LISTSERV mailing list manager LISTSERV 16.5

Help for TEI-L Archives


TEI-L Archives

TEI-L Archives


TEI-L@LISTSERV.BROWN.EDU


View:

Message:

[

First

|

Previous

|

Next

|

Last

]

By Topic:

[

First

|

Previous

|

Next

|

Last

]

By Author:

[

First

|

Previous

|

Next

|

Last

]

Font:

Monospaced Font

LISTSERV Archives

LISTSERV Archives

TEI-L Home

TEI-L Home

TEI-L  May 2005

TEI-L May 2005

Subject:

Re: Encoding to enable sorting of titles that start with articles

From:

Michael Beddow <[log in to unmask]>

Reply-To:

Michael Beddow <[log in to unmask]>

Date:

Mon, 30 May 2005 14:56:38 +0100

Content-Type:

text/plain

Parts/Attachments:

Parts/Attachments

text/plain (251 lines)

[LB]
[...]
> even though
> you haven't answered the questions I asked

That was maybe my tactful way of dissenting from the claim in the original
posting that

> The only issue on which I think debate is still necessary is the
> names for the two elements concerned

I think the general concepts you propose are fine and meet the requirements
neatly, so I'm not too worried about nomenclature, which I don't think
matters that much provided it follows consistent principles across different
tagset areas, isn't intrinsically confusing, and is backed up by clear
examples and suitable discursive equivalents of the formal definitions. But
in the current draft, I don't think the third of those provisos is yet met.

My problem with that draft (temporarily exacerbated of course by the
incorrect definition for the <index> element in the original version, now
corrected) is that I find two of the main examples problematic, and I am not
altogether happy about the general sequencing and emphasis of the
exposition.

My prime source of unease goes back to the P4 presentation of this topic. I
don't think that did justice to the highly skilled nature of index creation
(understandably perhaps, because academic presses increasingly feel able to
dispense with such skills and rely on copy-editors or, worse still, authors
themselves, to create indexes to scholarly texts). But that shortcoming is I
think considerably worsened in this new draft which makes rather a big thing
of automated multi-level index generation in a way which strikes me as
liable to raise misguided expectations and false hopes. At the same time,
the new draft relegates mention of automated creation of multiple separate
indexes (a very different thing from multi-level indexing, of course, and
one much easier to carry out satisfactorily by machine) to the presence of
one attribute in one example towards the end, whereas it was included early
on in the P4 discursive presentation as a matter in its own right. I would
say that the generation of multiple indexes, along with the attribute that
controls this matter, need earlier and more prominent mention here too.
Such multiple indexes are, after all, commonly used, and are easy to
generate automatically from markup which is straightforward to apply. That
makes them very different from multi-level indexes, which, unless the text
concerned has a very strict and consistent structure, can be generated
automatically only with great difficulty, and apposite application of the
necessary markup may lie outside the skills to be reasonably expected of
encoders, a point that goes unrecognised in the current draft.

I'll attempt in a moment to sketch an alternative approach to the structure
of this section (which will I think make clearer what I mean about the
skilled nature of index creation, so I won't expound that point any futher
here) but before that let me say in more detail why I think two of the
examples are unfortunate.

The first example in the corresponding section of P4 runs

<p>The students understand procedures for Arabic lemmatisation
<index level1="Arabic lemmatization"/>and are beginning
to build parsers.</p>

That is plain and wholly apposite. It assigns a value to the (in P4
mandatory) level1 attribute, and doesn't attempt to show assignment to more
than one index term entry. But in the new draft, this first example has been
repurposed to illustrate not only index markup in its most elementary form,
but to furnish in an example of multi-term (albeit single-level) assignment
as well, even though it is now introduced with a phrase "in the simplest
case..." which the elaboration has made rather inaccurate, since the
simplest case would assign to one term only, as did the P4 example.

