request from HTTP draft editors for charset registration element
Larry Masinter
procedures, e.g., draft-ietf-822ext-mime-reg-03.txt).
In the course of preparing the HTTP 1.1 draft, we wanted to suggest
some preferred names for charsets as registered by IANA.
Right now, charsets are registered with many aliases. However, for use
in HTTP, all of the aliases are really cumbersome. We would like
senders to preferentially send
ISO-8859-1
instead of any of the other aliases for that charset.
Is this a possibility?
Where is the revised charset registration specification, since it
isn't in reg-03?
Larry
Ned Freed
> (I'm told this is the right list to comment on MIME registration
> procedures, e.g., draft-ietf-822ext-mime-reg-03.txt).
> In the course of preparing the HTTP 1.1 draft, we wanted to suggest
> some preferred names for charsets as registered by IANA.
The concept of a "primary name" whose use is preferred over any other
name has been part of the draft procedure for some time. I think this
meets your criteria for a preferred name.
> Right now, charsets are registered with many aliases. However, for use
> in HTTP, all of the aliases are really cumbersome. We would like
> senders to preferentially send
> ISO-8859-1
> instead of any of the other aliases for that charset.
> Is this a possibility?
See above -- I think it is already a done deal. (Unless someone objects,
of course.)
> Where is the revised charset registration specification, since it
> isn't in reg-03?
It has been submitted as an I-D and should be out shortly.
Ned
Larry Masinter
I can't find any place that lists the preferred names that the HTTP
protocol implementors prefer. And the "name" in the current registry
for many encodings is unacceptable (e.g., Japanese EUC).
How can we go about either changing the preferred name, or having more
than one preferred name (preferred for different purposes).
i'm looking in draft-ietf-822ext-*.
Larry
Ned Freed
> I can't find any place that lists the preferred names that the HTTP
> protocol implementors prefer. And the "name" in the current registry
> for many encodings is unacceptable (e.g., Japanese EUC).
Uh, well, of course not -- the registration procedure is just that, a
procedure, not a registration in and of itself. The preferred names will have
to be defined once the procedure is adopted and the registry changed to match
the new procedure. Note that this isn't the only thing that has to be redone --
all registered character sets will have to indicate whether or not they are
valid as MIME text. I will make sure that the character set registered by the
base MIME specification indice a preferred name as well as all the aliases
for them.
Note also that the name you mention is unacceptable according to the new
procedure as well -- preferred names have to be in a form that can be used as a
MIME charset parameter without quotes or any other nonsense.
> How can we go about either changing the preferred name, or having more
> than one preferred name (preferred for different purposes).
> i'm looking in draft-ietf-822ext-*.
The new registration procedure is called draft-freed-charset-reg-00.txt; as
a result of the character set workshop it has been teased out of MIME and
now stands on its own.
Ned