Ira,
Maybe I'm a bit confused. I'm not nearly as active as you are in the
standards community as you are and definitely less knowledgeable with
XML than I should be. However, with the other standards that I have
worked, no matter where in the process, I seem to remember us working on
SOME version of the protocol. I would give status reports on IPP v1.1
and I never followed anything but SNMPv3. While the associate group was
working on the specs, they gave the FILE some long complicated name, and
this name usually appeared in the file so that if you just had a printed
version, you would still be able to verify that everyone was on the same
page (so to speak).
For me, I was always dealing with some file that was to trying to define
some version of a protocol. In the case of PSI, we are working on
version 1.0, for IPPFax we are working on version 1.1. To get to the
formal version, we will have numerous versions of the descriptive Word
document, but we are still working on protocol version 1.0 or 1.1.
Once the spec gets to its final, official, never to change again, state,
it seemed to get some "official" ID, say RFC 2911 or perhaps
5100.1-2001.
I realize that with PSI, we publish "temporary" directories so that
people can try to use definitions and find problems with the protocol
early. Right now these directories have some number where that number
also appears in the name of the protocol word document. Does this
really have to be "version"? Could it not just as easily be a date
(2003.01.31)?
I understand that there is a desire to make it easy to correlate
"current working version" of the protocol with some version of the
specification. In these cases, I would prefer that the name of the
protocol document contain the DATE AND "some other numeric identifier"
were that other identifier is used to create the appropriate name space.
I would include both because I do believe that is possible and likely to
change the specification without changing the fundamentals of the
protocol itself. (For example, when I go final and have to change then
name of the specification, I will need to change the name of the file.
However, this WOULD NOT require protocol rev. Nothing in the protocol
changed, only the name). Including both would allow us to store and
easily identify the desired file.
Gail
-----Original Message-----
From: McDonald, Ira [mailto:imcdonald at sharplabs.com]
Sent: Friday, January 31, 2003 3:04 PM
To: Gail Songer; McDonald, Ira; Harry Lewis
Cc: PWG (E-mail)
Subject: RE: PWG> "Draft Standard" is an oxymoron
Hi Gail,
If you want to introduce two different versions (the spec versus
the over-the-wire protocol), we can do that.
But at the risk of more misunderstanding from management and
customers. There is simply no precedent at all for that
distinction in the public standards of any other standards
organization that I've ever heard of.
For PSI, definitely the XML MUST change if the spec changes.
The XML is commented and has the definitive descriptive
clause for each attribute and each value. If the spec
changes enough to need a new version (and not just an
errata file in the PWG standards directory), then the
XML will have to change too.
Cheers,
- Ira McDonald
High North Inc
-----Original Message-----
From: Gail Songer [mailto:gsonger at peerless.com]
Sent: Friday, January 31, 2003 4:50 PM
To: McDonald, Ira; Harry Lewis
Cc: PWG (E-mail)
Subject: RE: PWG> "Draft Standard" is an oxymoron
Ira,
I was talking about the filename of written documents are stored on the
server. I think we should differentiate between the version of the
protocol and current file name of the specification.
I understand that XML files need some sort of unique name. This should
be the same as the version of the protocol but that might not be the
same thing as the filename that for the associated version of a spec.
(For example, if the spec underwent purely editorial changes that caused
no change in the XML, then the spec should be able to change without
having to rev the XML)
I feel like there are two different topics; how to name the specs and
how to version the protocol.
Gail
-----Original Message-----
From: McDonald, Ira [mailto:imcdonald at sharplabs.com]
Sent: Friday, January 31, 2003 1:59 PM
To: 'Harry Lewis'; Gail Songer
Cc: PWG (E-mail)
Subject: RE: PWG> "Draft Standard" is an oxymoron
Hi,
I agree with Harry - the major.minor.revision three-tier numeric
versioning is clearly needed for the schema files and the other
machine-readable artifacts that are required for PSI operation.
Date is not an acceptable form for the over-the-wire detailed
versioning of a protocol - and the two are the same thing for
PSI.
Cheers,
- Ira McDonald
PS - I'm happy to have date in the filename for working-drafts
and/or PWG 'standards track' documents.
-----Original Message-----
From: Harry Lewis [mailto:harryl at us.ibm.com]
Sent: Friday, January 31, 2003 3:18 PM
To: Gail Songer
Cc: PWG (E-mail)
Subject: RE: PWG> "Draft Standard" is an oxymoron
A point that (I think) Ira and Tom were trying to make is that "from the
outside looking in" major (and minor) version numbers such as v1.0 and
v1.1
have "meaning". Since we are so close to the standards development
process... any rolling counter will do for us (date or
major.minor.revision).
Also we need to have a format that makes sense for PSI and (potentially)
automatic reference and retrieval of schema, XML documents etc.
