RE: TT Req (translation markup)

Glenn,
 
OK, I see what you mean, I guess the questions then are
 
1) Does TT AF need to include a descriptive vocabulary that supports some or
all of the XLIFF mtype values?
2) Is the element name phrase appropriate - since phrase has a rather narrow
meaning.
3) Will TT AF allow extension to support user defined inline markup? Does TT
AF need a "get-out" clause (rather like the ruby parenthesis one?) 
 
regards John B

-----Original Message-----
From: Glenn A. Adams [mailto:glenn@xfsi.com]
Sent: 30 June 2003 14:26
To: Johnb@screen.subtitling.com; Luke-Jr@cox.net
Cc: public-tt@w3.org
Subject: RE: TT Req (translation markup)


 
The purpose of the phrase vocabulary item indicated in TT AF 1.0 Requirement
R209 is to satisfy the types of usage you have listed below. So I think we
are already covered here.
 
Regards,
Glenn
 

  _____  

From: Johnb@screen.subtitling.com [mailto:Johnb@screen.subtitling.com] 
Sent: Monday, June 30, 2003 6:10 AM
To: Luke-Jr@cox.net; Glenn A. Adams
Cc: public-tt@w3.org

Luke, Glenn, et al 

Actually this may be an important issue. 
It may be desirable to allow inline markup within a TT document that marks
sections of the text for specific purposes. I have in mind the XLIFF <mrk>
element.

The specification for XLIFF can be found at: 
http://www.oasis-open.org/committees/xliff/documents/xliff-specification.htm
<http://www.oasis-open.org/committees/xliff/documents/xliff-specification.ht
m>  

The mrk element has an mtype attribute 

<mrk mtype/> 

where the recommended values for the mtype attribute are as follow 

(this list is not exhaustive – and reference should be made to the XLIFF
specification): 
- abbrev = abbreviation, acronym, etc. 
- datetime = date or time information. 
- name = proper or common name. 
- phrase = sub-sentence level. 
- protected = text that should remain untouched during the process. 
- term = one or more words of a terminology entry. 

Some examples of a text element using mrk markup are shown below: 
<text id="T2" class="Charlie">Afternoon, <mrk mtype="abbrev">Dr</mrk>. <mrk
mtype="name">Sparrow</mrk>.</text> 
<text id="T27" class="Sir Lancelot">I don't want <mrk mtype="protected"
xml:lang="en-us">garbage</mrk> on my car.</text>

The first example illustrates the use of the mrk element to identify an
abbreviation and a proper name that would probably require special
processing by a machine translation system. In the second example the mrk
element protects the word 'garbage' from machine translation – thus this
word should remain as 'American English' regardless of any translation of
the remainder of the surrounding text.

Is there any reason why TT format would prohibit or prevent the use of such
markup?