W3C home > Mailing lists > Public > www-style@w3.org > May 2012

[CSSWG] Minutes and Resolutions Hamburg F2F 2012-05-11 Part II: XSLFO and CSS3 Layout

From: fantasai <fantasai.lists@inkedblade.net>
Date: Tue, 15 May 2012 17:38:14 -0700
Message-ID: <4FB2F6F6.9090501@inkedblade.net>
To: "www-style@w3.org" <www-style@w3.org>

   - Liam Quin presented an overview of XSL:FO and proposed folding
     the remainder of the XSL:FO WG into CSS

   - Bert Bos presented on a collection of various layout challenges and
     how they can or cannot be solved with proposals in the CSSWG

   - Steve Zilles started a discussion on harmonizing the various CSS3
     layout proposals

====== Full minutes below ======

XSLFO Overview

Scribe: vhardy

   <Liam> http://www.w3.org/2012/Talks/05-quin-paged-media/all.txt
   liam: thanks for giving me a chance to present.
   liam: XML activity lead at W3C
   liam: chair XML print and page layout WG
   liam: informal presentation. Not trying to persuade the group. Want to
         share background about how XSL came to what it is.
   liam: very brief comparison between CSS and XSL
   liam: brief notes on paged media requirements.
   liam: the big question is how do we go forward and what the overall goal is.
   liam: want great document formatting.
   liam: As we commoditize books, we should be able to do them with HTML and
         CSS. We should be able to do all of 70% of book formatting.
   liam: I do not want to be able to do 70% of each book. We want to do the
         entirety of 70% of the books.

   liam: SGML was created for publishing, became a <s>Rec</s> ISO standard
         in 1986
   liam:  every documentation on parts for automobile, cruise ships, airplanes,
          plants etc., are done in SGML
   liam: 3 parts of SGML on the web: Xlink, XML, XSL
   liam: won't talk about Xlink (dead), and you know about XML

   liam: Footnotes, back of the book index are important requirements and
   liam: The number of publishers using XSL-FO is increasing right now. They
         are not moving to CSS+HTML yet because there are things like
         back-of-the-book-index that are not available yet.
   liam: publishers are waiting for XSL-FO 2.0 but they are not joining the WG.
   liam: the work on XSL FO 2.0 has stopped, but people want the features.
   liam: how can we do this in HTML + CSS.

   Håkon: I have been publishing books in css since 2005.
   liam: yes, some people do that, but it is often done with extensions.
   fantasai: not everything is in the WG drafts, but a lot of it is.
   szilles: you are trying to help us identify features that, from your
            experience, a large user community would like. Some of the
            issues may have been solved, some not.
   liam: yes. The bigger problem is what should we do? I am thinking of
         stopping XSL-FO WG if we can agree that print is enough of a
         priority and we can address the needs with CSS+XHTML.
   liam: the main two possibilities are: flow becomes HTML and page masters
         and flow maps stay in XML. Or, some of the later part goes to CSS
         (e.g., regions, exclusions).
   liam: we are close to closing the XSL WG. The XSL FO WG is going to be
         closed in the next 6 months. There is a community group. The
         future for those users is CSS, for most of them.
   jdagget: if you see things that are needed in CSS, please point out what
            these are.
   liam: I have done some of that, more work is needed.
   tab: would be good to have your list and be able to tell you what is
        covered and what is not.

   liam: publishers take technology and do things with it. But they do not
         create the technology.
   liam: printers make innovation. You cannot get publishers in the WG.
   glazou: I think it would be good to have you in this WG and give input
           if what we are bringing to css is not in-line with what that
           user base wants.
   glazou: it is a new industry domain that HTML+CSS could address. Seems
           like an area of growth.
   glazou: Each time we touch a domain you have already seen in XSL-FO,
           please tell us if we are falling in traps.
   glazou: let us know if there is a better design you know of for some
   glazou: help us improve CSS.
   szilles: while XSL-FO was not conceived for browsers, it did look into
            things like progressive rendering.
   glazou: if you saw things during this meeting that you would like to
           comment on, please do it asap.
   liam: it is hard to catch up with all the current work.

   liam: Is it a sensible goal to replace XSL-FO with HTML+CSS?
   Håkon: I think it makes sense to publish books in HTML+CSS.
   Håkon: you should use at the GCPM specification.
   liam: how to we move on print media quickly so that it is useful for ePub.
   glazou: I think you should start with a requirements document for this
           working group.
   dbaron: don't assume that everything in GCPM is complete.
   liam: we spent 3 years talking about regions and exclusions.
   plinss: liam has a meta-questions. Print has not been addressed properly
           in css. What is the path forward? For him to fold things into
           CSS+HTML, he needs a commitment that we are going to work on that.
   plinss: GCPM has not gotten the required attention and level of completeness.
   plinss: I want GCPM driven to REC and not be a placeholder. Should be a
           formal priority for the group and be implemented in browsers.
   plinss: Liam wants a commitment from the group to take on that print focus.
   (comments about how recent contributions are already in the area of
    pagination content).

