W3C home > Mailing lists > Public > public-i18n-its@w3.org > April to June 2006

RE: withinText

From: Yves Savourel <yves@opentag.com>
Date: Wed, 26 Apr 2006 10:55:33 +0200
To: "'Andrzej Zydron'" <azydron@xml-intl.com>
Cc: <public-i18n-its@w3.org>
Message-ID: <001801c6690f$2e7233a0$640fa8c0@Breizh>

Hi Andrzej,

I'm CCing the list so all can follow the discussion. Sorry I forgot to do that in the first place.

>> <p><subflow>Some text</subflow>Text of the paragraph</p>
>> <p>Text of the paragraph <subflow>Some text</subflow></p>

OK, so if we need to treat these two cases the same way, then yes, a subflow attribute is probably needed to reduce the computing to
resolve the first case.

Now, the last question: What do we do when we get elements not declare as withinText inside a text run? (because no matter what any
guideline says we just know it will happen, so we might as well make sure we have planned for it): I see three possibilities:
A) we treat it as a break (spliting the parent content into 2 text unit), plus it own content in-between.
B) we treat it as a withinText,
C) we treat it as a sub-flow (current behavior from the spec).

The most likely cases of such un-declared elements would be the things like <p> inside a <li> (like in XHTML).

I'm almost tempted to say A) since it would allow people to choose how to deal with elements like <br/> (either as withinText or as
a break).


-----Original Message-----
From: Andrzej Zydron
Sent: Tuesday, April 25, 2006 5:13 PM
To: Yves Savourel
Subject: Re: withinText

Hi Yves,

Thank you very much for your reply. I think that we are getting 
somewhere. My replies below:

Yves Savourel wrote:
> Hi Andrzej,
> With regard to withinText:
> I've started the implementation as well and I'm running into different types of issues in cases where one would forget to declare
> element like <b> should be specify as withinText.
> For the sub-flow elements: let's say if we have a sub-flow attribute: will you have a different behavior if the preceeding sibling
> has a text node? For example, will we get two different representations for these two different entries?
> <p><subflow>Some text</subflow>Text of the paragraph</p>
> <p>Text of the paragraph <subflow>Some text</subflow></p>

No - I would always treat these in a uniform fashion by always 
extracting the subflow to a new trans-unit:

<trans-unit id="t1">
   <source><ph id="s1" xid="t2"/>Text of the paragraph</source>
   <target><ph id="t1" xid="t2"/>Text of the paragraph</target>
<trans-unit id="t2">
   <source><g id="sg1" xid="s1">Some text</g></source>
   <target><g id="tg1" xid="t1">Some text</g></target>
<trans-unit id="t3">
   <source>Text of the paragraph <ph id="s2" xid="t4"/></source>
   <target>Text of the paragraph <ph id="t2" xid="t4"/></target>
<trans-unit id="t4">
   <source><g id="sg2" xid="s2">Some text</g></source>
   <target><g id="tg2" xid="t2">Some text</g></target>

I hope the choice of XLIFF inline elements is OK.

> One reason I would see the use for the subflow attribute is that I'm not sure anymore that no-inline elements that are within a
> run should always be treated as subflow. Maybe there are cases where they should yield segmentation. For example, a <br/> element:
> currently it would be either a withinText or an empty subflow (so treated like an inline I suppose), but we would have no way to
> make it a segment breaker if we want to.
> In the other hand, there is still that <p> element in <li> that bother me, especially in cases like the OpenDocument footnote:
> <text:p text:style-name="Standard">
> The Palouse horses 
> <text:note text:id="ftn1" text:note-class="footnote">
> <text:note-citation>1</text:note-citation> 
> <text:note-body>
> <text:p text:style-name="Footnote">A Palouse hors is the same thing as an Applalosa.</text:p> 
> </text:note-body>
> </text:note>
> have spotted coats. 
> </text:p>
> Any thoughts?
<trans-unit id="t1">
   <source>The Palouse horses <ph id="s1" xid="t2"/>have spotted 
   <target>The Palouse horses <ph id="t1" xid="t2"/>have spotted 
<trans-unit id="t2">
   <source><g id="sg1" xid="s1">A Palouse hors is the same thing as an 
   <target><g id="tg1" xid="s1">A Palouse hors is the same thing as an 

Received on Wednesday, 26 April 2006 09:02:58 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:43:07 UTC