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


Help: OASIS Mailing Lists Help | MarkMail Help

dita message

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

Subject: Groups - DITA TC Meeting Minutes 11 June 2019 uploaded

Submitter's message
ACTION: Alan to set up DITAweb accounts for Zoe and Joyce

Minutes of the OASIS DITA TC
Tuesday, 11 June 2019
Recorded by Tom Magliery
link to agenda for this meeting:

11 AM ET open

1. Roll call
Attendance: Zoe Lawson, Kris Eberlein, Biil Burns, Robert Anderson, Stan Doherty, Alan Houser, Tom Magliery, Chris Nitchie, Scott Hudson, Joyce Lam, Jim Tivy

Regrets: Deb Bissantz, Dawn Stevens, Carsten Brennecke, Carlos Evia, Eliot Kimber, Eric Sirois, Nancy Harrison

2. Approve minutes from previous business meeting:
04 June 2019
https://lists.oasis-open.org/archives/dita/201906/msg00024.html (Harrison, 10 June 2019)
Moved by Kris, 2nd by Scott, approved by TC

3. Announcements:

4. Action items
(No new updates on any action items. Kris encouraged everyone to PLEASE try to make some progress on these items.)

5. DITA 2.0 stage one proposals

outputclass for DITAVAL properties
https://lists.oasis-open.org/archives/dita/201906/msg00004.html (Anderson, 04 June 2019)
Robert: This came out of me chatting with Jarno about DITA OT stuff. He said why do you have all these very specific styles but just don't give me a hook for my own CSS? We do that for everything with outputclass so it seemed obvious to do it with DITAVAL. Idea is to allow @outputclass on properties (prop) and revision properties (revprop). The outputclass gets applied to everything that matches the properties. It makes things more flexible and stable. In Jarno's mind we should deprecate all the other values, but I'm less sure about that because not everyone can do CSS.
Tom: I didn't really remember/know that you could do these things before.
Robert: Remember we expanded the styling to include all the stuff in the highlight domain in 1.3
Tom: Ah yes, I do remember when we implemented that in XMetaL.
Robert: This gives people a lot of flexibility at low cost
Scott: I really like the idea to use CSS
Chris: What about a mechanism for arbitrary CSS in the DITAVAL? I realize that might leave out some PDF processors. What about kind of like a foreign element for "inline styles go here"?
Robert: I'm worried about overcomplicating things
Chris: Most cases it'd be a block for CSS property set. The goal is to make the styling more flexible.
Robert: The goal here is heavily tied to CSS but largely because that's how most people use outputclass. You can use outputclass today to hang any kind of styles on elements. It doesn't have to be done with CSS. This update to outputclass would not have to be done with CSS.
Kris: We're not really envisioning to change what a DITAVAL is designed for (filtering, flagging). I would think that going to handle styling for inline elements is sort of expanding the scope. Or do you see that as part of flagging.
Chris: I see that as part of flagging.
Robert: In the end, it's adding a style. I'm less excited about that because it's not so simple.
Chris: It just seems like it might be nice to have have a general style mechanism
Robert: You'd get into processors having to validate CSS. Today some people already use outputclass for other styles. So there's no new processing. You just have to translate the token into an outputclass token.
Scott: I can still have a centralized way of applying styles across the document, even if the authors are using outputclass to identify things.
Kris: I think the only problem we have is the way that Chris raised about having an inline style
Chris: I'm busy and more or less convinced not to bother with it
Kris: Motion that we advance this proposal to stage 2
Scott: 2nd
(no objections)