   szilles: paginated content is not just for printing. It really is about
            pagination. Let's focus on that piece and printing will fall
            out of that.
   liam: mostly, but there are a few things that are specific to printing.
   liam: if we get the commitment, we can talk about the mechanics to do
         something together on this. Separate or same list?
   glazou: I would start with a list of requirements, then get traction
           and people will contribute.
   tab: could you start looking at what we have and ask us where we are at
        with each features.
   Håkon: this is a great initiative, you have a lot of expertise. My
          concern is that I would like this to be a CSS work item.
   liam: it would be.
   glazou agrees.
   liam: I would like to do all of books with what the CSS WG specifies.

   glenn: I have worked on the XSL FO for years, but I have not seen calls
          to use HTML+CSS. People are happy using it as is. Some people
          are going to work with the XPP community. I want to express that
          this is not exactly what Liam has been saying.
   liam: yes, some people are going to stick to XLS FO. But as browsers
         are commodatizing good printing, then it is worth changing.
   jdaggett: we should talk about pagination, not printing.
   glenn: I think that there are people willing to work on the next version
          of XSL FO. I would like to see work proceed in both independently
          and see cross affiliation as possible. There are different focuses
          and history in both communities.
   glenn: maybe they'll meet in the same place.
   Håkon: I do not have an opinion on closing XSL-FO or not.
   liam: the wg currently only has 2 members. We cannot fund that work.
   glazou: contributions and input from you on our spec. that touch on things
           you have worked on in FO, that would help. If you see things that
           we could add, please let us know too.
   plinss: we could also invite the two XSL FO invited experts to join the
           CSS WG.

   florian: I think this group cares about pagination.
   plinss: I want to see an increased level of commitment from the rest
           of the group.
   jdaggett: I do not think we can do that.
   plinss: we as a group set our priorities.
   tab: we are clearly committed to this, esp. with Adobe's increased
   plinss: we are bringing the rest of the expertise and folding it into css.
   glazou: can you invite your experts to join the css wg as invited experts.
   liam: yes, and I'll put together requirements.
   liam: I can answer questions on XSL.

   szilles: typically, there is a portion of the group that is not following
            some of the discussions. One of the possible solutions to this
            problem is to create a task force. The FX task force is an example.
            We founded a group to focus on a specific set of issues. The role
            of a task force is to develop proposals.
   szilles: we work faster if there are concrete proposals on the table.
   szilles: this may make a suitable task force.
   szilles: just as a way to make additional progress.
   jdaggett: the other alternative is to use the mailing list more.
   (discussion about next agenda items)

Complex Layout: Some Challenges

   <Bert> -> http://www.w3.org/Talks/2012/0509-CSS-ftf/
          Slides on complex layout and grid templates
   (bert presenting)
   bert: I will show you that GCPM is not enough. There is still a lot to do.
   bert: I have looked at common patterns that are not possible yet. Looking
         at where the problems are.
   bert: I believe the solution passes by grids.
   bert: they are a natural way to do things. easy to learn. I'd like to
         explain the latest editors of the grid template.
   bert: I would like to do as much as possible with css as long as it is
         easy to do.
   bert: e.g., Table of Content easier not done in CSS
   (slides on CSS processing model)

   (templates - basic ideas)
   vhardy: can cells in your grid overlap?
   bert: yes.

   (templates - second idea)
   (challenge 1: a simple grid?) slide
   (discussion about using a float:bottom for the last paragraph to be
    bottom aligned)
   bert: other examples where column content is 'justified' vertically.
   fantasai: could use a flex box in the template.
   bert: yes.
   bert shows the signature as a reverse run-in.
   vhardy: there is a proposal for that in the regions spec.
   (discussion about 'after' content that needs to reverse run-in)
   (challenge 1: HTML source) slide.
   bert: In the visual display, the images come before the title, but they
         come after the title in the logical order.
   (discussing about how aligning images at the top)
   (challenge 1: CSS rules) slide
   (challenge 1: other solutions?) slide

   (challenge 2: column span)
   (bert shows how all the content aligns with respect to the page bottom)
   bert: page number at the top: the number height is used to size the
   bert: we need new units to size content like this.
   tab: may be having text measure functions could be used as well.
   jdaggett: we already have that with canvas.
   szilles: that is not css.
   bert: the kerning in the running header is not easy to do.
   several: might be done with exclusions.
   jdaggett: I do not think there are automatic/scalable ways to do that.
   (challenge 2: CSS rules) slides
   (challenge 2: other solutions?) slide
   you do need GCPM top float.
   szilles: in XSL, float: top will make it automatically go to the next
            'top' because the float cannot precede the call-out
   stearns: this is where you may want different page templates depending
            on the number of columns.

   (challenge 4: centered floating images) slide
   bert: the floats are not in their logical position in the text.
   bert: but they still line up. They are not floats. Exclusions might be used or regions.
   bert: there are other ways.

