Re: Comments on the HTTP/1.0 draft.

Thus wrote: "Roy T. Fielding"
>Marc VanHeyningen writes:
>> Kindly show me where the MIME RFC requires conforming implementations to
>> check the registry; I can't seem to find it in my copy.
>A BNF defines how a conforming application should *parse* the input.
>RFC 1521 specifies that the parser should declare a syntax error if the
>MIME type is not registered with IANA.

I think I see what you're saying now; I see it as a normal distinction
between spec (which is tight) and practice (where an implementation is
welcome, even encouraged, to accept a larger set of potential inputs
than the grammar specifies if such can be done cheaply and without
introducing ambiguity or other such problems.)

But, if you really think it belongs in the grammar, OK.  I just
instinctively shy away from creating new grammars unless it's
absolutely necessary; goodness knows there are enough of them in the

>It already mentions these issues and includes explicit reference to how
>media types are registered.

Well, the issue I take is with the statement that HTTP should be
"allowed more freedom in the use of non-registered types."  I believe
one way to interpret this is "use whatever non-registered types you
want, as long as it's HTTP and not email that's no problem," and I'm
not sure that is what was meant.

- Marc
Marc VanHeyningen  <URL:>

Received on Friday, 9 December 1994 18:51:30 UTC