 <p>The students understand procedures for Arabic lemmatisation
  <index>
   <indexTerm>Arabic lemmatization</indexTerm>
   <indexTerm>Lemmatization</indexTerm>
  </index>and are beginning to build parsers.</p>

But there is a more serious problem as well. Thus revised, the example
shows a dubious indexation practice, of a kind I would not accept if applied
by an publisher to anything authored by me. It confuses multi-term
assignment and multi-level assignment. It would be unobjectionable to assign
this point in the text to two top-level index terms: "Arabic" and
"Lemmatization". Alternatively (although not yet capable of being
illustrated by the markup being examined at this introductory point in the
section) this text point could be assigned to two second-level entries in a
hierarchical index, viz" Lemmatization, Arabic", and "Arabic,
lemmatization". An index such as this example would create, having two
top-level entries, one labelled "Arabic lemmatization", and the other simply
"Lemmatization" would be defective qua index. Now the proposed revision does
indeed allow multi-level assignments to be encoded rather elegantly; but
that is not the topic that this example claims to address, nor could it do
so without introducing more concepts and elements that are appropriate for
the introductory specimen.

The other example I am unhappy about shows a different kind of problem.
Unlike the first one, it does indeed exemplify the point it claims to
address, but all the same I think it misrepresents the feasibility of
actually applying the type of markup illustrated in many likely encoding
situations.

As it stands, one can imagine project directors instructing their encoders
to mark up a text so as to generate a multi-level index of the type
exemplifed from Burton's A of M, on the grounds that the Guidelines tell you
how to do it. In fact they don't; and can't, so shouldn't claim to, at least
not without considerable qualification.

It is easy enough with a specimen end-product to hand such as the example
cites from an index to the Anatomy, to "reverse engineer" the sort of markup
that might generate such index entries. What is much more difficult (indeed,
to my mind virtually impossible) to envisage is how an encoder (even an
author-encoder) could reliably lay the markup eggs that would produce this
sort of output chicken, unless s/he had access to a pre-existent index to
the text already drawn up along these lines.

The example invites a critique along the lines Wittgenstein offers of
Augustine's account of infantile language learning. In order to learn
language by the means Augustine describes in his Confessions, Wittgenstein
points out, the infant would already have to know a language in which to
interpret the ostentative definitions of objects which according to
Augustine are the sole vehicle of language acquisition. Similarly, in order
to use the <index> elements to insert markup that would allow an application
to generate a multi-level index of the type exemplified here, the encoder
would need to have to hand a manually-prepared multi-level index prepared by
someone who had deployed a human indexer's skill to comb the text and
marshal the resultant gleanings. I find it barely conceivable that an
encoder, working linearly through the copy text and alighting on the phrases
indexed in the examples, could deploy the prodigious powers of conspectus
required to say "ah yes, this phrase belongs in level three zzz, below level
2 yyy, which is below level 1 xxx" and hence confidently insert the correct
recursive markup. But if the encoder would instead need a manually
pre-prepared index in order to assign the correct hierarchical index markup,
then why bother with tagging for supererogatory automated index generation
at all, rather than just marking up the pre-prepared index as back matter?

Not that I am denying that support for automatic generation of a
hierarchical index via markup is conceptually impossible. My contention is
that the correct application of such markup is likely to be a practical
impossibility in much real-world encoding, except where the text to be
encoded has a highly disciplined and explicit hierarchical structure (e.g.
is written on the Germanic PhD model with profuse analytical subheadings so
that the locus of any given phrase within the overall conceptual scheme
would be relatively easy for an encoder to identify reliably and
consistently). So yes, support for generation of multi-level indexes does
belong in the tagset, and examples of its deployment belong in the
Guidelines, but not with the material currently chosen as an exemplar. And
since such a facility is in practice not likely to be used very often, it
doesn't in my view merit the prominence it has in the present draft among
the "things you need to know" about indexation-related markup.

With those two particular criticisms, I hope, explained, I now offer an
outline rewrite of the bulk of what the draft covers. As well as adding
material (and altering the sequence of the exposition) I have also made
various detailed changes in phrasing when taking over text from the draft:
since I hope the reasons for those changes will be either self-evident or
deducible from what I have written above, I have not explicitly commented on
them.

=============

6.8.2 Index Entries

Indexing scholarly texts is a skilled activity best performed by human
beings with the requisite knowledge and experience. Where existing texts are
being digitized, it should not be automatically assumed that standard
searching and information retrieval software will be able to meet all the
needs fulfilled by a well-crafted manual index, and where such an index is
present it may be advisable to retain and encode it. The <div1> element or a
<div> element at an appropriate level may be used to demarcate the index,
and the index itself transcribed as a structured list or table.

