This message is from the T13 list server.


Don,

You did not read my message carefully.  I said that it was obsolete in
ATA-4, approved in August 1997, but first finds its way into the working
draft in June of 1996 with revision 5.  It was submitted as a proposal in
February of 1996 (D96125), which was revised 5 times before being included
in the working draft.  The issue of obsolescing commands was discussed in
the minutes of March 27-29, 1996, and indeed a vote was taken at that time
on also including the SEEK command (which lost).  All of these documents are
on the web site.

The way standards bodies work is that proposals are submitted (and available
on the web site) on a specific topic, and then are debated.  Almost all
proposals go through several iterations as a standalone document (some have
taken up to 10 revisions) before the committee accepts it.  This one went
through 5 revisions, not uncommon.

At that point the editor of the draft document will schedule its inclusion
in the next revision of the working draft. Once that is produced, it becomes
the baseline that future proposals reference.  At this stage a question may
come up from the editor or people reviewing the working draft that was not
apparent in the original proposal (for instance, a subtle interaction with
another section of the document).  These are then resolved, sometimes by the
editor taking direction from the committee, sometimes by a new proposal (if
the changes are significant). Note that it is still possible to submit a
proposal reversing the original proposal, although obviously doing that
earlier would have had more of a chance of success (typically new
information is needed to justify a reversal).

Finally, a draft is produced that is reviewed on a line by line basis.  For
ATA-4 this happened in March of 1997 with revision 10.  The committee spends
several meetings (often special extra meetings) reviewing the draft document
before "forwarding" it to the parent committee.  So these reviews produced
more drafts, up to draft revision 16 in August of 1997.  Forwarding produces
letter ballot comments and the public review stage, which generated revision
17 in October of 1997.  Finally, revision 18 incorporated editorial changes
people later discovered in August of 1998.

And then the process begins again.  The committee has actually been revising
the standard on a 12-18 month of so cycle.  Too long a cycle and things get
approved as proposal, but not communicated as a standard (we had a bad
experience once with SCSI-2 that took 5 years to approve).  To short a
cycle, and people get confused and not enough change is presented to really
warrant a new standard.

This is why I take exception to some concept that the committee acts on a
"whim."  It takes so long to get something in the standard, from concept to
publishing, that frequently the "new" topics covered in ATA-(N) were
originally presented to the committee while it was still working on
ATA-(N-1)!  Their are several layers of review, all of the meetings are
public, all paperwork is posted on the web site.  In this particular case
the topic was begun in February of 1996 and concluded (with forwarding
ATA-4) the following summer - 15 months.  It took 18 months if you include
the comment resolution phase.  Hardly a "whim."

Jim

PS I was actually on the committee at the time, although I was transitioning
my more active duties to others.  The people on T13, particularly the
officers and editors, work quite long and hard and produce something of
substantial value to the industry.


-----Original Message-----
From: don clay [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, March 20, 2002 7:59 AM
To: ata reflector
Subject: [t13] Re^19 R/W (Long) Commands


This message is from the T13 list server.


On the experience issue, I have been intimately involved in product planning

both with manufacturing and with customers as well and am just as well
versed 
with those costs and implications as well.  Once again, the R/W Long command

in all of these considerations as well as all others is insignificant
(that's as in 
zero, not measurable, nada, doesn't count, any way that you want to say
it!!!) 
when compared to the whole of an ATA Interface.  

When you obsolete a command that is currently being used by those in the ATA

Community, you put them at risk to the whims of the committee.  

Since you say that R/W Long was obsoleted in August 1997, I went back to 
June 1997 and re-read all of the reflector email from 6/97 through the end
of 
10/97.  There was no mention of any discussion of R/W Long in either the 
reflector email, the meeting minutes, nor any of the ad hoc meetings, 

3/19/02 6:08:05 PM, "McGrath, Jim" <[EMAIL PROTECTED]> wrote:

>
>Don,
>
>The ATA committee regularly exercises due diligence - all proposals are
>available on the T13 web site for public viewing long before they are
>approved by the committee.  All meeting minutes are also posted on the web
>site.  This is usual practice for most standards bodies, which are required
>by law to be public bodies (since otherwise standard setting in private by
>companies could be considered an antitrust violation).
>
>On the experience issue, the problem is that I think you are looking at it
>from an engineer's perspective (my experience has been in program 
management
>and marketing as well as engineering).  The cost of getting products to
>market is far higher than just the direct engineering cost.  And once
again,
>the risks of making changes when there is no benefit is too high for our
>business.  By making these commands obsolete the standard gives 
comapanies
>the freedom to make those tradeoff themselves.
>
>Jim
>
>
>
>



Reply via email to