Romain Roure | 2 Jun 17:42 2008
Picon

tei_all.xsd validation problem

Hello everyone,

I am currently using the tei_all.xsd schema provided there : 
http://www.tei-c.org/release/xml/tei/custom/schema/xsd/tei_all.xsd.
It seems the schema itself is not valid (I've been browsing through the 
mailing list archive for corresponding posts but didn't find anything 
related).
Basically, tei_all.xsd imports teix.xsd. teix.xsd imports tei_odds.xsd 
which does not exist physically at the given url.
I've been trying to workaround this by generating my own tei_odds.xsd 
using Oxygen and tei_odds.rng.
It still fails as tei_all.xsd and tei_odds.xsd are sharing the same 
namespace (http://www.tei-c.org/ns/1.0) and thus tei_odds.xsd never get 
loaded by XML validators/parsers.
I kinda am stuck here and would appreciate any help you can provide.

Thanks in advance,

Romain

Courtney Michael | 2 Jun 22:57 2008

att.timed and timecode

Hello.

We are just beginning to explore using TEI  and the guidelines for  
Transcriptions of Speech to encode transcripts of interviews  
originally taped for broadcast. for one of our projects. We are  
planning to digitize the video interviews as well. We would like to  
include timecode in the TEI document and link it to the video files at  
regular intervals.

Would the <timeline> element be the proper place for timecode  
information? Does anyone have experience with this that they'd be  
wiling to share?

Many thanks,
Sincerely,

Courtney Michael
Project Manager
WGBH Media Library & Archives
One Guest Street
Boston, MA 02138
p. 617-300-2673
f. 617-300-1056
courtney_michael <at> wgbh.org

McAulay, Elizabeth | 2 Jun 23:48 2008

Re: att.timed and timecode

Hi Courtney,

We're entering timecodes into our oral history transcripts (to match
corresponding audio or video). After posting to the TEI in Libraries
listserv, we learned that there isn't any built-in place to put the
timecodes, so we need to extend the TEI to add start/stop to an element
as an attribute. We're adding an element within the <sp> (for "speech")
tag for that purpose now. We also considered using the <milestone> tag.
So our current encoding looks like this:

FRAGMENT:
     <sp id="3">
                <start>00:00:33</start>
                <speaker> ARKIN </speaker>
                <p>Well, I was born in Detroit, Michigan, at Women's
Hospital, delivered by my uncle
                    Al, who was my mother's uncle, so I guess -- no,
mother's cousin. And -- on
                    February 18, 2007. So I was, for the first few years
of my life, I think that we
                    lived in the very central city, and when I was about
three, we moved to a
                    suburban area where there were no roads. I mean,
there were roads, but most of
                    the roads were dirt roads, and there weren't all
that many houses, although I
                    imagine it was early [tracked] housing in that
neighborhood. And this was in
                    Detroit, Michigan, I think I mentioned. Auto City
and future city of soul music,
(Continue reading)

McAulay, Elizabeth | 3 Jun 00:23 2008

FW: att.timed and timecode

Just wanted to share this input from Holly Wang who is leading this
project:

-----Original Message-----
From: Wang, Holly 
Sent: Monday, June 02, 2008 2:56 PM
To: McAulay, Elizabeth; 'Courtney Michael'
Subject: RE: att.timed and timecode

Courtney,

I have looked into all the time related tags in TEI and couldn't find
the proper one to use. The most close one is <u> which has a start and
stop attribute and could also be used for the speaker change. We are
currently using <sp> for the speaker change during speech and <u> is not
the child element of <sp>. For right now, we have modified the TEI
schema so <sp> could have the start and end attribute. Here is an
example:

<sp id="1" start="00:00:03" end="00:00:05">
	<speaker> ARKIN </speaker>
	<p>how are you? </p>
</sp>

Let me know if you need more example

Holly Wang 
Digital Library Program of UCLA hollyw <at> library.ucla.edu 310-206-4044

-----Original Message-----
(Continue reading)

Jarom McDonald | 3 Jun 02:53 2008

Re: att.timed and timecode

Obviously, the information of the start and stop points are relative to one particular audio/video file,
and so your TEI document would have to somehow make reference to that file also. While simple extension of
certain TEI elements (as mentioned by a couple of replies) is definitely one solution (which you'd want to
augment with setting up a target/ptr relationship so end users would have access to the reference point
for those timecodes), you might want to give thought to a more complex solution such as incorporating a
more thorough time-based annotation model from a different XML project. There are two that I use in
various projects. One is the Continuous Media Markup Language (CMML), first developed by the same group
that created the Ogg multimedia container and the Theora/Vorbis codec
 s. The other is the MPEG-7 standard (as its name suggests, put out by the same Moving Pictures Experts Group
that gave us mpg, mp3, mp4, etc.), formally known as the Multimedia Content Desc!
 ription Interface (actually, this isn't so much as a schema as it is a set of standards, descriptors and
extensions to the concept of a schema).

