Re: Question re: RFC 8011 Section 7.2 Attribute Extensibility
"Kennedy, Smith \(Wireless & IPP Standards\) via ipp" <[email protected]>
| Newsgroups | gmane.ietf.ipp |
|---|---|
| Message-ID | <[email protected]> |
Thanks, Mike! That explanation makes sense to me. Cheers, Smith --- Smith Kennedy [email protected] > On Feb 10, 2024, at 6:09 PM, Michael Sweet <[email protected]> wrote: > > CAUTION: External Email > > Smith, > > The disadvantage of doing vendor prefixes on all attributes, member attributes, and values is that it makes it harder to convert to a standard implementation. > > So for a new attribute, prefix the attribute, e.g.: > > smiNNN-foo (type2 keyword) > smiNNN-foo-default (type2 keyword) > smiNNN-foo-supported (1setOf type2 keyword) > > while the values are not prefixed: > > none > bar > baz > > For a new member attribute, the member attribute and the supported attribute are prefixed: > > media-col (collection) > smiNNN-media-scent (type2 keyword) > > smiNNN-media-scent-supported (1setOf type2 keyword) > > For a new collection attribute, prefix the top-level attributes but not the member attributes: > > smiNNN-foo-col (collection) > foo-id (integer) > bar (type2 keyword) > > smiNNN-bar-supported (1setOf type2 keyword) > smiNNN-foo-col-default (collection) > smiNNN-foo-col-supported (1setOf keyword) > smiNNN-foo-id-supported (rangeOfInteger) > > >> On Feb 10, 2024, at 3:00 PM, Kennedy, Smith (Wireless & IPP Standards) <[email protected]> wrote: >> >> >> Hi Mike and Chris, >> >> Since this is a convention and not something that is normatively defined or validated, I suppose the vendor can do what it wants. >> >> But what about the "xxx-supported" / "xxx-default" attributes that correspond to the member attributes in the collection? >> >> If at HP I define: >> >> smi11-whatever-col (collection) >> whatever-number (integer) >> whatever-type (type2 keyword) >> >> so that I can make: >> >> smi11-whatever-col = { >> whatever-number = 42 >> whatever-type = whale >> } >> >> there could likely be these accompanying attribute definitions: >> >> whatever-number-supported (rangeOfInteger) >> whatever-number-default (integer) >> whatever-type-supported (1setOf type2 keyword) >> whatever-type-default (type2 keyword) >> >> Is that how we want this to be? Or would it be better like so: >> >> smi11-whatever-col (collection) >> smi11-whatever-number (integer) >> smi11-whatever-type (type2 keyword) >> >> and then >> >> smi11-whatever-number-supported (rangeOfInteger) >> smi11-whatever-number-default (integer) >> smi11-whatever-type-supported (1setOf type2 keyword) >> smi11-whatever-type-default (type2 keyword) > > ________________________ > Michael Sweet > _______________________________________________ ipp mailing list [email protected] https://www.pwg.org/mailman/listinfo/ipp