W3C home > Mailing lists > Public > www-i18n-comments@w3.org > November 2014

Re: bug qa-translate-flag

From: Richard Ishida <ishida@w3.org>
Date: Wed, 26 Nov 2014 17:27:24 +0000
Message-ID: <54760D7C.4040002@w3.org>
To: Gunnar Bittersmann <gunnar@bittersmann.de>
CC: www-i18n-comments@w3.org
hi Gunnar,

see below...


On 01/09/2014 11:26, Gunnar Bittersmann wrote:
> Hi Richard,
> With deepest apologies, there are still things that slipped through
> while reviewing the article. In order of importance, AFAIS:
>
>
>
> A. http://www.w3.org/International/questions/qa-translate-flag#attributes
>
> »»
> This can, of course, cause problems in cases where you do want the
> attribute values to be translated but not the element content, or vice
> versa. In some cases those situations can be mitigated by nesting the
> markup concerned. For example, you could have an outer span element with
> translate set to yes that carries the title attribute you want to avoid
> translating.
> ««
>
> It’s not clear that “you want to avoid translating” should refer to the
> “outer span element”. As written, it refers to the “title attribute”
> which is wrong. This sounds as if the title attribute value should not
> be translated while in fact it should be. It’s the element content that
> should not be translated. Only then the following makes sense:
>
> »»
> Inside that span you could put another span with translate set to no and
> containing the element content. This is how articles in this series
> handle links to translated versions of a page – the title attribute of
> the outer element carries the name of the language pointed to, and the
> inner element carries the name of that language in the language itself
> (which should not be changed). This also helps when labelling the
> language using the lang attribute.
> ««
>
> Please reformulate the sentence in question.

Thanks for spotting that. Reworded.


> B. http://www.w3.org/International/questions/qa-translate-flag#why
>
> »»
> <p>Click the Resume button on the Status Display or the
> <span class="panelmsg" translate="no">CONTINUE</span> button
> on the printer panel.</p>
> ««
>
> I think this sample would be clearer if the output of a translation
> (where “CONTINUE” and surrounding text would differ in language) would
> be shown below, e.g. into German:
>
> <p>Drücken Sie Fortsetzen in der Statusanzeige oder die
> Taste <span class="panelmsg" translate="no">CONTINUE</span>
> an Ihrem Drucker.</p>
>
> The HTML source code would be:
>
> <p>In German translation this will become:</p>
> <figure class="example"><p><code>&lt;p&gt;Drücken Sie Fortsetzen in der
> Statusanzeige oder die<br>
> Taste &lt;span class="panelmsg" translate="no"&gt;<span
> translate="no">CONTINUE</span>&lt;/span&gt;<br>
> an Ihrem Drucker.&lt;/p&gt;</code></p>
> </figure>

Note that the translate=no attribute should go on the figure element, 
not the span around CONTINUE, since this is billed as a German 
translation. It should therefore remain a German translation.

> I’ve added this into my translation
> http://dev.bittersmann.de/International/questions/qa-translate-flag.de#why
> You might want to consider having it in the English original.


Ok. Added.


> C.
> http://www.w3.org/International/questions/qa-translate-flag#yyyshortcomings
>
> »»
> For example, we may want to allow the natural language text of the above
> source code to be translated, while protecting the code itself (ie. the
> keywords such as label, for, postcode, input, etc.). We could do that by
> surrounding the natural language text with elements that have the
> translate attribute.
> ««
>
> sounds as if this would work with current online translation services,
> while down below in
> http://www.w3.org/International/questions/qa-translate-flag#stickyness
> it says
>
> »»
> Microsoft and Google's translation engines also don't translate content
> within code elements. Note, however, that you don't seem to have any
> choice about this – there don't seem to be instructions about how to
> override this if you do want your code element content translated.
> ««
>
> How about adding the info that translating natural language in code
> samples is wishful thinking current online translation services already
> in the first place?


Before rewording the text I (rewrote and) re-ran the tests, and updated 
the results page. Now it's only Microsoft services that don't do the 
yes-inside-no thing.  I think the text is therefore ok as is.


> D. still
> http://www.w3.org/International/questions/qa-translate-flag#yyyshortcomings
>
> Shouldn’t the section identifier "yyyshortcomings" be renamed to
> something that fits to the heading “When to use translate="yes"”?


Yes. Done.


> E. still
> http://www.w3.org/International/questions/qa-translate-flag#yyyshortcomings
>
> »»
> <p>The <code class="kw">yes</code> value of the <code
> class="kw">translate</code> attribute is mostly used to override the
> effect of setting translate to <code class="kw">no</code>.
> ««
>
> The second “translate” should also be marked-up as keyword, should it?
>
> <p>The <code class="kw">yes</code> value of the <code
> class="kw">translate</code> attribute is mostly used to override the
> effect of setting <code class="kw">translate</code> to <code
> class="kw">no</code>.

Fixed.



> F. http://www.w3.org/International/questions/qa-translate-flag#how
>
> »»
> <a
> href="http://www.w3.org/TR/its20/#EX-translate-selector-1">Internationalization
> Tag Set (ITS)</a> specification.
> ««
>
> and below
>
> »»
> on the Internationalization Tag Set specification
> ««
>
> The specification title should be set in italics, i.e. marked-up using a
> cite element?

Fixed. Also for Metadata for the Multilingual Web – Usage Scenarios and 
Implementations.


> G.
> http://dev.bittersmann.de/International/questions/qa-translate-flag.de#why
>
> In the example with “french pain”, shouldn’t “pain” be tagged as French
> using <span lang="fr"> (giving a good example)?


That's the actual code used, so I don't want to change it.


> H. http://www.w3.org/International/questions/qa-translate-flag#attributes
>
> Using <sup>1</sup> leads to increased line height, destroying the
> vertical rhythm.
>
> Suggestion: Use Unicode character ¹ or asterisk * as sidenote indicator.

Should be ok now, with the new styling.


> Hopefully, I’ve found all the nits to pick now.

Actually I found a few more things to change while I was at it. I assume 
you have a diff tool that can easily show changes, rather than me having 
to list everything. I'll send the new source file off-list.

Thanks for the comments. And apologies for taking so long to address them.

ri
Received on Wednesday, 26 November 2014 17:28:39 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:20:17 UTC