I have some unstated assumption that would become more obvious with a map context.
If each definition is in a separate topic, then assembling a glossary for a deliverable is just part of the usual map approach of selecting the appropriate content from a pool of available content:
<bookmap>
...
<glossarylist>
...
<glossentryref href=""ac-air.dita"/>
<glossentryref href=""ac-electric.dita"/>
</glossarylist>
...
</bookmap>
Different deliverables will have different maps, so it's easy for a deliverable to include either or both of the definitions for AC.
If both definitions are in the topic, then I need a separate topic for each permutation of the definitions for the term and need to share the definitions across topics by conref. In this case, that's three permutations (ac-air.dita, ac-electric.dita, and as-airelectric.dita).
<glossentry id="ac-airelectric.dita">
<glossterm id="term">AC</glossterm>
<glossdef id="air-def">Air conditioning</glossdef>
<glossdef id="electric-def">Air conditioning</glossdef>
</glossentry>
<glossentry id="ac-air.dita">
<glossterm conref="ac-airelectric.dita#ac-airelectric/term"/>
<glossdef id="ac-airelectric.dita#ac-airelectric/air-def"/>
</glossentry>
<glossentry id="ac-electric.dita">
<glossterm conref="ac-airelectric.dita#ac-airelectric/term"/>
<glossdef id="ac-airelectric.dita#ac-airelectric/electric-def"/>
</glossentry>
Of course, the number of permutations goes up with the number of definitions per term.
In other words, it's an example of the familiar tradeoffs between managing reuse with a map on topics as opposed to conref within topics.
Either way, I think people will want the ability to select the definitions that are pertinent to their deliverable.
Hoping that clarifies,
Erik Hennum
ehennum@us.ibm.com
"Jennifer Linton" <jennifer.linton@comtech-serv.com>
02/20/2006 09:20 AM |
|
Yes, that sounds good. But wouldn't we get the same amount of reuse if we did it this way:
<glossentry>
<glossterm>AC</glossterm>
<glossdef id="ac-air">Air conditioning</glossdef>
<glossdef id="ac-electric">Air conditioning</glossdef>
</glossentry>
This way people don't have to go through multiple glossary entries for the same term to figure out which one they want to produce.
Jen
Jen Linton
Comtech Services, Inc.
Senior Consultant and Web Manager
710 Kipling St. Suite 400
Denver, CO 80215
P: 303-232-7586
F: 303-232-0659
skype: jenlinton
From: Erik Hennum [mailto:ehennum@us.ibm.com]
Sent: Monday, February 20, 2006 10:14 AM
To: Jennifer Linton
Cc: dita@lists.oasis-open.org; Esrig, Bruce (Bruce); Paul Prescod
Subject: RE: [dita] Last call for top DITA issues
Hi, Jen:
I think it's a reuse issue. Some information deliverables might use both ac-air and ac-electric. Others might use one or the other alone. So, we're better off letting the process collate and format the selected definitions rather than having writers manually assemble each glossary entry.
Down the road, one could imagine a process that
1. Builds a list of terms based on occurrences of <term> within the content
2. Pulls the definitions for those terms from a pool of available glossary definitions
3. Collates the definitions by term
4. Formats the definitions for publication
Thanks,
Erik Hennum
ehennum@us.ibm.com
"Jennifer Linton" <jennifer.linton@comtech-serv.com>
02/20/2006 08:54 AM |
|
I may have missed this, but why would we not just have both definitions in the same entry as the term and apply ids or other clarifying metadata to each definition?
Jen
Jen Linton
Comtech Services, Inc.
Senior Consultant and Web Manager
710 Kipling St. Suite 400
Denver, CO 80215
P: 303-232-7586
F: 303-232-0659
skype: jenlinton
From: Erik Hennum [mailto:ehennum@us.ibm.com]
Sent: Monday, February 20, 2006 9:43 AM
To: Esrig, Bruce (Bruce)
Cc: dita@lists.oasis-open.org; Paul Prescod
Subject: RE: [dita] Last call for top DITA issues
Hi, Bruce:
My apologies for neglecting to clarify the term concerns -- I think the existing glossary proposal already handles both the case where the same term has multiple senses:
<glossentry id="ac-air">
<glossterm>AC</glossterm>
<glossdef>Air conditioning</glossdef>
</glossentry>
...