I guess I'm casting my preference toward v1.x.x. with clear-cut rules as
to
where to start and end the sequence as we somewhat articulated on the
call
yesterday.
----------------------------------------------
Harry Lewis
IBM Printing Systems
----------------------------------------------
"Gail Songer" <gsonger at peerless.com>
Sent by: owner-pwg at pwg.org
01/31/2003 01:22 PM
To: "PWG (E-mail)" <pwg at pwg.org>
cc:
Subject: RE: PWG> "Draft Standard" is an oxymoron
I prefer identifying documents by date rather than by version number.
For IPPFax, it was sometime into the process when we decided what
"version"
to use. (Should we make this the first version 1.0 or align our version
with the version of IPP we were going to require V1.0) There also seems
to
be some discussion on the version of IPP for the Document object....
Gail
-----Original Message-----
From: Harry Lewis [mailto:harryl at us.ibm.com]
Sent: Friday, January 31, 2003 12:03 PM
To: Hastings, Tom N
Cc: Farrell, Lee; PWG (E-mail)
Subject: RE: PWG> "Draft Standard" is an oxymoron
I propose we draft a working draft of a proposal for drafting draft
standards proposals. Oh.. that's right.. we did that once...
Seriously... can we move off the topic of how the brain and tongue work
together and focus on what appeared to be the issues with substance from
yesterday's call?
1. 3 tier or 2 tier. We had a 3 step process but I'm willing to reduce
this
to 2 steps based on our experience
- We used to call our 3 step process Proposed, Draft and Standard
- We can call our 2 step process anything but I think Proposed and
Standard
were the most vocal (Draft and Standard does fit better in my brain...
but
then there is this endless debate.. anyone got a coin)?
2. Versioning
- <major>.<minor>.<revision>
- date coded
We have documents in or nearing last call which really depend on
closure.
----------------------------------------------
Harry Lewis
IBM Printing Systems
----------------------------------------------
"Hastings, Tom N" <hastings at cp10.es.xerox.com>
Sent by: owner-pwg at pwg.org
01/31/2003 12:35 PM
To: "Farrell, Lee" <Lee.Farrell at cda.canon.com>
cc: "PWG (E-mail)" <pwg at pwg.org>
Subject: RE: PWG> "Draft Standard" is an oxymoron
Lee and Bill,
The problem is what do you call successive versions of the Draft
Standard,
before you are ready to send it out for Last Call?
Working Drafts of the Draft Standard?
Using "Draft" in two different senses in the same sentence to identify a
document is pretty confusing. And we know that people in normal
conversion
like to drop the adjectives and just talk about the "Draft". So which
do
they mean when they say the "Draft is ...".
Tom
-----Original Message-----
From: Farrell, Lee [mailto:Lee.Farrell at cda.canon.com]
Sent: Thursday, January 30, 2003 18:33
To: PWG (E-mail)
Subject: RE: PWG> "Draft Standard" is an oxymoron
Duh.
[If people can understand "jumbo shrimp" without losing sleep, I don't
see
why "draft standard" would cause a problem.]
lee
-----Original Message-----
From: Harry Lewis [mailto:harryl at us.ibm.com]
Sent: Thursday, January 30, 2003 6:05 PM
To: Hastings, Tom N
Cc: pwg at pwg.org
Subject: Re: PWG> "Draft Standard" is an oxymoron
Er... Um... so why is it so hard to put the definition to use and
realize
that a "Draft Standard" is a preliminary version of a "Standard"?
----------------------------------------------
Harry Lewis
IBM Printing Systems
----------------------------------------------
"Hastings, Tom N" <hastings at cp10.es.xerox.com>
Sent by: owner-pwg at pwg.org
01/30/2003 04:24 PM
To: pwg at pwg.org
cc:
Subject: PWG> "Draft Standard" is an oxymoron
Here is why I think that "Draft Standard" is an oxymoron. Draft is too
fleeting. Standard is meant to be more stable.
So I looked up the word "Draft" in the dictionary. Webster's Seventh
Collegiate Dictionary says:
"a preliminary sketch, outline, or version".
We all use the word "draft" (or "working draft") to mean the document
that
we update rapidly to get to a version that we all consider stable enough
to
have a Last Call.
So one of the appealing suggestions made at today's call was to just
remove
section 3.4 Draft Standard and have only 3.4 Proposed Standard and 3.6
Standard. Both have to have a series of drafts to be reviewed to lead
up to
being an approved Proposed Standard or an approved Standard. And both
need
to have a draft that is considered good enough to both trying a Last
Call
and then the Last Call has to actually pass.
I think much of our trouble is terminology, so fixing the terminology,
and
deleting a step seems to be a good thing to do and is NOT abandoning the
process or overturning turnips.
Tom