Re: [weirds] feedback on draft-designteam-weirds-using-http

[Francisco Obispo <[e-mail]>]

Did a bit more research and http://www.w3.org/International/questions/qa-controls proposes what I had in mind, which is encoding the content using something like base64 and decoding it before presentation. (ugly)

I might be missing something, but why are we arguing about this? I don't think we're planning to pass control codes within the XML.. 

Francisco


On May 17, 2012, at 1:24 PM, Julian Reschke wrote:

> On 2012-05-17 22:13, Francisco Obispo wrote:
>> If you need to specify non-ascii characters in an XML document, you either have to:
>> 
>> a) Encode the characters in their XML entities ñ ->  ñ  or
> 
> 1) You can use those "as-is", if you declare the document encoding properly (such as by using UTF-8).
> 
> 2) That doesn't change the repertoire of characters you can use.
> 
> 3) Also, it's ASCII control characters which are a problem.
> 
>> b) use a CDATA instruction: Unparsed Character Data [1]
>> 
>> [1] http://www.w3schools.com/xml/xml_cdata.asp
> 
> -> <http://w3fools.com/>
> 
>> The latter approach seems to be more efficient for some.
> 
> And it doesn't help at all. In CDATA, you don't need to escape "<" and "&", but in turn there are *other* sequences you can't use. See <http://www.w3.org/TR/REC-xml/#sec-cdata-sect> -- you can't have the string "]]>" inside CDATA because it ends the CDATA section.
> 
> My recommendation is never ever to use CDATA unless when hand-editing *and* when you absolute positively know what you do. If you generate with a program, don't ever use it.
> 
> Best regards, Julian

Francisco Obispo 
email: [e-mail]
Phone: [phone] || INOC-DBA *3557* NOC
PGP KeyID = B38DB1BE