At 09:26 09/11/97 PDT, Jay Martin wrote:
>Considering we have just declared a "Functionality Freeze", is this
>something we really need to do?
I think it is more of a "plugging a hole" than new fuctionality
and its only a Printer description attribute. See below to see if
you agree.
>>Can you cite some examples/scenarios in which *not* having this
>new attribute would make life very difficult (or impossible) to
>implement?
As I understand the spec and agreements, we've agreed that validation
of a document-uri cannot be part of a Validate operation, since the
Validate operation is identical to a Print-Job operation which cannot
have a document-uri attribute. So there is no way for a client to know
what scheme to try for a Print-URI or Send-URI scheme, except to try it,
in the Print-URI or Send-URI operation.
Then we have also been debating whether a Printer has to validate the
document-uri or not on the Print-URI or Send-URI operation. We've agreed
that the standard will NOT require the Printer to validate the URI.
So the user may NOT even get an error back on the Print-URI or the Send-URI
operation that the scheme is not supported (though maybe we should
add such an error and at least recommend that a Printer at least validate
that the URI scheme is supported, if it doesn't validate the entire
URI.)
Finally, I suspect that it will be an end-user, rather than a program,
that will be supplying URIs in Print-URI and Send-URI operations. The
end-user knows about document xxx and wants to submit it by reference
to Printer yyy, so that the end-user should be able to find out which
schemes are supported.
We already have a notification-schemes-supported, so that
"document-uri-schemes-supported" is very similar.
On the other hand, those more familiar with HTTP servers can tell us
whether most HTTP servers support most URI schemes that an end-user might
need, such as HTTP:, FTP:, and HTTPS:
Alternatively, we could recommend a list of URI schemes for a Printer
to support. Or would that get out of date? And require a lot discussion
and debate?
So adding a "document-uri-schemes-supported" Printer attribute seemed
an easy solution. The complete proposed spec is given below.
Tom
>> ...jay
>>----------------------------------------------------------------------
>-- JK Martin | Email: jkm at underscore.com --
>-- Underscore, Inc. | Voice: (603) 889-7000 --
>-- 41C Sagamore Park Road | Fax: (603) 889-2699 --
>-- Hudson, NH 03051-4915 | Web: http://www.underscore.com --
>----------------------------------------------------------------------
>>>Tom Hastings wrote:
>>>> In the vein of "plugging holes":
>>>> ISSUE: How can a client determine which uri schemes the Printer supports
>> for use in the "document-uri" attribute in the Print-URI and Send-URI
>> operations?
>>>> How about adding a printer description attribute:
>>>> document-uri-schemes-supported (1setOf uriScheme)
>>>> This attribte specifies the URI schemes that the Printer supports for use in
>> the "document-uri" attribute in the Print-URI and Send-URI operations.
>> If the Printer does not support either of these operations, the
>> "document-uri-schemes-supported" attribute SHALL not be supported.
>>