Hi Tom,
Personally, I prefer having a single octet for the attribute. In all of the
other cases of (1setOf text(MAX)) the values have been discrete values and the
order of the values was not important. If we were to split the conneg string
across multiple octet strings, then order would be VERY important and we no
longer have one octet string representing a complete value.
Gail
"Hastings, Tom N" <hastings at cp10.es.xerox.com> on 03/06/2001 11:02:37 AM
To: Paul Moore/AUCO/US at AUCO
cc: ifx at pwg.org, ipp at pwg.org (bcc: Gail Songer/AUCO/US)
Subject: RE: IFX> New data types [for IPP WG agenda]
Paul and IPP FAXers,
I wasn't at the Maui meeting (and haven't seen any minutes).
And I'm sorry not to be at the meeting today.
However, I question the idea of adding new long data types. I thought it
made much more sense to make the attributes be (1setOf text(MAX)) and
(1setOf octetString(MAX)). Then you could have an arbitrary number of 1023
octet strings. There would be no 32K or 64K limit. We'd never have to add
new data types. AND MOST IMPORTANT, IPP FAX would be compatible with
current IPP (which was one of our main goals). Then current Printer
implementations could support the added features of IPP FAX, but would not
have to add fundamental mechanisms, such as new data types.
If you think it is a problem to break up CONNEG strings at arbitrary
boundaries, and alternative would be to have an IPP FAX Printer attribute
that is the URL to the conneg text file. Advantages:
a. The conneg file could also be any length.
b. The conneg file could be on another server.
c. A client implementation that also supports Internet FAX would be dealing
with the same CONNEG text strings as a unit that come back from the server.
As Paul suggests, this should be also discussed at the IPP WG meeting,
tommorrow, March 7.
Tom
-----Original Message-----
From: pmoore at netreon.com [mailto:pmoore at netreon.com]
Sent: Monday, January 29, 2001 09:42
To: ipp at pwg.org
Cc: ifx at pwg.org
Subject: IFX> New data types
Fellow IPPers,
In the IPP Fax meeting in Maui (1/26/01) we decided that we needed two new
data
types to deal with large lumps of data.
BigText. This is just like Text except that it is up to 65535 bytes long
BigOctetString. This is just like OctetString except it is up to 65535 bytes
long.
BigText is used for vcard representation of users. [Note to ifx attendees -
this
does need to be text since it contains localized human readable data - my
mistake].
BigOctetString is used for CONNEG data. We are alos considering using it for
carrying digital certificates.
We considered several alternatives but this seemed the cleanest solution.
Can I therefore ask that this be a topic for your next meeting so that you
can
bless this change (or not of course!)
THnaks
Paul Moore
IPP Fax Chair