

[16:56] Svante Schubert: Because every component (logical object) will be counted.

[16:56] Patrick: but I am asking how do we point to a sibling inside the paragraph?

[16:57] Svante Schubert: If there is the text "Hello World!" within the 4th paragraph, the W of World would be "/4/7"

[16:57] Svante Schubert: Note: Every letter is a logical unit (component)

[16:58] Svante Schubert: Or the first component the table has 3 rows with 4 columns. The counting would /1/2/3 for the table being the first component, the 2nd row and the 3rd cell

[16:59] Svante Schubert: The advantage of the typeless pointer are that the OT, the movement back and forth dependent if something is deleted or inserted in front of the component is not type dependent, neither.

[17:01] Patrick: <text:p>Hello World!</text:p>

[17:02] Patrick: <text:section><text:p>Hello World!</text:p></text:section>

[17:04] Patrick: use the descendant axis only and each element in the descendant axis is counted and then for text content, count each character but only after the start element that begins the text string

[17:06] Patrick: <text:p>Hello World!</text:p> and I had <text:p>Hello <text:alphabetical-index-mark-start>World!<text:alphabetical-index-mark-end></text:p> then it would be /1/2/1 for the W

[17:06] Svante Schubert: It depends if <text:section> would be defined as logical component for itself. For instance, text:span would not count.

[17:06] Patrick: not counting all elements

[17:07] Svante Schubert: We are counting in general the start-tag of such a logical block.

[17:07] Svante Schubert: While there are start/end markers, such as text:span

[17:07] Svante Schubert: which do not count

[17:08] Patrick: descendant-axis doesn't include milestones - XPath doesn't include

[17:09] Svante Schubert: <draw:frame><draw:image .../> </draw:image.../> </draw:frame> would be only one logical component, although there multiple (alternatives) images

[17:16] Patrick: Essence of proposal is a sub-set of ODF elements that will be counted for the construction of a sub-tree that is used for change tracking

[17:17] Svante Schubert: Usually I mapped the sub tree to a list of changes to create this subtree

[17:19] Svante Schubert: The browser received the sequence of changes, not knowing if this document was created by a user or loaded from a server

[17:20] Patrick: hand shake on what change tracking features are supported, text, tables, images, etc.

[17:21] Patrick: doesn't that mean that ODF must define the feature sets for each type of change tracking?

[17:22] Patrick: Svante says define default subset or full or nothing

[17:30] Svante Schubert: Feature delta to the default feature set would be possible, but only later interesting for merging/real-time collaboration. For instance, in our VI vs. MS Office example.

office-collab message