I don’t yet have a firm opinion on whether we should try
to fix the conref push feature by adding a new attribute or whether we should drop
the conref push feature entirely. However, I’d like to point out that conref
push is just one of the features that depends on the new attribute. If I
understand the specs correctly, the new attribute is also needed in order to
use the keyref feature and the anchor/anchorref feature in all the ways they are
intended to be used.
Su-Laine
Su-Laine Yeo
Interaction Design Specialist
JustSystems Canada, Inc.
Office: 778-327-6356
syeo@justsystems.com
http://na.justsystems.com
From: Michael Priestley
[mailto:mpriestl@ca.ibm.com]
Sent: Wednesday, February 04, 2009 7:32 AM
To: Grosso, Paul
Cc: dita@lists.oasis-open.org
Subject: RE: [dita] Map structure and possible new attribute for conref
push and anchorref
Hi Paul,
I agree this is
an important change, and it's unfortunate it wasn't caught before. As the
original proposer of the feature, I thought the TOC attribute would be
sufficient. But I think Su-Laine makes a good case for creating a new
attribute.
Having
discovered a shortcoming in the feature, we do indeed have two choices:
- fix it
- remove it
We haven't had
the discussion about removing it, but you're right that we should have that
discussion. However, I don't think removing the entire feature (as planned and
as already partially implemented by some) is necesarily less disruptive than
adding one new attribute with clear semantics and general usefulness.
Michael
Priestley, Senior Technical Staff Member (STSM)
Lead IBM DITA Architect
mpriestl@ca.ibm.com
http://dita.xml.org/blog/25
Sorry I just noticed this, but I was quite
surprised
to see we're talking about a major change/addition to
DITA 1.2. I thought we had closed 1.2 to new things
quite some time ago.
How did the discussion go that had us deciding we could
change our minds at this late date and add something new
to DITA 1.2?
If this is "just" part of the conref push proposal, then
why did we just think of it? The conref push proposal
was supposed to be completed months ago. Perhaps this is
evidence that conref push is not ready for DITA 1.2.
paul
>
Original Message-----
> From: Su-Laine Yeo [mailto:su-laine.yeo@justsystems.com]
> Sent: Tuesday, 2009 February 03 13:42
> To: dita@lists.oasis-open.org
> Subject: RE: [dita] Map structure and possible new attribute
> for conref push and anchorref
>
> Hi everyone,
>
> In this morning's TC call we discussed the requirement to be
> able to say
> "these topics need to be processed but do not contribute
> directly to any
> output result". I said I would follow up by writing a proposal
for
> fulfilling this requirement, that the group could discuss and
> then vote
> on. This is a proposed substantive change to the 1.2 specification:
>
> Overview:
> The discussion this morning centered on fulfilling this requirement by
> adding a new attribute to the topicref attribute set. The new
> attribute
> could be called "localdisplay" and would function in an
> analogous way to
> the toc, print, and search attributes. There are other
> possible ways to
> fulfill the requirement, such as the one suggested by Eliot
> in a branch
> of this thread, however the idea of a new "localdisplay"
> attribute seems
> like a clean and reasonable approach, so I will expand on it here.
>
> Proposed scope:
> The new attribute should be added to the %topicref-atts; and
> %topicref-atts-no-toc; attribute sets described here:
> http://docs.oasis-open.org/dita/v1.1/CD02/langspec/common/topi
> cref-atts.
> html
>
> Proposed name:
> "localdisplay". In previous discussions I've suggested
a name of
> "appear" or "display" but "localdisplay"
is now my favourite (I think
> Gershon suggested it).
>
> Proposed description:
> Specifies whether the topic should be displayed in output at the
> location of the topic reference. For example, if a topic is
> not intended
> to be used directly in output, but must be included in a map
> because it
> contains content that is to be transcluded into another topic via the
> conref push mechanism, set the localdisplay attribute to "no"
> to prevent
> the topic from appearing twice in output. If the value is not
> specified
> locally, but it specified on an ancestor, it will inherit the value of
> the ancestor.
>
> yes
> Include the topic at the location of the topicref
>
> no
> Do not include the topic at the location of the topicref
>
> -dita-use-conref-target
> See Using the -dita-use-conref-target value for more
information.
>
> Usage example:
> Continuing on my example from below, the author of Folder2
> can create a
> third map called masterbuild.ditamap, with contents as follows:
>
> <map>
>
<topicref href = "Folder1/default.ditamap" format =
"ditamap"/>
>
<topicref href = "Folder2/pushmap.ditamap" format
="ditamap"
> localdisplay ="no"/>
> </map>
>
> The author of Folder2 and of masterbuild.ditamap can then process
> masterbuild.ditamap to create deliverables. Any topics referenced by
> pushmap.ditamap will not appear at the end of a deliverable,
> however the
> contents of the pushed topics will appear in appropriate locations
> within the topics that are referenced by defaultmap.ditamap.
