RE: ACTION-447: Make a batch transformation of the test suite to xliff

Hi Felix,

See inline..

From: Felix Sasaki [mailto:fsasaki@w3.org]
Sent: Thursday, February 21, 2013 12:44 PM
To: Yves Savourel
Cc: public-multilingualweb-lt@w3.org
Subject: Re: ACTION-447: Make a batch transformation of the test suite to xliff

Am 21.02.13 21:24, schrieb Yves Savourel:
> Independent of whether this is normative or not: how would the guidance look like?

Maybe something like:

In order to have identical default behavior, ITS processors SHOULD apply the following default ITS rules when processing HTML5 (and the link to the rules file).

OK ... but in the example below Fredrik said that Okapi specifies keywords to be translatable as a default. Can we expect that this is a default for all HTML filters? And related: Fredrik wrote "One of the default global html5 rules (in Okapi) specifies <meta name="keywords"…’s content to be translatable". Is this really a global rule (which somebody who knows where it is stored could modify), or a default, non ITS rules based processing?

[F: Those two examples are rules in the Okapi implementation so I guess it’s up for discussion if they should are should be generally suggested or not. In Okapi default global rules is in an external file but it’s not linked from within the html5 files (href) so in that sense it’s invisible. But from within Rainbow you can see the ITS rules that are applied by default and you can chose to modify them (create a copy and modify, without modifying the html file itself). Maybe the problem is with Okapi in the way we’re using those invisible rules? But at the same time we want to isolate the user from ITS if not needed.]

Also, would the link be to a rules file in the spec or in the wiki (so that more easily it could be updated)? For the need to do that see e.g. here: HTML5.1 will have new elements, e.g.
https://dvcs.w3.org/hg/html-extensions/raw-file/tip/maincontent/index.html

and we state we are covering HTML5 or its sucessor with ITS2.

[F: I would think that it would be linked to from the wiki with some change history, I don’t think the spec should dictate which data categories should be applied to individual elements of a specific format implementation]

Also ... in the example below two data categories are processed by Okapi: Translate and Domain. What do we expect from a tool that only implements domain, with regards to the defaults of data categories not being implemented?

[F: If a tool only implements domain then it can only use a suggested default rule for domain. For other categories not implemented content should be passed through as is without expectations of defaults since the tool is not doing anything with it.]

And asking again (I may have missed the answer, in that case sorry for that): why was this not needed for ITS 1.0? See
http://lists.w3.org/Archives/Public/public-multilingualweb-lt/2013Feb/0159.html

("what I still don't get" part).

[F: At Enlaso we didn’t use ITS 1.0 to process html files, only xml. That’s why it didn’t come up earlier.]

Finally, let me ask about the

[F: yes? ☺]

- Felix


-ys



From: Felix Sasaki [mailto:fsasaki@w3.org]
Sent: Thursday, February 21, 2013 1:05 PM
To: public-multilingualweb-lt@w3.org<mailto:public-multilingualweb-lt@w3.org>
Subject: Re: ACTION-447: Make a batch transformation of the test suite to xliff

Am 21.02.13 20:55, schrieb Fredrik Liden:

Hi Marcin and Yves,



Regarding the comment for domain4html.html. Trans-unit id=1 is actually the extraction of the content of <meta name="keywords"…

One of the default global html5 rules (in Okapi) specifies <meta name="keywords"…’s content to be translatable and a second rule specifies that it contains the default domain for the entire <html> element.



The <script> rules applies the combined rules for the <body> element only. So what you’re seeing is the <p> (trans-unit=2) with the combined domains. But <meta name="keywords"… being outside the body only has the default domain applied.



Another example why we need some guidelines/expectations for html 5 behavior. ☺

Independent of whether this is normative or not: how would the guidance look like?

Best,

Felix






Cheers,

Fredrik

domain4html.html
<!DOCTYPE html>
<html>
<head>
<meta charset=utf-8>
<meta name="keywords" content="SPORTS LAW, Judicial Matters"/>
<meta name="x-mykeywords" content="Sport, Law "/>
<script type="application/its+xml">
<its:rules xmlns:its="http://www.w3.org/2005/11/its"<http://www.w3.org/2005/11/its> xmlns:h="http://www.w3.org/1999/xhtml"<http://www.w3.org/1999/xhtml> version="2.0">
<its:param name="domainParam">keywords</its:param>
<its:domainRule selector="/h:html/h:body" domainPointer="/h:html/h:head/h:meta[@name='x-mykeywords' or @name=$domainParam]/@content" domainMapping="'sports law' LAW, 'labor law' LAW, 'contract law' LAW, 'competition law' LAW,'tort law' LAW"/>
</its:rules>
</script>
</head>
<body>
<p>Some text about sport and law.</p>
</body>
</html>