Changes to indexing
https://lists.oasis-open.org/archives/dita/201906/msg00017.html (Eberlein, 08 June 2019)
Kris: This came out of spec editing, I've been looking at element ref topics thinking about what content to move to more conceptual topics. It became clear that it'd be helpful to make changes to indexing. It's kind of messy because there was functionality added in 1.1 index-see and index-see-also, and had to create a new element index-base. I'm suggesting we move see and see also back into the base spec and get rid of the indexing domain. That includes getting rid of index-sort-as. It's duplicated by the sort-as element from 1.3. If we do that we can also get rid of index-base. So the proposal is to remove a domain and remove index-base.
Robert: I like this idea. I think the focus when we created index-base was "we don't want to add things to the base because it makes things too complicated". Then we added three new elements. For overall complexity it wasn't a good balance. index-base has been annoying to people. Particularly since sort-as was added, there's text in the spec that says it's equivalent to index-sort-as. Simplifying is good, gets rid of confusing and duplicate elements. From reading the indexing topics, it's really impossible to define how index processing works without treating them as part of the base.
Scott: For those people not wanting to use them it's still fairly easy to constrain out.
Robert: In fact it's exactly the same. You only need to constrain out indexterm, which removes everything inside of it.
Tom: Are we concerned with Eliot's point on the email list?
Kris: I wasn't quite sure what it referred to.
Robert: I'm not sure it was quite related to this proposal, but Kris's suggestion didn't remove any functionality. Eliot's suggestion looked like it added functionality.
Kris: Motion to advance the proposal to stage 2
Scott: 2nd
(no objections)

Allow in same locations as

https://lists.oasis-open.org/archives/dita/201906/msg00026.html (Hudson, 10 June 2019)
https://lists.oasis-open.org/archives/dita/201906/msg00027.html (Kimber, 10 June 2019)
https://lists.oasis-open.org/archives/dita/201906/msg00028.html (Eberlein, 10 June 2019)
Scott: One of the things I run into with my authors is the need to insert examples in different locations than the end of the document. They need something a little bit more block-oriented that can be used in more places. Examples are also useful in lists, sometimes tables, maybe as part of info in step, and maybe more. The current model is very limiting. I suggested div because it's allowed in many places, but more important is the ability to identify something as semantically an example in many more contexts. If it's too hard to maintain content-model wise maybe we could just add example to div.
Kris: I wish Eliot was here but since he isn't please open his email link.
Scott: Eliot suggested a new element examplediv. I'm not so much in favor of that if we could allow something that is already present as a block but remove some restrictions.
Kris: I think Eliot is saying example is currently logically a specialization of section, although not technically.
Robert: It's not a specialization, but in how it's defined in the spec it was always written that way. Definitely up through DITA 1.2 the grammar files said "For content model of example, use what's in section". I don't know whether it's that history that makes me blind here but it makes me cringe to think of example inside of div because I think of example as a large titled block. But does it need to be.
Tom: Eliot's suggestion would get around that
Robert: But then we have two elements for the same thing, something that we do a little grudgingly.
Tom: Do we have to hold our noses and change what example is?
Kris: That makes me feel wrong too. Scott I can understand wanting more semantic markup. Would we want to consider an example token (for @type) on note?
Stan: I support that, there are people doing that already on note with othertype.
Scott: For translation purposes I don't want it tied into an attribute
Zoe: One question I have is I've always wanted example in more places. In my brain section and example are the same thing. Having to put an example at the end of a topic is just weird. And what happens with stepxmp?
Kris: Scott you mentioned wanting to put example in info in tasks. That's what stepxmp is for.
Scott: Yeah I forgot that and we do use it.
Chris: If I were inventing example from scratch I'd model it more on figure than section.
Robert: Figure has a weird content model. Example is more free-flowing section content. Doing it with note feels "off". Doing it with examplediv seems "off".
Scott: We also need inline examples, so maybe we need a block one and an inline one.
Chris: Am I right that div is not titled
Robert: yes, and I definitely don't want to add a new titled element
Chris: given that we have an existing example element that is modeled after section what's the harm in adding it to more content models
Kris: The big problem is that if we make a titled element within section we'll give people the ability to do nesting that we didn't want
Chris: but an example within a section is a very common use case