<glossentry id="ac-electric">
<glossterm>AC</glossterm>
<glossdef>Alternating current</glossdef>
</glossentry>
It would be up to the processing to collate by term and format definitions on output.
The proposal also includes a <glsynonym> element to specify synonym terms with the same meaning.
If one of the themes of DITA 1.1 / DITA 1.0.5 is enabling books, I'm wondering whether the shortdesc enhancements would be important for the narrative glue.
Hoping that's interesting,
Erik Hennum
ehennum@us.ibm.com
"Esrig, Bruce (Bruce)" <esrig@lucent.com>
02/20/2006 03:49 AM |
|
1. How about a simple generalization of task to allow tables and graphics in parallel with steps.
That would leave the following compatible extensions for the future:
- the div elements
- nested steps
2. For glossary, I'd still like to close on the issue of terms with multiple senses. This was raised privately with Erik Hennum earlier, but I'd be glad to discuss it more broadly in case it could easily cause incompatible changes if we decided to tackle it after a simple glossary mechanism was already fielded.
Specifically, we have cases where a project uses the same term in multiple senses, and being able to gloss the correct sense would be really helpful. For example: AC := i. Air Conditioning, ii. Alternating Current. It's just an abbreviation, but I can imagine it happening with terms as well. Quick, define "interface" in a way that is specific enough to indicate how you mean it in three interesting contexts and yet not ambiguous when glossed. Fun challenge, right?
The opposite also happens: multiple terms for a single sense. This might be easy for DITA to handle because of conref.
Best wishes,
Bruce
====
To merge the threads ... JoAnn Hackos wrote:
Which index-related issues would you prioritize?
- #44 Keep indextermref (or redefine its function)
- #45 Add See, See Also indexing elements
- #45a Add sort order indexing elements
- #45b Add page range indexing elements
I wouldn't leave the others out if we can manage it.
- -----Original Message-----
From: Erik Hennum [mailto:ehennum@us.ibm.com]
Sent: Friday, February 17, 2006 6:39 PM
To: Paul Prescod
Cc: dita@lists.oasis-open.org
Subject: Re: [dita] Last call for top DITA issues
Hi, Paul:
As Don was pointing out to me today, the <data> element presumably rides in on the coat tails of bookmap because the book metadata depends on it, but that should be okay because <data> is good to go (unless anyone has thought of ways to improve it while it has been sitting on ice).
Does glossary also ride in with bookmap as a requirement for having a complete book story?
Any other hangers on?
Thanks,
Erik Hennum
ehennum@us.ibm.com
"Paul Prescod" <paul.prescod@blastradius.com>
02/17/2006 03:22 PM |
|
So the current list of proposals is:
#20 Extensible metadata attributes
* customers ask for it and the current situation is just not
workable for many
#38 Bookmap / bkinfo revision
* in widespread quasi-standard use already
* often requested by customers
#35 Support foreign content vocabularies such as MathML and SVG
* MathML in particular is very popular, and has been hacked by
XMetaL customers
* graphic (image) scaling improvements (no issue number?)
Now is the time to jump in with any other DITA 1.1 issue that you
consider absolutely urgent and therefore a candidate for "early
release".
And then SOME STILL-TO-BE-RANKED SUBSET of indexing requirements:
- #44 Keep indextermref (or redefine its function)
- #45 Add See, See Also indexing elements
- #45a Add sort order indexing elements
- #45b Add page range indexing elements
I'll send another email about ranking the indexing requirements.
Paul Prescod
[attachment "graycol.gif" deleted by Erik Hennum/Oakland/IBM] [attachment "ecblank.gif" deleted by Erik Hennum/Oakland/IBM]
- References:
- RE: [dita] Last call for top DITA issues
- From: "Jennifer Linton" <jennifer.linton@comtech-serv.com>
- RE: [dita] Last call for top DITA issues
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]