domain4html.html.xlf
<?xml version="1.0" encoding="UTF-8"?>
<xliff version="1.2" xmlns="urn:oasis:names:tc:xliff:document:1.2" xmlns:okp="okapi-framework:xliff-extensions" xmlns:its="http://www.w3.org/2005/11/its"<http://www.w3.org/2005/11/its>>
<file original="/Copy of domain4html.html" source-language="en-us" target-language="fr-fr" datatype="html">
<body>
<trans-unit id="1" okp:itsDomain="SPORTS LAW, Judicial Matters">
<source xml:lang="en-us">SPORTS LAW, Judicial Matters</source>
<target xml:lang="fr-fr">SPORTS LAW, Judicial Matters</target>
</trans-unit>
<trans-unit id="2" okp:itsDomain="Sport, Law, SPORTS LAW, Judicial Matters">
<source xml:lang="en-us">Some text about sport and law.</source>
<target xml:lang="fr-fr">Some text about sport and law.</target>
</trans-unit>
</body>
</file>
</xliff>






-----Original Message-----
From: Yves Savourel
Sent: Tuesday, February 19, 2013 6:03 AM
To: 'Mārcis Pinnis'; 'Multilingual Web LT Public List Public List'
Subject: RE: ACTION-447: Make a batch transformation of the test suite to xliff



Hi Mārcis,



I missed a few comments in your docx file.

Here are the file with my additional notes (nothing major).



(BTW: your comment about Domain  in domain4html.html is interesting.

I'll try to look at the test output and see if it matches the info output in the XLIFF file.

If it does, this may be an interesting overriding case.)



-ys







-----Original Message-----

From: Mārcis Pinnis [mailto:marcis.pinnis@Tilde.lv]

Sent: Tuesday, February 19, 2013 5:24 AM

To: Yves Savourel; 'Multilingual Web LT Public List Public List'; Dave Lewis (dave.lewis@cs.tcd.ie<mailto:dave.lewis@cs.tcd.ie>)

Cc: Felix Sasaki (fsasaki@w3.org<mailto:fsasaki@w3.org>)

Subject: RE: ACTION-447: Make a batch transformation of the test suite to xliff



Hi Yves, all,



I had a look at the examples. I believe that either I am missing something (not understanding where the ITS 2.0 data is in the XLIFF documents) or there is some backwards compatibility of content lost when converting from the HTML/XML examples to XLIFF.



1. I had a look at the Terminology part and I could not find ITS 2.0 related terminology annotation in the XLIFF documents. I have attached my findings to this e-mail.



2. With the Locale Filter I see that instead of having ITS 2.0 mark-up, the whole fragment has been removed and replaced with a placeholder (is that because it is not possible to add Locale Filter mark-up in XLIFF at all?). This does not preserve the content, but filters out fragments based on ITS 2.0 consumption/production Use Case scenarios (which is I guess an internal process and not for data exchange purposes). And ... it actually does not show an XLIFF document with the Locale Filter data category metadata in it (that was what we wanted to see, but the examples, I believe do not show that). Is this because XLIFF would not be able to handle ITS 2.0 annotation or because of some other reasons (I am a bit confused here ... so I would like to clarify)?



Some other findings (more in the attached file) 3. The Language Information as I understand it, will be fully passed on to xml:lang (that is clear).

4. The Domain metadata seems to be transformed from ITS into an OKAPI internal structure.

5. The Elements Within Text information as I understand it, is just structural, so no mark-up is necessary (that is clear).



Maybe I have just misunderstood what the XLIFF examples would contain? I had the understanding that the transformation to XLIFF would preserve ITS 2.0 metadata. Did I understand it wrong?



Then ... I had a look also at the files in the "roundtrip-example" directory. As I understand from Yves e-mail, these are not valid XLIFF files, right?!



I still had a look at the examples that contained terminology annotation. I believe Terminology is used incorrectly:

<mrk its:terminology="yes" its:termInfoRef="#ge1">Arizona</mrk>

The attribute is its:term="yes" rather than terminology... (or am I again missing out some information?)



The files seemed not to have Domain and LocaleFilter metadata in them - it would be great to see these categories in action as well.



Best regards,

Mārcis ;o)



-----Original Message-----

From: Yves Savourel [mailto:ysavourel@enlaso.com]

Sent: Monday, February 18, 2013 4:52 PM

To: 'Multilingual Web LT Public List Public List'

Subject: ACTION-447: Make a batch transformation of the test suite to xliff



Hi all,



I've done this action item.



A batch file as well as the XLIFF output have been added to GitHub:

https://github.com/finnle/ITS-2.0-Testsuite/commit/294018ba576799dcbee7b9566da83837dd69f4ae




Notes:



-- The XLIFF outputs are often identical because the test files are just different ways to markup the same content.



-- The XLIFF output often make little sense because the input exercises only one data category. For example, a storage size limitation set on a span ("inline") element will not show up on an inline element in XLIFF because there is no information in the input file that says the span element is 'within text' (since the test case is about the storage size). IHMO the output are rather useless.



-- Most data categories have output, but only when the extraction use them. For example there is no output for directionality because, while the Okapi ITS engine process and provides that data category, the filter does nothing with it.



Cheers,

-yves

Received on Thursday, 21 February 2013 23:52:14 UTC