It can also be useful, however, to generate a new index from a
machine-readable text, whether the text is being written for the first time
with the tags here defined, or as an addition to a text transcribed from
some other source. Depending on the complexity of the text and its subject
matter, such an automatically-generated index may not in itself satisfy all
the needs of scholarly users. However it can assist a professional indexer
to construct a fully adequate index, which might then be post-edited into
the digital text, marked-up along the lines already suggested for preserving
pre-existing index material.

Indexes generally contain two main types of reference: to specific pages and
to page ranges. The first type of reference is generated by placing one or
more <index> elements at appropriate points in the text. The second type
requires the use of an <indexSpan> element to mark the start of the range to
be targeted, with an <anchor> element marking its end, the two boundaries
being associated by an Xpointer.

[examples]

It is also common practice to compile more than one index for a given text.
A biography of a poet, for example, may offer an index of references to
poems by the subject of the study, another index of works by other writers,
an index of places or historical personages etc. These Guidelines offer a
means of assigning index terms and locations to one or more specific indexes
by means of the indexName attribute.

[examples]

Additionally, indexes may be multi-level. That is to say, as well a entries
such as "TEI", or "markup", they may contain articulated items like "TEI,
markup practices, index terms", with the top level entry "TEI"
being followed by a number of second-level subcategories, any or all of
which may have a third-level list attached to them and so on. In order to
reflect such a hierarchical index listing, the <index> elements may be
nested to the required depth.

[examples]

One final feature frequently found in manually-created indexes to printed
works cannot readily be encoded by the means provided here, namely cross
references internal to the index term listing. For example, the entry "TEI"
might not contain a reference to the he text proper, but instead read simply
"see Text Encoding Initiative", under which entry the actual page references
would then be grouped. Where such internal cross-references are regarded as
essential, they would need to be added by post-editing and encoding an
auto-generated intermediate index.

Although the position of the <index> or <indexSpan> element in the text
furnishes the target location that will be specified in the generated index
entry, no part of the text itself is used to construct that entry. Index
terms appearing in the entry come solely from the content of <indexTerm>
elements, which consequently may have to repeat words or phrases from the
text proper. They need not do so verbatim,which gives scope for
normalization of spelling and other modifications which may assist
generation of an index in a desired form or sequence. Sometimes even a
normalized form of an index term may be insufficent for an application to
order it exactly as desired, in which case a sortKey attribute may be placed
on the <indexTerm> element and given an appropriate value. (In such
instances, a sorting application will need to process <indexTerm>s which
have such an attribute differently from those that lack it: it would be
helpful to the developers of such applications if this point could be made
explicit at a suitable place in the teiHeader.)

[examples]


[...]

=====================================


Re the sortKey attribute, which I would indeed say is a good idea, it does
rather raise the question of what should be the future of the various other
attributes named "key" throughout the tagset. P4 tends to gloss these as
representing either a sort key or the key to some form of database, but
those are really rather different things. Now this present instance really
is proposed for use as a sorting aid, not as a "key" in the database sense,
but I rather wonder if there isn't a case for either similarly
disambiguating, where appropriate, all the other "key"s elsewhere, or for
leaving the "mostly harmless", but consistent, ambiguity here and just
calling it "key" like all the rest of the clan?

