IPP Mail Archive: Re: IPP> MIME-types and alignment

Re: IPP> MIME-types and alignment

Larry Masinter (masinter@parc.xerox.com)
Wed, 20 Aug 1997 04:39:54 PDT

I asked:

> I would like to inquire if there are clients of printer MIBs
> >that use the enum list of printer interpreters for anything other
> >than displaying to the user of a system monitoring device for
> >informational display

and you answered with a hypothetical ("may be") rather than an actual
("is, in the following product"):

> Printer discovery may be coupled with programmatic use of the
> Interpreter table to facilitate selection of the correct driver
in a printing system.

I understand clearly how it could be used, what I'm asking is whether
it *is* used. Isn't it a misuse of SNMP to use it for discovery. Are
there systems that actually use the Printer MIB, rather than some kind
of directory service, for device discovery?

> Also, there are disadvantages to MIME for memory constrained systems
> (typical devices or end-nodes). An integer representation is much
> more concise and alleviates storing translations.

I find this incredible: an end-node has enough memory to hold the
4 bytes for an integer enumeration, but not enough to hold
the string "application/postscript"? I don't know what you mean
by 'storing translations'. MIME types are protocol elements, and
not user strings.

> Granted, most PDL names are not likely to require translation
> but I'm not entirely certain for all of them (use "PAGES" as
> an example").

MIME types are not "PDL names". They are registered strings instead
of registered entities.

The issue is not the advantage of form, the issue is that we need
a single registry of media types, so that we don't have to coordinate
parallel registries or deal with the consequences of not doing so.

Larry