Re: Requesting IESG Approval for the Media Type application/xslt+xml, application/xquery+xml, and application/xquery

[Alexey Melnikov]

Alexey Melnikov wrote:

> Philippe Le Hegaret wrote:
>
>> Hello Alexey,
>
Hi Philippe,
I've done a detailed review of the 3 registrations. application/xslt+xml
and application/xquery+xml look fine to me (I might have some non
blocking nits I report separately). I have some further comments
regarding application/xquery, I am sorry that I haven't spotted them
earlier.

So, how do you want to proceed?
I suggest that I separate IESG review of application/xslt+xml, from
review of application/xquery+xml and application/xquery. This way
application/xslt+xml can be approved on September 24th and the remaining
2 can be dealt at a later date.

>> I don't see the registration of those media types on the agenda of the
>> IESG:
>> https://datatracker.ietf.org/iesg/agenda/
>
> Hi Philippe,
> I didn't realize you wanted IESG to approve the updated registration.
>
>> What is the next step to get those registered?
>
> I can add this as a management item for the next IESG telechat. After
> the telechat I can tell you if there are any comments from IESG.

[Alexey Melnikov]

On Sat, Sep 12, 2009 at 2:50 PM, Alexey Melnikov
<alexey.melnikov@...> wrote:

> Alexey Melnikov wrote:
>
>> Philippe Le Hegaret wrote:
>>
>>> Hello Alexey,
>>
> Hi Philippe,
> I've done a detailed review of the 3 registrations. application/xslt+xml and
> application/xquery+xml look fine to me (I might have some non blocking nits
> I report separately). I have some further comments regarding
> application/xquery, I am sorry that I haven't spotted them earlier.
>
> So, how do you want to proceed?
> I suggest that I separate IESG review of application/xslt+xml, from review
> of application/xquery+xml and application/xquery.
> This way
> application/xslt+xml can be approved on September 24th

This is done.

> and the remaining 2
> can be dealt at a later date.

[Alexey Melnikov]

Alexey Melnikov wrote:

> Alexey Melnikov wrote:
>
>> Philippe Le Hegaret wrote:
>>
>>> Hello Alexey,
>>
> Hi Philippe,
> I've done a detailed review of the 3 registrations.
> application/xslt+xml and application/xquery+xml look fine to me (I
> might have some non blocking nits I report separately). I have some
> further comments regarding application/xquery, I am sorry that I
> haven't spotted them earlier.

In particular:

1).
<http://www.w3.org/TR/2009/PER-xquery-20090421/#id-registration-of-mime-type>
is using own order of sections (and sometimes uses different section
names) from the template prescribed in Section 10 of RFC 4288. I thought
information about Magic numbers is missing, but I finally found section
"G.4 Recognizing XQuery Files", which covers that. In the future, I
would ask W3C to use the correct template, as it makes checking
registrations easier.

2).

> The syntax of XQuery is expressed in Unicode but may be written with
> any Unicode-compatible character encoding, including UTF-8 or UTF-16,
> or transported as US-ASCII or Latin-1 with Unicode characters outside
> the range of the given encoding represented using an XML-style ෝ
> syntax.

Is there any good reason for allowing Latin-1? IETF pretty much settled
on only using US-ASCII, UTF-8 (and rarely UTF-16).

[Liam Quin]

On Thu, Oct 01, 2009 at 10:14:57PM +0100, Alexey Melnikov wrote:
> <http://www.w3.org/TR/2009/PER-xquery-20090421/#id-registration-of-mime-type>
> is using own order of sections (and sometimes uses different section  
> names) from the template prescribed in Section 10 of RFC 4288. I thought  
> information about Magic numbers is missing, but I finally found section  
> "G.4 Recognizing XQuery Files", which covers that. In the future, I  
> would ask W3C to use the correct template, as it makes checking  
> registrations easier.

I may have used an outdated template - XQuery was either the first or one
of the first specs to use this new combined process.

