Re: ACTION-725: Send a list of content-types for which content transformation applies

Looking again at the list, I wonder if we should use the opposite view 
and list content types that a CT proxy must/should not transform. Below 
is a try to rationalize this for tomorrow's discussion:

1. Mobile content-types
-----------------------
Unless the user agent doesn't support these formats, a proxy MUST NOT 
restructure or recode responses that are in:
1. text/xml, if not XHTML or if identified as mobile content in DOCTYPE
2. application/xml, if not XHTML or if identified as mobile content in 
DOCTYPE
3. application/json
4. application/octet-stream
5. text/vnd.wap.wml
6. text/vnd.wap.wmlc


... where the list of mobile content doctypes typically is the one 
proposed by Sean [1]:
<!DOCTYPE html PUBLIC "-//OMA//DTD XHTML Mobile 1.2//EN"
"http://www.openmobilealliance.org/tech/DTD/xhtml-mobile12.dtd">
<!DOCTYPE html PUBLIC "-//WAPFORUM//DTD XHTML Mobile 1.1//EN"
"http://www.wapforum.org/DTD/xhtml-mobile11.dtd">
<!DOCTYPE html PUBLIC "-//WAPFORUM//DTD XHTML Mobile 1.0//EN"
"http://www.wapforum.org/DTD/xhtml-mobile10.dtd">
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML Basic 1.1//EN"
"http://www.w3.org/TR/xhtml-basic/xhtml-basic11.dtd">
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML Basic 1.0//EN"
"http://www.w3.org/TR/xhtml-basic/xhtml-basic10.dtd">

In other words: don't touch "data", and don't touch mobile content. 
Since it starts with a "Unless...", I guess we could simply have "A 
CT-proxy SHOULD NOT" instead of "Unless..., a CT-proxy MUST NOT"


2. Non-mobile content-types
---------------------------
The list of proprietary MIME types is an ever-growing one. We may list 
some of them as examples, but the list cannot be exhaustive.
I don't think we need to add anything regarding other content-types, in 
the sense that they may all be transformed, and we already have some 
text in 4.4. "A proxy SHOULD strive for the best possible user 
experience that the user agent supports [...]"

We may still want to add something along the lines of:

For content-types other than the ones listed above, a proxy SHOULD NOT 
restructure or recode:
a) a content-type that was explicitly listed in the Accept header of the 
request
b) a content-type that the CT-proxy knows is supported by the user agent
... unless:
i] network/memory considerations make the response too big
ii] some features of the underlying format, in use in the page, are 
known not to be supported by the user agent

In other words: don't touch formats that are supported by the user 
agent, unless not doing so would result in a poor user experience.



Note I'm always talking about restructuring and recoding, not 
optimizing. I think optimizing may still be done in all cases.

Francois.

[1] http://lists.w3.org/Archives/Public/public-bpwg-ct/2008May/0000.html



Sean Patterson wrote:
> For text/plain the main thing I had in mind was segmentation of long pages.  (RFCs and W3C recommendations for example :-)  Headers and/or footers might also be added.
> 
>> -----Original Message-----
>> From: Francois Daoust [mailto:fd@w3.org]
>> Sent: Tuesday, April 29, 2008 5:23 AM
>> To: Sean Patterson
>> Cc: public-bpwg-ct
>> Subject: Re: ACTION-725: Send a list of content-types for which content
>> transformation applies
>>
>> Thanks Sean!
>>
>> I was about to react on "text/xml" but remembered that it is indeed
>> possible to serve XHTML pages using this generic content type, as
>> explained in:
>> http://www.w3.org/TR/xhtml-media-types/
>> (I only mention it so that I don't forget about it again ;-))
>>
>> I'm more puzzled by "text/plain". What are the transformations that a
>> proxy may do to resources served with that content type? A conversion to
>> an HTML response?
>>
>> François.
>>
>>
>> Sean Patterson wrote:
>>> Here is a list of content types that a CT proxy may want/need to
>> transform:
>>> ***_Text and markup:_*
>>>
>>> text/html
>>>
>>> application/xhtml+xml  (if not identified as mobile content in DOCTYPE)
>>>
>>> application/xml  (if XTHML and not identified as mobile content in
>> DOCTYPE)
>>> text/xml  (if XTHML and not identified as mobile content in DOCTYPE)
>>>
>>> text/plain
>>>
>>> application/rtf
>>>
>>> ***_De facto_****_ standards and proprietary types_****_:_****__*
>>>
>>> application/pdf
>>>
>>> application/msword
>>>
>>> application/msexcel
>>>
>>> application/vnd.ms-excel
>>>
>>> application/vnd.ms-powerpoint
>>>
>>> application/mspowerpoint
>>>
>>> application/x-shockwave-flash
>>>
>>> ***_CSS:_*
>>>
>>> text/css
>>>
>>> ***_Images:_*
>>>
>>> image/gif
>>>
>>> image/jpeg
>>>
>>> image/png
>>>
>>> ***_JavaScript:_*
>>>
>>> application/x-javascript
>>>
>>> text/javascript
>>>
>>> I think this includes the most common content types that would be
>>> transformed, but it's possible I've left some off.  Of course there are
>>> a large number of uncommon content types that a particular CT proxy may
>>> want to transform.
>>>
>>> I've left multimedia types (audio, video) off the list for now.  Not
>>> sure if we want to say anything about those types.
>>>
>>> Sean
>>>
> 

Received on Monday, 5 May 2008 12:16:22 UTC