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:

Proportional Font

LISTSERV Archives

LISTSERV Archives

TEI-L Home

TEI-L Home

TEI-L  March 2002

TEI-L March 2002

Subject:

Re: teixlite.dtd

From:

Lou Burnard <[log in to unmask]>

Reply-To:

[log in to unmask]

Date:

Thu, 21 Mar 2002 18:36:38 +0000

Content-Type:

TEXT/PLAIN

Parts/Attachments:

Parts/Attachments

TEXT/PLAIN (74 lines)

For what it's worth, my understanding of the underlying logic behind this
seemingly nonensical state of affairs is as follows.

1. <s> was originally provided for one purpose only: the end-to-end
segmentation of linguistic corpora. This means it should be used with the
following constraints:
  -- every word in the textt must be contained by one <s> only
  -- <s> elements may not be nested inside other <s> elements

2. <seg> was provided for those, not a few, who considered the above
   restrictions unreasonable. In particular,
  -- <seg>s may appear at arbitrary places within a text
  -- <seg>s may be nested within other <seg>s

Where things get complicated, is when you start trying to fit <s> and
<seg> into the TEI class structure. Clearly both <s> and <seg> are
phrase-level, in the sense that they can occur multiple times within a
single chunk-level item (such as a <p> or an <ab>): this defines where
they are permitted in the DTD.

There is however a clear inconsistency in their contents: <s> can contain
only other phrase-level elements, whereas <seg> can contain either
phrase-level or inter-level elements such as <q> and <bibl>. Why that
should be, I am not sure, but presumably it is because of the greater
generality of application envisaged for <seg>: you might well want to
tag as a single segment something containing a <q>. You may retort that
the same is true of <s>, and maybe you're right, but that's not the way
<s> was originally envisaged (see above) -- the assumption was that if
you found something new inside an <s> (such as a <q> or a <bibl>) that,
ipso facto, was enough to warrant a new <s>.

As with all other such questions, this is one that can be reconsidered by
the Tech Council, now that we have one.

Lou

p.s. yes I know that the P4X model for <s> now includes <s> even though it
says in the text it's not allowed. This is a consequence of xmlification,
which doesn't permit exclusion exceptions. In P5, this restriction,
if still required, will be achieved by a schema rule


On Tue, 19 Mar 2002, Guy Aston wrote:

> <p>
> <s>"As it is", Guy said, "it seems confusing". </s>
> <s>And confusion, as the bible tells us (TEI guidelines, Chicago & Oxford)
> is to be avoided.</s>
> </p>
>
> The dtd states:
>
> 1.
> <s> cannot contain <q> or <bibl>
> <s> can contain <seg>
>
> 2.
> <seg> can contain <q> and <bibl> (and, interestingly,  <s>)
>
> In other words,  I can tag a stretch inside an <s> as <q> or <bibl>
> providing I put it inside a <seg>.
>
> <p>
> <s><seg type="sbit_q"><q>"As it is"</q></seg>, Guy said, <seg
> type="sbit_q"><q>"it seems confusing"</q></seg>. </s>
> <s>And confusion, as the bible tells us <seg type="sbit_bibl">(<bibl>TEI
> guidelines, Chicago & Oxford</bibl>)</seg> is to be avoided.</s>
> </p>
>
> So why shouldn't I? Or is this an error in the dtd?
>
> Guy
>

Top of Message | Previous Page | Permalink

Advanced Options


Options

Log In

Log In

Get Password

Get Password


Subscribe or Unsubscribe

Subscribe or Unsubscribe


Archives

ATOM RSS1 RSS2



LISTSERV.BROWN.EDU

CataList Email List Search Powered by the LISTSERV Email List Manager