> 2).
>> The syntax of XQuery is expressed in Unicode but may be written with  
>> any Unicode-compatible character encoding, including UTF-8 or UTF-16,  
>> or transported as US-ASCII or Latin-1 with Unicode characters outside  
>> the range of the given encoding represented using an XML-style ෝ  
>> syntax.
>
> Is there any good reason for allowing Latin-1? IETF pretty much settled  
> on only using US-ASCII, UTF-8 (and rarely UTF-16).

To the best of my memory it was for in response to a comment regarding
HTTP cpmpatibility, but, there is also no good reason to forbid it at
this point, with over 50 XQuery implementations in the field.

For XQuery 1.1 we could possibly disallow Latin-1, but I don't think
we'd gain anything now.  If we were still in 2005, we'd be in a
position to change such things, and I think do a better job.
Removing it could only hurt interop now I think.

Thanks for the comments!

Liam

--
Liam Quin, W3C XML Activity Lead, http://www.w3.org/People/Quin/
http://www.holoweb.net/~liam/ * http://www.fromoldbooks.org/

["Martin J. Dürst"]

On 2009/10/02 10:30, Liam Quin wrote:
> On Thu, Oct 01, 2009 at 10:14:57PM +0100, Alexey Melnikov wrote:
>> <http://www.w3.org/TR/2009/PER-xquery-20090421/#id-registration-of-mime-type>

>> 2).
>>> The syntax of XQuery is expressed in Unicode but may be written with
>>> any Unicode-compatible character encoding, including UTF-8 or UTF-16,
>>> or transported as US-ASCII or Latin-1 with Unicode characters outside
>>> the range of the given encoding represented using an XML-styleෝ
>>> syntax.
>> Is there any good reason for allowing Latin-1? IETF pretty much settled
>> on only using US-ASCII, UTF-8 (and rarely UTF-16).
>
> To the best of my memory it was for in response to a comment regarding
> HTTP cpmpatibility, but, there is also no good reason to forbid it at
> this point, with over 50 XQuery implementations in the field.
>
> For XQuery 1.1 we could possibly disallow Latin-1, but I don't think
> we'd gain anything now.  If we were still in 2005, we'd be in a
> position to change such things, and I think do a better job.
> Removing it could only hurt interop now I think.

I think
http://www.w3.org/TR/2009/PER-xquery-20090421/#dt-encoding-declaration
makes it clear that XQuery documents can be in very much any encoding.
So the mention of Latin-1 in the Mime type registration is just an
example. As such, it can stay in the registration template or be removed
without any effect on interoperability.

[charset expert reviewer hat on]
However, there is no 'charset' "Latin-1" registered at
http://www.iana.org/assignments/character-sets. So I strongly suggest to
replace it with ISO-8859-1, or leave it out.
[hat off]

Regards,    Martin.

--
#-# Martin J. Dürst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@...

[Alexey Melnikov]

Liam Quin wrote:

>On Thu, Oct 01, 2009 at 10:14:57PM +0100, Alexey Melnikov wrote:
>  
>
 [...]

>>2).
>>    
>>
>>>The syntax of XQuery is expressed in Unicode but may be written with  
>>>any Unicode-compatible character encoding, including UTF-8 or UTF-16,  
>>>or transported as US-ASCII or Latin-1 with Unicode characters outside  
>>>the range of the given encoding represented using an XML-style ෝ  
>>>syntax.
>>>      
>>>
>>Is there any good reason for allowing Latin-1? IETF pretty much settled  
>>on only using US-ASCII, UTF-8 (and rarely UTF-16).
>>    
>>
>To the best of my memory it was for in response to a comment regarding
>HTTP cpmpatibility, but, there is also no good reason to forbid it at
>this point, with over 50 XQuery implementations in the field.
>
>For XQuery 1.1 we could possibly disallow Latin-1, but I don't think
>we'd gain anything now.  If we were still in 2005, we'd be in a
>position to change such things, and I think do a better job.
>Removing it could only hurt interop now I think.
>  
>

If this is done for backward compatibility, this is Ok with me.

Please address Martin's concern about Latin-1 not being registered.

