dita message
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]
Subject: RE: [dita] indexing question
- From: "Grosso, Paul" <pgrosso@ptc.com>
- To: <dita@lists.oasis-open.org>
- Date: Fri, 21 Jul 2006 10:38:07 -0400
Actually, the point specified by an indexterm in the prolog
never gets associated with any page since the prolog never gets put into the
output (under the usual styling of "prolog"), so per the standard indexterm
processing, an indexterm within a prolog would simply be ignored (i.e., never
generate an index entry).
So I misstated when I said such might create an index entry
pointing to the second page of the topic.
But I was correct that an indexterm in a prolog is useless
(under the standard indexterm processing), so we should certainly suggest people
don't do this.
Bruce is suggesting something akin to what Erik (I believe)
implied a while ago: that an indexterm within a prolog have "special semantics"
and "act as if" there is an indexterm on the first page of the
topic.
We could do this. And perhaps this would address the
user requirement to be able to have an index entry that points to the first page
of a topic and therefore imply we don't need to allow indexterm within title and
shortdesc and such. I'm just not sure if having to insert a prolog containing a
keyword containing an indexterm--as opposed to just putting an indexterm in the
title--will be acceptable to users.
Users on (and represented by members of) the TC--what
do you say?
paul
Paul,
The
"point" specified by an indexterm in the prolog isn't necessarily after the
title or shortdesc.
We
can interpret that when the source markup contains an indexterm in
the prolog, the indexterm is understood to be tied to a point at the beginning
of the title.
That's just to preserve the primary method, in which
an indexterm that applies to the entire topic is entered in the
prolog.
Best
wishes,
Bruce
We could resolve the issue this way.
The problem with this solution is the case where you have a long
shortdesc (which precedes the prolog) so that the point-wise, subflow
indexterm in the prolog ends up on the second page of the topic so your
index entry does not have the page number of the first page in the topic.
So, in fact, an indexterm in a prolog really has no useful purpose, since
you cannot guarantee that its effect is going to be the desired
one.
So I could live with this solution provided that we
allow indexterm in other places that it is currently not allowed (e.g.,
title) so that a user can ensure they get an index entry pointing to the
first page of a topic.
And once we do that, we would then issue a "best
practices" statement saying not to put indexterm within
prolog.
paul
From what I understand, the issue of
treating indexterm differently in topic prolog vs content was due to the
misunderstanding that the content of indexterm in content actually
appears as part of that content. Since we have now clarified
that indexterm's content is always a subflow, we can treat indexterm
uniformly in both topic prolog and content. Is my understanding correct
that this issue is now resolved?
Chris
Hi, Rudolfo, Dave, and other index enthusiasts:
A lightbulb went
off for me. I think we're conflating two cases here.
In the case of
<keyword>, the element is an inline within content and a subflow in
the prolog as Rudolfo has stated.
In both cases, <keyword>
identifies a word from a vocabulary. In the inline case, the vocabulary
word is delimited within the flow. In the prolog case, the vocabulary word
is identified as potential metadata for search engines.
In neither
case do we have a base processing expectation of producing published
indexes for the <keyword> element. We have sometimes speculated
about the possibility of generating indexes from inline mentions (in
Eliot's term) of vocabulary words, but I believe we've always deferred
that.
By constract, <indexterm> is a subflow in all cases as
Chris has stated. The contents of <indexterm> must be translated,
but the translation of the content in which <indexterm> is embedded
isn't affected in any way by the positioning of <indexterm>. That
is, <indexterm> does not delimit part of the flow.
In the
prolog, <indexterm> is specified as serving two purposes: feeding
index terms to search engines as part of the metadata and indexing the
topic.
Digression: The case could be made to treat <term> in
exactly the same way as <keyword>, using <keyword> for words
from formal languages and <term> for words from cultural or social
vocabularies. That would require adding <term> to the prolog.
Hoping that clarifies,
Erik
Hennum
ehennum@us.ibm.com
"Rodolfo M. Raya"
<rodolfo@heartsome.net>
"Rodolfo M. Raya"
<rodolfo@heartsome.net>
07/18/2006 02:15
PM |
|
On Tue, 2006-07-18 at 15:03 -0500, David Walters
wrote:
Hi,
New example:
The <p> text is
complete.
<topic>
<prolog>
<indexterm>term one</indexterm>
<indexterm>term
two</indexterm>
</prolog>
<body>
<p>Paragraph that contains term one <indexterm>term
one</indexterm>
and term two <indexterm>term two</indexterm>
inside.</p>
</body>
</topic>
If the content of
<indexterm> is
completely ignored when the topic is published as XHTML, PDF or any other
format, then this element should be completely ignored at translation
time.
The content of <indexterm> doesn't need to be
translated if it is only a location marker. The whole element can be
replaced with a tag by the the translation tool.
Best
regards,
Rodolfo
-- The
information in this e-mail is intended strictly for the
addressee, without prejudices, as a confidential
document. Should it reach you, not being the addressee,
it is not to be made accessible to any other
unauthorised person or copied, distributed or disclosed
to any other third party as this would constitute an
unlawful act under certain circumstances, unless prior
approval is given for its transmission. The content of
this e-mail is solely that of the sender and not
necessarily that of Heartsome.
| | |
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]