Both of these allow for not only timecode referencing, but also for annotation of those time codes, along
with general information about audio/video codecs, bitrates, framerates, sampling rates, video
width/height, and so forth. Approaching your tagging of the interviews by utilizing schemas from
multiple namespaces (i.e. textual information with the tei namespace, video information with the cmml
namespace which is much simpler to deal with for low-complexity projects) could potentially make your
digital objects much more useful.

Then again, it might be overkill as well.

Jarom McDonald
Assistant Research Professor
Humanities Technology and Research Support Center
1163G JFSB
Brigham Young University
Provo, UT 84602
(801) 472-0993
jarom_mcdonald <at> byu.edu
(Continue reading)

David Sewell | 3 Jun 03:27 2008

Re: tei_all.xsd validation problem

There are a couple of things going on here that I can't fully diagnose
(Sebastian Rahtz will probably need to weigh in). As for the files at 
http://www.tei-c.org/release/xml/tei/custom/schema/xsd, they are the 
ones generated by running "make" on the P5 sources:

  http://tei.svn.sourceforge.net/viewvc/tei/trunk/P5/Exemplars/

but the Makefile there seems to treat tei_odds.xsd as a temporary file 
as it is deleted after other schemas are generated. Hence it is not 
included among the release files.

Running command-line roma on tei_odds.xsd produces valid, working .xsd 
files. I'll email them to you off-list.

The Web-based Roma also has some problems as it produces the single file 
tei_all.xsd but does not also generate the included teix.xsd.

DS

On Mon, 2 Jun 2008, Romain Roure wrote:

> Hello everyone,
>
> I am currently using the tei_all.xsd schema provided there : 
> http://www.tei-c.org/release/xml/tei/custom/schema/xsd/tei_all.xsd.
> It seems the schema itself is not valid (I've been browsing through the 
> mailing list archive for corresponding posts but didn't find anything 
> related).
> Basically, tei_all.xsd imports teix.xsd. teix.xsd imports tei_odds.xsd which 
> does not exist physically at the given url.
(Continue reading)

Syd Bauman | 3 Jun 05:25 2008

Re: att.timed and timecode

Hi Lisa! Hope all is well in sunny L.A.

EM> I'm interested in hearing any comments from the group.

OK. I'm going to discuss Lisa's post first, then tackle the original
poster's question.

I'm sorry I didn't catch this discussion earlier on the Library SIG
list. Must have missed it, sorry. Anyway ...

I have inserted line numbers in the following for reference.

EM>  1. <sp id="3">
EM>  2.   <start>00:00:33</start>
EM>  3.   <speaker> ARKIN </speaker>
EM>  4.   <p>Well, I was born in Detroit, Michigan, at Women's
EM>  5.   ...
EM>  6.   would reminisce about growing up.</p>
EM>  7.   </speaker>
EM>  8. </sp>

This snippet brings up lots of thoughts.

First, the use of id= as opposed to xml:id= on line 1 makes me think
you're using P4 instead of P5. Especially if you are customizing, I
recommend switching to P5 as soon as reasonable. It really is
(overall) a lot better.

Second, whether you're using P4 id= or P5 xml:id=, "3" is not a valid
value. IDs must start with a letter.
(Continue reading)

Sebastian Rahtz | 3 Jun 10:49 2008
Picon
Picon

Re: tei_all.xsd validation problem

I am gearing up to a new release of the TEI in a couple
of weeks, so this is just in time to maybe get fixed.

However, it's more than a missing file; the
schema barfs even then. I am looking at it.

I strongly recommend using the RELAXNG
schemas, if you can.

--

-- 
Sebastian Rahtz      
Information Manager, Oxford University Computing Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431

Sebastian Rahtz | 3 Jun 10:56 2008
Picon
Picon

new release of TEI P5 imminent

We hope to release TEI P5 1.1 in a  few weeks, containing
a variety of small fixes and improvements

There is no chance now to fix extra bugs, but if anyone
has small typos they have not yet reported,
there is still time to get those cleaned up.

--

-- 
Sebastian Rahtz      
Information Manager, Oxford University Computing Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431

Torsten Schaßan | 4 Jun 19:06 2008
Picon

question concerning author identification

Hi all,

I searched the Guidelines and archives but couldn't find an explanation
or example for the hopefully simple question:

How would I best link from <author> to any resource containing 
standardized information about authors like the Personennamendatei (PND) 
or the LoC Name Authority File?

Would it be reasonable to add to the definition of <author> the
att.naming in order to have  <at> key available for this purpose or do you 
prefer another way like another element inside <author>?

Best, Torsten

---
Torsten Schassan
Herzog August Bibliothek, Postfach 1364, D-38299 Wolfenbuettel
Tel.: +49-5331-808-130, schassan {at} hab.de
    http://www.hab.de; http://www.hab.de/forschung/projekte/weiss64.htm


Gmane