I agree with Pat 100% about the SNMP having just implemented Mk.1 on a
AMD186 for a photocopier/printer regime (Mk.II SNMP being FAR worse). This
is a *monster* and is being bent and warped to do things that are well
beyond the original scope it was invented for. The front end requirements
is not much better, and it does not address security either. HTML browsers
are all either already supporting or about to support IMAP in place of
POP3, and so we have a ready made and robust protocol at our feet ready to
use. But the matter does not end there.
We must be sure that we don't settle for SNMP just because it represents
the path of least resistance and in-fighting amongst manufacturers - the
lowest common denominator like RS232 vs. RS422.
I would respectfully invite eveyone to very closely look at the IMAP 4
(Interactive Mail Access Protocol) RFC 1730 and related protocols (such as
Myers IMSP draft) that covers authentication and scaling problems where
multiple servers are involved. It has a proven functionality and is well
suited to the matter of printing IMHO - especially where a user is VERY
likely to never use the same workstation twice, in which case it answers
most of the problems of synchronicity, queues and error conditions.
This should be looked at immediately as the basis for the protocol - it has
public stacks already available, is easy to implement, is stingy on
bandwidth and promises to soon overtake POP3 because of its enhanced
functionality and inherent nature of explicitly dealing with mobile users
who appear and disappear from the network - the core of what the networked
world is all about. This is the reason many browsers already have it or
soon will have it, and so it is a natural fit and one that does not tread
on any proprietory toes.
At 6:56 PM 3/3/97, Patrick Powell wrote:
>The problem with assuming that you will sent HTML as printer status
>is that it is highly unstructured. If all you want to do is have
>ummm... wetware... read and process the status then you don't need
>to worry. People are very adaptible. However, if you want to use
>the status information for managment tools then you will need something
>>And before you talk about SNMP as the answer, note that SNMP has a HORRIBLE
>overhead in terms of development, bulk, etc., that simple text based systems
>appear not to have.
>>I like text, simple formats, and the ability to parse it with a minimum
>of effort. I would even reject LALR grammars - simple keyword based
>tag/formats would be nice.
>>Patrick ("Browser De Jour Syndrome") Powell