Re: [widgets] P&C 1.0 Last Call WD: localization comments

Hi Jere,  
On 1/14/09 3:28 PM, "Jere Kapyaho" <jere.kapyaho@nokia.com> wrote:
> Hi Marcos,
> 
> I have (still) a couple of concerns about the localization section of
> Widgets Packaging & Configuration Last Call WD of 20081222.
> 
> /1/ Is the following statement in [1] as it should be?
> 
> "Author requirements: Localized folders must be at the root of the widget (a
> localized folder not at the root of the widget will be treated as an
> arbitrary folder)."
> 
> I think it should now read:
> 
> "Localized folders must be placed inside the container for localized content
> (...)".

Fixed. 

> 
> /2/ I was looking at the localized widget example in the same section (it's
> non-normative, but important nonetheless), and it seems that the left and
> right sides of the example don't agree. The paths on the right are absolute
> from the root, not the container.

Fixed. 

 
> The right side shows en-au, but that is not found on the left side. The
> first bullet of the example mentions Korean, but that does not seem to be
> present. Should it be Spanish instead?

Fixed.  
 
> Finally, it's not obvious which of the files shown (if any) is the start
> file, because the content of config.xml is not shown.
> 

I've now added /config.xml to the right side. It includes a <content/>
element which hopefully makes things more clear.

> /3/ This is a potentially confusing statement:
> 
> "At runtime, the widget user agent will set the (HTML or XML) base of the
> start file to the localized folder (even if the start file does not reside
> inside the localized folder)."
> 
> I assume in this case "the localized folder" means the one determined in
> Step 6, right? This might not be obvious from the context, it requires a
> trip to the text of Step 6.

This is correct. I've added a link to step 6 in the text: "please refer to
step 6" (does that help at all or should I elaborate more on it in the
spec?)
 
> /4/ Since BCP47 tags are case-insensitive, it might be good to normalise all
> their occurrences to lowercase, to avoid any confusion.

Ok, I wrote into step 3 that the wua-language list must be normalized to
lowercase form. I modified step 6 also so the comparison is done in lower
case. Can you please check that it is ok?
 
> /5/ Is there value in being able to have multiple config.xml documents,
> instead of just one, and tagging the relevant elements with a BCP47 tag? The
> example mentions "different author and license", which could be expressed in
> one file. If you have a pointer to some earlier discussion that resolves
> this, it's fine. Or do you think it would make the config.xml document too
> bulky?

Although we have had these discussions in the past, I don't have pointers
but I am happy to summarize our thinking here. There are a number of reasons
why we went with the multiple config file approach:

 1. the XML I18n best practices guidelines says it that we should have
multiple documents. See http://www.w3.org/TR/xml-i18n-bp/#DevMLDoc
 2. Like you said, makes the config files less bulky and, IMHO, easier to
maintain.
 3. Makes processing the XML easier and more predictable.

Also, in a separate email I noticed you raised concerns about the use of
"not required" in the Widgets digsig spec. I noticed that I had introduced
the same problem into the the P&C spec. I have gone through and changed all
instances of "not required" to use the word "optional". I'll make sure that
doesn't happen in the other specs too.

For the purpose of LC disposition of comments, can you please indicate if
you are satisfied with the working group's response.

Kind regards,
Marcos

Received on Saturday, 17 January 2009 20:57:02 UTC