Kris: I like the idea of a new untitled block element. Inline element makes me kind of queasy, though.
Chris: I'm not saying infinitely nested example. It's allowed within section. It's a one layer deep titled block. I suppose people could use it for 2nd level nesting, but the chances are low.
Robert: I think putting it in section is not going to give Scott what he wants. Then again you can also do that in figures. Nobody would think that's wise and your results might be garbage but you can do it. So that on its own isn't the reason to say we can't do that with example. We have titled things that you can make nested.
Tom: Aren't there both titled and untitled things.
Robert: There's no "best practice" in the spec. I don't think there's even a consistent best practice in usage.
Scott: What if we allowed example wherever figure and table are allowed. All that's doing is removing the restriction that when you use section or example, you're ONLY allowed to follow up with more of those.
Kris: You mean example as currently defined?
Scott: Yes. Even in a ref or refbody you have the same constraint.
Kris: I think we need to keep the focus not on concept/reference but on example as it works in topic. A lot of the things in regard to concept/reference really have to do with the content model of those.
Robert: It's just concept. In reference it's just like in topic: unless you have a constraint of your own you can use example more places. The restriction is that they can only be a child of body. Concept is the only one that has that restriction. It has confused poeple but that's tangent to this dicsucssion.
Zoe: I understand that constraints exist that it's supposed to be at the same level as section but I'm trying to think of where I'd use example and I'm getting lost. Do I want to tag "For example, blah" all the time? Sometimes I do but why does it have to be at the same level as a section? That hurts a lot of people's brains.
Scott: It makes it harder for reuse, too, because you wanted that example to stay with the section. You need two conrefs instead of one.
Robert: I think your core question is at what level is it wise/useful for people to mark up examples. If you have a sentence in the middle of a paragraph, you don't want to set that sentence as its own block, so maybe that's where Scott was thinking about an inline element. In my VERY limited experience that tends to be where you've got a codeblock and I wouldn't want people to change that to an example element.
Kris: In the spec we use otherprops=example on both some blocks and some inline phrases. We do that because have had requests to be able to filter out examples in the spec.
Robert: A lot of those we have that outputclass attribute on elements that have their own semantics.
Tom: Would blessing some kind of global attribute solve this whole problem?
Scott: the challenge would be what if we wanted a formal example mixed into the paragraph model.
Kris: Maybe we should step back and look at the requirements. I'm hearing the ability to clearly identify content that is an example, anywhere. Another one might be to generate some kind of styling or label. Am I missing any other use caes/requirements?
Scott: I think those are the main concerns. Semantic idenification and styling. And ease of re-use.
Scott: I really do like the idea of the importance attribute. That could help with semantic identification. I'd also like to be able to use example higher up in the conbody or refbody.
Kris: Where do we want to go with this? Clearly there is interest in addressing these use cases. But perhaps not consensus on how to do that. Shall we continue discussion next week, and in the meantime mull over our options.
Scott: the only limit on the discusson is I'm on vacation next week until July.
Kris: Ok, maybe we'll continue the discussion but reach no conclusions until you're back, since you'll likely be the champion of the proposal.
Scott: I'm interested in continuing discussion on the list as well.

6. DITA 2.0 editorial work
[skipped for time]

7. Draft comments in latest working draft from TC attention
[skipped for time]

8. DITA for 'To be named' version 2.0
[skipped for time]

9. Multimedia domain
Kris: Alan, are you able to set up a DITAweb rview for these topics?
Alan: Is the content ready to go in github?
Kris: I might need to publish something e.g. a PDF but I should be able to do that this week. Are you available later this week?
Alan: Yes, that should be fine.
Kris: Heads up to everyone, we'll open up this review soon, just for the multimedia topics, we do have some draft-comments in place. The idea is to do a review to get them locked down and able to be published in the committee note. We don't want the CN editors (Carlos, Bill) to have any questions about those topics.
ACTION: Alan to set up DITAweb accounts for Zoe and Joyce

11:57 ET close

-- Mr. Tom Magliery
Document Name: DITA TC Meeting Minutes 11 June 2019

No description provided.
Download Latest Revision
Public Download Link

Submitter: Mr. Tom Magliery
Group: OASIS Darwin Information Typing Architecture (DITA) TC
Folder: Meeting Notes
Date submitted: 2019-06-11 14:59:25

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