Michael Beddow

Top of Message | Previous Page | Permalink

Advanced Options


Options

Log In

Log In

Get Password

Get Password


Search Archives

Search Archives


Subscribe or Unsubscribe

Subscribe or Unsubscribe


Archives

July 2019
June 2019
May 2019
April 2019
March 2019
February 2019
January 2019
December 2018
November 2018
October 2018
September 2018
August 2018
July 2018
June 2018
May 2018
April 2018
March 2018
February 2018
January 2018
December 2017
November 2017
October 2017
September 2017
August 2017
July 2017
June 2017
May 2017
April 2017
March 2017
February 2017
January 2017
December 2016
November 2016
October 2016
September 2016
August 2016
July 2016
June 2016
May 2016
April 2016
March 2016
February 2016
January 2016
December 2015
November 2015
October 2015
September 2015
August 2015
July 2015
June 2015
May 2015
April 2015
March 2015
February 2015
January 2015
December 2014
November 2014
October 2014
September 2014
August 2014
July 2014
June 2014
May 2014
April 2014
March 2014
February 2014
January 2014
December 2013
November 2013
October 2013
September 2013
August 2013
July 2013
June 2013
May 2013
April 2013
March 2013
February 2013
January 2013
December 2012
November 2012
October 2012
September 2012
August 2012
July 2012
June 2012
May 2012
April 2012
March 2012
February 2012
January 2012
December 2011
November 2011
October 2011
September 2011
August 2011
July 2011
June 2011
May 2011
April 2011
March 2011
February 2011
January 2011
December 2010
November 2010
October 2010
September 2010
August 2010
July 2010
June 2010
May 2010
April 2010
March 2010
February 2010
January 2010
December 2009
November 2009
October 2009
September 2009
August 2009
July 2009
June 2009
May 2009
April 2009
March 2009
February 2009
January 2009
December 2008
November 2008
October 2008
September 2008
August 2008
July 2008
June 2008
May 2008
April 2008
March 2008
February 2008
January 2008
December 2007
November 2007
October 2007
September 2007
August 2007
July 2007
June 2007
May 2007
April 2007
March 2007
February 2007
January 2007
December 2006
November 2006
October 2006
September 2006
August 2006
July 2006
June 2006
May 2006
April 2006
March 2006
February 2006
January 2006
December 2005
November 2005
October 2005
September 2005
August 2005
July 2005
June 2005
May 2005
April 2005
March 2005
February 2005
January 2005
December 2004
November 2004
October 2004
September 2004
August 2004
July 2004
June 2004
May 2004
April 2004
March 2004
February 2004
January 2004
December 2003
November 2003
October 2003
September 2003
August 2003
July 2003
June 2003
May 2003
April 2003
March 2003
February 2003
January 2003
December 2002
November 2002
October 2002
September 2002
August 2002
July 2002
June 2002
May 2002
April 2002
March 2002
February 2002
January 2002
December 2001
November 2001
October 2001
September 2001
August 2001
July 2001
June 2001
May 2001
April 2001
March 2001
February 2001
January 2001
December 2000
November 2000
October 2000
September 2000
August 2000
July 2000
June 2000
May 2000
April 2000
March 2000
February 2000
January 2000
December 1999
November 1999
October 1999
September 1999
August 1999
July 1999
June 1999
May 1999
April 1999
March 1999
February 1999
January 1999
December 1998
November 1998
October 1998
September 1998
August 1998
July 1998
June 1998
May 1998
April 1998
March 1998
February 1998
January 1998
December 1997
November 1997
October 1997
September 1997
August 1997
July 1997
June 1997
May 1997
April 1997
March 1997
February 1997
January 1997
December 1996
November 1996
October 1996
September 1996
August 1996
July 1996
June 1996
May 1996
April 1996
March 1996
February 1996
January 1996
December 1995
November 1995
October 1995
September 1995
August 1995
July 1995
June 1995
May 1995
April 1995
March 1995
February 1995
January 1995
December 1994
November 1994
October 1994
September 1994
August 1994
July 1994
June 1994
May 1994
April 1994
March 1994
February 1994
January 1994
December 1993
November 1993
October 1993
September 1993
August 1993
July 1993
June 1993
May 1993
April 1993
March 1993
February 1993
January 1993
December 1992
November 1992
October 1992
September 1992
August 1992
July 1992
June 1992
May 1992
April 1992
March 1992
February 1992
January 1992
December 1991
November 1991
October 1991
September 1991
August 1991
July 1991
June 1991
May 1991
April 1991
March 1991
February 1991
January 1991
December 1990
November 1990
October 1990
September 1990
August 1990
July 1990
June 1990
April 1990
March 1990
February 1990
January 1990

ATOM RSS1 RSS2



LISTSERV.BROWN.EDU

CataList Email List Search Powered by the LISTSERV Email List Manager