RE: [dita-translation] Translating required-cleanup

["Farwell, Kevin" <Kevin.Farwell@lionbridge.com>]

Hi,

I wonder if this is more a development/localization process question.
According to the reference, "As the element name implies, the intent for
authors is to clean up the contained material and eventually get rid of
the <required-cleanup> element." and "Because the content of
<required-cleanup> is not considered to be verified data..." I take
those to mean the content needs editing, even if that's only tag or
attribute editing. I wouldn't think such content even be a candidate for
translation. I would expect it to be looping through some editorial
process. In the example below, the paragraph that has unverified data
should not be translated either. Assuming the content is correct, the
tagging errors would be propagated over the whole language set. Assuming
it isn't, both tagging errors and content errors are multiplied.

If it should make it out the door, I would take option one. That would
force people to choose to get risky content translated.

Kevin
  

-----Original Message-----
From: Robert D Anderson [mailto:robander@us.ibm.com] 
Sent: Wednesday, April 19, 2006 7:31 AM
To: dita-translation@lists.oasis-open.org; mambrose@sdl.com;
pcarey@lexmark.com; rfletcher@sdl.com; bhertz@sdl.com; 'Richard Ishida';
tony.jewtushenko@productinnovator.com; 'Lieske, Christian'; 'Jennifer
Linton'; Munshi, Sukumar; Charles Pau; dpooley@sdl.com; Reynolds, Peter;
'Felix Sasaki'; 'Yves Savourel'; Dave A Schell; 'Bryan Schnabel'; Don
Day
Subject: [dita-translation] Translating required-cleanup


This question has come up a couple of times outside of the list, so I'd
like this group to address it:

The definition of the translate attribute indicates that it applies to
the current element and all nested elements, unless or until the value
changes in a nested element. We also have conventions that
<draft-comment> and <required-cleanup> should not be translated by
default, because they are not included in the output.

My understanding is that if a user places translate="yes" on a
required-cleanup element, this means that it should be translated. In
this case, the user knows best - the contents will be used for some
purpose, and should be translated.

What about this case?
<p translate="yes"> ...translatable text ...
  <required-cleanup> text in here </required-cleanup> </p>

From the definition of @translate, it seems that requried-cleanup will
inherit translate="yes", making it translatable. It also seems that, in
most cases, this is not the desired behavior. I'm wary of making this an
exception to the rule, though, because exceptions just make DITA more
difficult to implement.

I think what's actually wanted is that required-cleanup should have a
default setting of translate="no", set within the DTDs and schemas. This
means that the only way to make it translatable is to explicitly set the
attribute on the element. Specializations that are used for translatable
content, such as <reusableContent>, could change the default to "yes"
for that element.

What do others think? I think to clear up the confusion, we have to do
one of the following (I would vote for the first):
1. Give required-cleanup and draft-comment a default of @translate="no"
2. Clarify that the current @translate behavior always applies - it even
inherits for elements that do not usually get translated 3. Provide a
list of exceptions where @translate does not inherit

Any thoughts?

Thanks-

Robert D Anderson
IBM Authoring Tools Development
Chief Architect, DITA Open Toolkit
(507) 253-8787, T/L 553-8787