OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

office-collab message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: IRC log from today's meeting - 2013-05-08

Please find the IRC log of today's meeting below.
Our next meeting will be in about two weeks on the 22th of May.



Attendees: John Haug, Jos van den Oever, Patrick Durusau, Thorsten, Svante.

[15:30] Svante Schubert: Hi 
[15:34] Patrick Durusau: Dialing in now
[15:34] Thorsten: hi there!
[15:36] Svante Schubert: I have been working on prototyping ODF operations, take a look at https://www.ox.io/ox_text
[15:37] Svante Schubert: ^^The reason why I have specified so little in the past weeks, as they are working on the first release.
[15:38] Svante Schubert: In opposite to a former browser office from Sun, where we send HTML from the server to the client, here operations are send via JSON among the participants.
[15:39] Jos van den Oever: no, not simultaneous editing atm
[15:39] Jos van den Oever: webodf prototype does do that btw, but is limited in # of operations atm
[15:39] Patrick Durusau: Is this synchronous exchange of edits?
[15:39] Jos van den Oever: http://git.open-xchange.com/
[15:40] Svante Schubert: @Patrick: Not in the first versions. atm they are using something like a "stick" a user gets. Only the one with the stick is allowed to do changes.
[15:41] Jos van den Oever: example of testing operations in webodf http://gitorious.org/webodf/webodf/blobs/master/webodf/tests/ops/operationtests.xml
[15:41] Jos van den Oever: describe before, operations, and after
[15:41] Jos van den Oever: the spec could take same approach
[15:41] Patrick Durusau: So there is a continuum of 1) who has the stick, 2) synchronous exchange of edits, 3) asynchronous tracking of edits, Yes?
[15:42] Svante Schubert: @Jos: Exactly that is a great principle for testing!
[15:42] Jos van den Oever: only the implementors complain about all the tests 
[15:42] Svante Schubert: @Patrick: AFAIK it is not planned for the first shoot, but on the list for later results.
[15:49] Svante Schubert: Jos: WebODF is using operations as well
[15:52] Patrick Durusau: Are the operations described against ODF markup or against an abstraction of the ODF markup?
[15:53] Jos van den Oever: in webodf, operations are against ODF
[15:54] Jos van den Oever: it would be useful to use a computer parsable set of examples of operations so they can be reused in test code easily
[15:54] Patrick Durusau: Working against markup though prevents cross-application tracking of changes
[15:54] Svante Schubert: @Patrick: Against the markup, like insertParagraph would be the insertion of a <text/>
[15:55] Patrick Durusau: B/c <p> in an ODF editor isn't the same <p> in an HTML editor, different structure
[15:55] Svante Schubert: We should add, that we would define our change on a normalized ODF, like namespace prefix, etc. is irrelevant.
[15:56] Jos van den Oever: @Patrick: yes, but we are speccing for odf, right?
[15:56] Jos van den Oever: our editor also contains odf, the document is not html, the changes we send apply on odf only
[15:56] Jos van den Oever: to let them work on odf and html (and other formats) would be really hard and limiting
[15:57] Svante Schubert: I agree, we (the ODF TC/SC) would specify operations against ODF XML
[15:59] Patrick Durusau: Disagree, gently, because an abstract grove model (see mail earlier today) is easier to specify the operations. When Svante talks about normalized ODF, he is choosing what I would call grove components to be counted for locating changes. This may be a question of terminology.
[15:59] Svante Schubert: Nevertheless we should have an eye on the other formats, but our first change-tracking operation set is likely to work well in all formats
[15:59] Thorsten: yup, agreeing with patrick - a conceptual step above raw odf xml
[16:00] Jos van den Oever: i'm fine with that if we also include the step from intermediate to raw odf xml
[16:00] Svante Schubert: Yes indeed a component (a bundle of one or more XML elements) is an abstraction from the XML
[16:01] Svante Schubert: Yes, I skipped on step. There is an abstraction.
[16:03] Svante Schubert: It seems we have a full consensus here!
[16:04] Svante Schubert: We do not start to define the operations directly on the XML, but on components.
[16:04] Jos van den Oever: If we specify in terms of Grove, that would mean we specify a kind of runtime model, right?
[16:04] Svante Schubert: All possible ODF DOMs are already defined by the ODF RelaxNG Schema
[16:05] Jos van den Oever: Similar to the way XML has a DOM.
[16:06] Patrick Durusau: @Jos, I would say no. The grove specifies the abstract operations. We map that to the required XML output. How your runtime is only required to produce the defined output. How it gets there, runtime model, isn't our issue.
[16:07] Patrick Durusau: Yes, the model reflects the XML model.
[16:07] Jos van den Oever: @Patrick: so it's like a combined xml infoset or a combined dom of the documents in the zip?
[16:08] Patrick Durusau: @Jos, I would have to think about "combined dom." Groves can be a subset of the features we want to track for change tracking purposes.
[16:08] Jos van den Oever: perhaps when we start writing the operations, we can exchange comments to improve the operations
[16:09] Patrick Durusau: @Jos, +1!
[16:09] Jos van den Oever: to get the subset part, i'll have to read the papers first
[16:09] Svante Schubert: yes of course, Jos!
[16:09] Patrick Durusau: We don't track every change to the markup.
[16:10] Jos van den Oever: the example i pasted above is a bit deceptive since it's a file which of course is markup 
[16:11] Svante Schubert: @Patrick: We could if we wanted to, but the cost/benefit ratio might not be good, as some changes are used very seldom
[16:11] Jos van den Oever: it does not compare strings to determine if the tests passes, but odf
[16:13] Patrick Durusau: We do have a wiki, yes?
[16:13] Jos van den Oever: ah, where is it?
[16:17] Svante Schubert: byebye
[16:17] Jos van den Oever: tx&bye

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]