[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Registering a charset alias
On 2009/08/15 0:33, Erik van der Poel wrote:
> (1) Adding new aliases. Relatively uncontroversial, but at some point,
> we have to stop. That stopping point may be slightly controversial.
Before we stop, we have to start. I suggest Anne go ahead. If anybody
else wants to help, that's fine, too.
> (2) Merging charsets. Some IANA charsets have distinct names, but are
> treated as aliases in major implementations.
If "major implementations" means "major browser implementations", then
that's way not enough, because the IANA charset registry has much wider
usage.
> (3) Supersets. Perhaps in a separate file, not in the same file as the
> normal charset names and aliases.
>
> (4) Preferred MIME names. Some charsets have aliases but do not have
> "preferred MIME name" attached to any of them.
That's fine. I think the rule should be:
1) If the Name or an Alias is labeled as "(preferred MIME name)", then
use that.
2) Else, use the name.
Anne (or somebody else), can you make sure that gets into the authors
section of HTML5?
> It might be a bit too ambitious (or even unnecessary) to include all
> of these in a single coordination call.
Please don't expect a coordination call to solve any actual problems.
What it may help is to get a feeling for who might do what.
> We also need to remember that charsets are used in protocols other
> than Web protocols (e.g. email).
Very true indeed.
Regards, Martin.
> Erik
>
> 2009/8/14 Patrik Fältström<paf@cisco.com>:
>> On 14 aug 2009, at 16.34, Anne van Kesteren wrote:
>>
>>> One of our main goals with HTML5 has been to make the browser space more
>>> competitive.
>> Which of course is a good goal!
>>
>>> This includes documenting in great detail what rules browsers need to
>>> implement in order to render legacy Web documents. This in turn includes
>>> such rules as treating iso-8859-1 as if windows-1252 was specified and
>>> having x-x-big5 be an alias for big5.
>> Well, I want to once again point out that they should not be aliases. One
>> might be a superset of the other, but not an alias.
>>
>> For example, the Windows-1252 codepage coincides with ISO-8859-1 in the code
>> ranges 0x00 to 0x7F and 0xA0 to 0xFF, but not for the range 0x80 to 0x9F.
>> And that is a very important distinction that from my point of view
>> _absolutely_ should be pointed out in any text that suggest such
>> equivalences.
>>
>> But, lets take this up at a coordination call.
>>
>> Patrik
>>
>>
>
--
#-# Martin J. Dürst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp mailto:duerst@it.aoyama.ac.jp