Re: Character set registration

On 12/17/95 at 3:31 PM, Ned Freed wrote:
>> As I remember, many
>> of them were registered en masse a long time ago, and most of them are
>> useless for (i.e. unrecognized by any) MIME application.
>
>The usefulness or uselessness of a given character set is NOT a criteria for
>registration. For one thing, it clearly isn't IANA's job to figure out what
>character sets are useful to a given application. IANA lacks both the delusion
>of sufficient expertise and the foolhardiness necessary to attempt such a
>thing.

I agree. On the other hand, there is a registration procedure defined in that draft which does involve community review.

>Allow me to provide you with a small example of the problems inherent in this
>area. You assert that most of them aren't supported by any MIME application.
>This is, in fact, provably false, because I have a counterexample -- *my* MIME
>application supports almost all of them! In fact I know of at least one
>other application that provides support comparable to what I provide in mine.

I retract the statement, especially because if I have my way, yet another implementation may exist which can support all of these. However, I will say that whether or not a particular implementation supports many of the character sets, that is no indication of the usefulness in producing messages which them. See below.

>Registration is a way of making a given definition known, describing one key
>characteristic we all need to know, and giving that definition one or more
>names. That's it. Any exceptation that this process is going to say (or has
>ever said) more than this is totally specious.
>
>The particular issue of whether or not a given character set is appropriate for
>a given application is entirely out of scope...

Then perhaps I need you to explain to me what's going on in section 5.1.4:

>5.1.4.  Usage and Implementation Requirements
>
>Use of a large number of character sets hampers
>interoperability.  However, the use of a large number of
>undocumented and/or unlabelled character sets hampers
>interoperability even more.
>
>A character set should therefore be registered ONLY if it adds
>significant functionality that is valuable to a large
>community, OR if it documents existing practice in a large
>community. Note that character sets registered for the second
>reason should be explicitly marked as being of limited or
>specialized use.

This sounds to me very much like a requirement for "appropriateness" in some sense of the word. So, if we as a community determine that a particular character set is not valuable to a large community and there is not a widespread use of the character set in e-mail, objections to that character set may lodged during the review period. My take then from section 5.2.2 is that the character set reviewer may reject the registration due to "significant objections raised on the list or objections raised externally."

Is this not the intended use of the review process?

>...and when it's necessary to do
>this it should be dealt with either by getting the IAB to come up with their
>long-awaited guidelines for character set usage or by writing an applicability
>statement for a given class of applications and getting it approved.

Absolutely. Personally, I'm not altogether comfortable with the review process being totally "guidelineless", since I can imagine instances where people will try to torpedo character sets they consider "politically incorrect" (I'll leave that to your imagination). Getting IAB to pick up the ball would be good.

>I wish you the best of luck on either of these enterprises, should you
>attempt to follow up on getting them done...

I will keep back from the flame for the time being, though I have been known to bite off more than I can chew in the past.

pr
--
Pete Resnick <mailto:presnick(_at_)qualcomm(_dot_)com>
QUALCOMM Incorporated
Home: (217)337-1905 / Fax: (217)337-1980