>
> Regards,
> Su-Laine
>
> Su-Laine Yeo
> Interaction Design Specialist
>
> JustSystems Canada, Inc.
> Office: 778-327-6356
> syeo@justsystems.com
> http://na.justsystems.com
>
>
>
>
>
Original Message-----
> From: Su-Laine Yeo [mailto:su-laine.yeo@justsystems.com]
> Sent: Wednesday, January 21, 2009 12:10 PM
> To: dita@lists.oasis-open.org
> Subject: [dita] Map structure and possible new attribute for
> conref push
> and anchorref
>
> From the Jan 20, 2009 minutes:
>
> 8. New ITEM: "Conref push for static output formats"
discussion
> * http://lists.oasis-open.org/archives/dita/200812/msg00006.html
> * http://lists.oasis-open.org/archives/dita/200812/msg00007.html
> Discussion from Su-Laine Yeo and Michael Priestley ensued. Michael
> clarified that there would be a master build map (owned by the
> information architect for the product) that the processor would call.
> How to handle content that should not be displayed anywhere? Su-Laine
> stated that this is a problem for anchorref also. Possibly need
a new
> attribute. Discussion to follow via e-mail.
>
> ------------------
>
> Here is a summary of the overall user goal we discussed in
> the meeting:
>
>
> Using this example:
> - Folder1 containing a file called defaultmap.ditamap and a
> file called
> default.dita. The map contains a topicref to default.dita.
> - Folder2 containing a file called pushmap.ditamap and a file called
> push.dita, where the push.dita topic has a <step> to be pushed
into a
> set of steps in default.dita. The map contains a topicref to
> push.dita.
>
> We want the author of the Folder1 contents to not have to
> know anything
> about the contents of Folder2. The author of Folder2 must be able to
> produce a deliverable that resolves conrefs pushed from push.dita into
> default.dita. One way for the author of Folder2 to do this is
> to create
> a third map called masterbuild.ditamap, with contents as follows:
>
> <map>
>
<topicref href = "Folder1/default.ditamap" format =
"ditamap/>
>
<topicref href = "Folder2/pushmap.ditamap" format
="ditamap"/>
> </map>
>
> The author of Folder2 and of masterbuild.ditamap can then process
> masterbuild.ditamap to create deliverables.
>
> A problem with using master build maps is that, given current
> processing
> rules, processing masterbuild.ditamap, the contents of pushmap.ditamap
> would appear in output. Using the toc attribute will not solve the
> problem as the toc attribute suppresses topics from the table of
> contents but not the body of the deliverable. In PDF output
> the contents
> of pushmap.ditamap would appear to be dumped at the end of
> the document.
> Also, the toc attribute does not necessarily prevent the topics of
> pushmap.ditamap from appearing twice in search results or indexes, as
> the behaviour of the toc attribute with respect to search and
> indexes is
> undefined. Therefore, a new @appear attribute which would completely
> suppress pushmap.ditamap from the toc, the body of the document, and
> other access points such as search and index, would be
> useful. There are
> other ways to solve this problem, but a new attribute is one obvious
> solution.
>
> I briefly mentioned in the meeting that we have a similar problem with
> anchorref. To explain further, the <anchorref> element proposed
here:
> http://www.oasis-open.org/committees/download.php/26092/IssueM
> apConvenie
> nces12047.html is similar to conref push in that it is also
> intended to
> allow pushing of content from one independent content set
> into another.
> A master build map would also be needed with topicrefs to both the
> default map containing the <anchor> and the pushing map
containing the
> <anchorref>.
>
> The <anchorref> proposal currently says, " The copied child
> elements can
> be suppressed in the current context by setting the toc
> attribute on the
> child elements to "yes" and on the <anchorref> element
to "no"." As I
> explain above, setting a toc attribute would not suppress
> anything from
> the current context, so I think the @appear attribute is what
> we'd need
> on this as well.
>
> As an aside, I am not sure why we need to add a new
> <anchorref> element
> when it seems to do the same thing as conref push.
>
> Regards,
> Su-Laine
>
> Su-Laine Yeo
> Interaction Design Specialist
>
> JustSystems Canada, Inc.
> Office: 778-327-6356
> syeo@justsystems.com
> http://na.justsystems.com
>
>
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe from this mail list, you must leave the OASIS TC that
> generates this mail. Follow this link to all your TCs in OASIS
at:
> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgr
> oups.php
>
>
> ---------------------------------------------------------------------
> To unsubscribe from this mail list, you must leave the OASIS TC that
> generates this mail. Follow this link to all your TCs in OASIS
at:
> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgr
> oups.php
>
>
---------------------------------------------------------------------
To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail. Follow this link to all your TCs in OASIS at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php