[Liam Quin]

On Fri, Oct 02, 2009 at 05:02:19PM +0900, "Martin J. Dürst" wrote:
> I think
> http://www.w3.org/TR/2009/PER-xquery-20090421/#dt-encoding-declaration
> makes it clear that XQuery documents can be in very much any encoding.  
Yes.

> So the mention of Latin-1 in the Mime type registration is just an  
> example. As such, it can stay in the registration template or be removed  
> without any effect on interoperability.
>
> [charset expert reviewer hat on]
> However, there is no 'charset' "Latin-1" registered at  
> http://www.iana.org/assignments/character-sets. So I strongly suggest to  
> replace it with ISO-8859-1, or leave it out.
> [hat off]

I think we can make that change :-)

Thanks for the comments!


Liam

--
Liam Quin, W3C XML Activity Lead, http://www.w3.org/People/Quin/
http://www.holoweb.net/~liam/ * http://www.fromoldbooks.org/

[Liam Quin]

On Fri, Oct 02, 2009 at 11:42:18AM +0100, Alexey Melnikov wrote:
[...]
>> For XQuery 1.1 we could possibly disallow Latin-1, but I don't think
>> we'd gain anything now.  If we were still in 2005, we'd be in a
>> position to change such things, and I think do a better job.
>> Removing it could only hurt interop now I think.
>>  
>>
> If this is done for backward compatibility, this is Ok with me.
>
> Please address Martin's concern about Latin-1 not being registered.

I've filed an issue/bug against the document [1].

Liam

[1] http://www.w3.org/Bugs/Public/show_bug.cgi?id=7797

--
Liam Quin, W3C XML Activity Lead, http://www.w3.org/People/Quin/
http://www.holoweb.net/~liam/ * http://www.fromoldbooks.org/

[Liam Quin]

On Fri, Oct 02, 2009 at 05:02:19PM +0900, "Martin J. Dürst" wrote:

> I think
> http://www.w3.org/TR/2009/PER-xquery-20090421/#dt-encoding-declaration
> makes it clear that XQuery documents can be in very much any encoding.  
> So the mention of Latin-1 in the Mime type registration is just an  
> example. As such, it can stay in the registration template or be removed  
> without any effect on interoperability.
>
> [charset expert reviewer hat on]
> However, there is no 'charset' "Latin-1" registered at  
> http://www.iana.org/assignments/character-sets. So I strongly suggest to  
> replace it with ISO-8859-1, or leave it out.
> [hat off]

Today the XQuery WG agred to make that change , both for
XQuery 1.1 (in development) and as an errata to XQuery 1.0.

Public record of decision:

http://www.w3.org/Bugs/Public/show_bug.cgi?id=7797

Liam

--
Liam Quin, W3C XML Activity Lead, http://www.w3.org/People/Quin/
http://www.holoweb.net/~liam/ * http://www.fromoldbooks.org/

[Alexey Melnikov]

Hi,

Liam Quin wrote:

>On Fri, Oct 02, 2009 at 05:02:19PM +0900, "Martin J. Dürst" wrote:
>  
>
>>I think
>>http://www.w3.org/TR/2009/PER-xquery-20090421/#dt-encoding-declaration
>>makes it clear that XQuery documents can be in very much any encoding.  
>>So the mention of Latin-1 in the Mime type registration is just an  
>>example. As such, it can stay in the registration template or be removed  
>>without any effect on interoperability.
>>
>>[charset expert reviewer hat on]
>>However, there is no 'charset' "Latin-1" registered at  
>>http://www.iana.org/assignments/character-sets. So I strongly suggest to  
>>replace it with ISO-8859-1, or leave it out.
>>[hat off]
>>    
>>
>Today the XQuery WG agred to make that change , both for
>XQuery 1.1 (in development) and as an errata to XQuery 1.0.
>
>Public record of decision:
>
>http://www.w3.org/Bugs/Public/show_bug.cgi?id=7797
>

Thanks.
Let me know when the document is updated with new text and I will ask
IESG to review the registration.