   (challenge 5: mixed horizontal & vertical) slide

   (challenge 6: manuscript) slide
   bert: there are several flows.
   bert: 3, interleaved: one around the outside, two side-by-side inside.
   bert: There are drop caps which extend between regions.
   liam: this is still done in publication.
   liam: today.
   bert: not sure how to best model this.
   stearns: we can do that with regions and exclusions.
   (discussion about different solutions)

   (challenge 7: seven different sets of slots)
   jdaggett: there are multiple examples of grids where the axis of the
             grids are rotated and not aligned with other grids. If you
             look into some of the Russian constructionism, you'll find
             these designs.
   stearns: is there any limitation to a grid item that prevents it to
            being transformed?
   (discussion on that topic).
   bert: would be good to do rectangular grids easily.

Complex Layout: Pulling it all together

   (Grid Template Layout - summary, May 2012 1/2) slide.
   bert: flow property, equivalent to flow-into in regions spec.
   bert: grid cells can overlap. If you want an exclusion behavior, then
         you need something like exclusions.

   (Grid Template Layout - summary May 2012 2/2) slide.
   bert: @template looks a lot like @page
   bert: chaining slots together is like in region.
   bert: however, the chain defines the region ordering explicitly.
   (discussion about if it is best to move a region to a chain or declare
    the chain specifically)

   (Grid Template Design) slide
   szilles: I observe that in this particular area, there seem to be that
            GCPM, regions, templates etc… solve parts of the problem but
            not the whole problem very well.
   szilles: proponents seem stuck on their particular solution and
            unwilling to bring things together.
   szilles: there are efforts in that particular direction, but I would
            like to propose that we set-up a week of dedicated to do
            education between the people pushing the different proposals
            and see if we can find a way to combine the pieces and put
            them together in a complete solution.
   szilles: grid, template, regions, exclusions, flex box, ... all the
            layout specs.
   szilles; + regions and exclusions.
   jdaggett: are you saying they should be unified or aligned?
   szilles: aligned at least, unified at best.
   szilles: this comment is triggered by having listened to the different
   glazou: this is a general comment, not on bert's presentation in particular.
   szilles: we are not making progress if we are attached to particular
   dbaron: we are going to need multiple tools. I don't think a single tool.
   szilles: I agree, but I'd need a set of pieces that I can put together.
   fantasai: understanding how things work together makes sense.
   jdaggett: in some cases, I hear things overlap enough that they should
             be unified.

   Håkon: the fundamental issues is that the problem we are trying to
          resolve is very large. We need to agree on the subset we address.
   Håkon: I have talked about the use cases we want to address.
   Håkon: scans are good, but resizing needs to be addressed.
   florian: I am not sure scans will let us solve the issue.
   Håkon: we could see which use cases we want to address.
   glazou: the problem with the 20 use cases is that you'll need a week
           to agree on the use cases.
   ted: yesterday, during the discussion about improving our working mode,
        I think we found there are different expertise areas.
   ted: the different editors could meet on a dedicated layout telecon
        for example.
   fantasai: I found that spending a couple days with Bert was useful.
   fantasai: I did the same with Microsoft.
   fantasai: I am trying to align the grid module and the template specs.
   ted: this is nice, but we need more general coordination.
   florian: having use cases is nice, and we should see how to solve the
            issue in general, not just compare solutions.
   glazou: I have a problem with use cases when we collect them ourselves.
           I would prefer use cases from users.
   glazou: apple is doing magazines, Opera does things with publishing etc...
   glazou: these people have live cases that do not reach us.
   glazou: e.g., Wired magazine are on the web.
   arno: this is after talking with users that we brought regions/exclusions
         to the css wg.
   Håkon: we know where to go to scan Wired for example.
   glazou: this won't tell us what wired needs the most.
   arno: Wired is a good example because they have complex print layouts.
         They want to translate their layouts to digital form.
   arno: They understand there is a difference.
   arno: They understand they need adaptive layout.
   arno: Their challenge is to find the tools to express that.
   arno: Looking at the printed page is not ideal because the goal is not
         to replicate the printed page
   arno: This is the way the designers look at adaptation that is important.
   krit: The use cases change over time too.
   glazou: it is related to fashion, yes.
   glazou: goal not to replicate what's on paper, but replicate the coolness
           of the magazine
   stearns: when I was collecting examples, people sent me a printed scan
            and the online version which had different layouts.
   stearns: they are interested to take the layouts on-line.
   Håkon: there is the 95% cases that we need to address.

   fantasai: with regards to the template and grid module, several of us
             would like to discuss this in a task force. It would be nice
             that anyone who wants to be involved can be involved.
   fantasai: right now it is bert, myself, dholbert, phil
   glazou: I am interested.

<br type=lunch>
Received on Wednesday, 16 May 2012 00:38:47 UTC

This archive was generated by hypermail 2.4.0 : Monday, 23 January 2023 02:14:15 UTC