Our next meeting will be on the 10th of February:
http://www.timeanddate.com/worldclock/meetingdetails.html?year=2016&month=02&day=10&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/apps/org/workgroup/office-collab/event.php?event_id=39488
Please find the IRC log of today's meeting below:
Attendees: Patrick, Peter, Svante
[16:26] Svante Schubert: Hello Peter, dialing in
[16:30] Peter Rakyta: Hi Svante
[16:31] Svante Schubert: The quite long XPath _expression_ of the <text:span> containing the 'A' within the list:
/office:document-content/office:body/office:text/text:list/text:list-item/text:list/text:list-item/text:list/text:list-item/text:list/text:list-item/text:list/text:list-item/text:list/text:list-item/text:list/text:list-item/text:list/text:list-item/text:list/text:list-item[1]/text /text:span
How does a character selection work? [16:33] Patrick: /text)[9]/text:span[1] [16:34] Patrick: (doc("content.xml")//text:p)[9]/text:span[1][16:37] Patrick: string(doc)("content.xml")//text:p)[9]/text:span[1][16:38] Svante Schubert: If span has multiple character or multiple span the above example gets more complicated for XPath
[16:39] Svante Schubert: And if I want to insert a table:table and there are other components like images there would be an ambiguity, righ? [16:40] Svante Schubert: like paragraph, image, image and paragraph and we want to insert inbetween the images we need to use an xpath with images [16:41] Patrick: <text:p><image><text:p> [16:43] Svante Schubert: OT gets more complicated. If a paragraph becomes a heading it would influence other XPath, right? [16:45] Patrick: /text:p)[9]/text:span[1] [16:47] Svante Schubert: The advantage of your solution
is inde
ed that the referencing is working on other XPath implementations. [16:50] Svante Schubert: The number referencing I have proposed is the least common denominator abstracting from XML (even ODF & OOXML in my mail example) and as well application models. [16:53] Svante Schubert: Patrick: As long we are an XML standard, we should use XPath as XML referencing model. [16:54] Svante Schubert: Patrick: As it is defined already and used often [16:57] Svante Schubert: Svante: Patrick's approach would work when using this subset. [16:58] Svante Schubert: Patrick: Did not want to
use XPat
h in its full glory, but only a small subset [17:00] Svante Schubert: Patrick, the notation with numbers only is just a abbreviation of XPath and can be expanded.. [17:02] Svante Schubert: well, I guess it is more, as if there is an image with multiple alternative images it is still only one image component. So you can not count draw:image [17:02] Svante Schubert: The abbreviation for //*[text:p | table:table | .. ][1]/ is /1 [17:02] Svante Schubert: You can count draw:frame as there are other frames, which are not images [17:05] Svante Schubert: I guess you can address ever
y part o
f an XML document with XPath [17:06] Svante Schubert: It is only not easy readable and interpretable by humans and applications not knowing Xpath so far [17:07] Svante Schubert: ODF Style inheritance might be problematic [17:09] Svante Schubert: What about we are using an abbreviation for the serialization and you are defining the XPath in the specification? [17:09] Svante Schubert: I can not remember: Is the XPath Whitespace normalizing compatible with the one from ODF? [17:10] Svante Schubert: How do you map it? [17:10] Svante Schubert: Patrick: Best way is not to use character counting, but use regex expressions [17:1
1] <
b>Svante Schubert: Remember that we should be able to alternate the order of operations and adopt operations to keep creating the same document. [17:23] Svante Schubert: -- [17:23] Svante Schubert: Peter: It is not very convenient to merge changes of different documents, when only using UNDO changes. [17:25] Svante Schubert: We are using the UNDO operations as we only tweaked the existing ODF application methods using XML before state. [17:26] Svante Schubert: Peter: The downside of the existing approach you always have to read the content and the operations. [17:28] Svante Schubert: The design choice was to stick close existing applications model and to not save redundant DO operations. [17:29] Svante Schubert: The idea was that Merge is far less being used to make this necessary. But it is true it is no
t very c
onvenient to the merge szenario. [17:30] Svante Schubert: More get more complexity by keeping DO and UNDO stack in synch. [17:31] Svante Schubert: The goal was to keep the design as simple as possible.. [17:32] Svante Schubert: Peter: Only wants not UNDO & DO, but only DO [17:32] Svante Schubert: Than we have the compatibility problem that the content is not representing the latest state. [17:34] Svante Schubert: The next meeting will be at the 10th of February. |