Sorry for not commenting sooner. We have been discussing the printer scheme
further and I think we need to further elaborate the pages-per-minute
attribute. Specifically, it is almost universal that color printers
advertise two speeds: the optimal monochrome speed and the optimal full
color speed (though sometimes vendors cheat on the full color speed by
advertising the CMY but not K speed, which is typically faster but of lower
quality on four pass laser engines). Anyway, these two speeds should suffice
for service advertising. I don't think duplex speeds (or other non-optimal
speeds) are needed because they are almost universally not advertised. So, I
suggest the page-per-minute attribute be an ordered list of integers, where
the first integer is the monochrome speed and the second integer is the full
color speed. The second value should be optional (i.e. not present or set to
zero for monochrome-only printers). And speaking of ordered lists, this
brings me to my second comment.
According to draft-ietf-srvloc-service-scheme-12.txt, section 126.96.36.199 on
"The values for multivalued attributes are an unordered set.
Deletions of individual values from a multivalued attribute are not
supported, and deletion of the attribute causes the entire set of
values to be removed."
Yet, some of the attributes defined in the draft printer template are
specified as ordered lists (e.g. uri-supported and uri-security-supported).
Is this legal? If not, then I guess we need two attributes for speed:
page-per-minute-monochrome and pages-per-minute-color.