[t13] op x12 "INQUIRY" "vendor id" how distinct

Pat LaVarre p.lavarre at ieee.org
Sun Jun 13 12:50:45 PDT 2004


This message is from the T13 list server.


Jeff G:

> Double-quote chars are used in email messages (and C code) to delimit 
> strings.

And elsewhere, yes.

> Submitting the string "ATA     " verbatim would be quite silly.

Ouch, I see I forgot to quote the most significant (and FIFTH) field of
the form, specifically:

Requested Vendor ID: ATA

I meant to submit eight chars with blanks only on the right, not blanks
anywhere else, no quotes, and not ten chars.  I'm confident that the
human reader of the form will understand that.

> > t10 Vendor ID's will not contain x22 double-quote chars, and
> > 
> > "While the Vendor ID list includes both upper and lowercase characters, 
> > requests for new Vendor IDs that are the same  as an existing ID except 
> > for the character case will not be accepted."
> 
> Er, huh?

Sorry, thanks for asking, yes we've got threads tangled here, my fault
more than anyone's.  I now believe I should have chosen to post twice,
not once.  I will try to disentangle:

Relevant algorithms I've never before or elsewhere seen clearly
specified include:

1) How to quote a vendor ID.

That piece of t10.org says we can quote a vendor ID with double quotes
and never collide with any standard content.

2) How to say if a vendor ID is or is not equal to a suggested string.

My own background in human factors tells me I should strip leading and
trailing blanks.  I should compress any other run of blanks to a single
blank.  I should not distinguish blank/ nul/ tab.  I should ignore case.

What makes me question my background are the dozens of people who in
fact massively distribute less friendly choices.

What confuses this issue is that t10.org, unlike the later usb.org
effort for example, has no separate merely numeric vendor ID.  We have
only the human readable ID which has to serve both as the human readable
form and also as the machine readable form.

What's news to me is that a piece of t10.org divides all possible ID's
into three categories, not two:

a) left-justified byte-for-byte matches
b) loose matches that agree over the printable bytes
c) mismatches

Category (b), while arguably distinct from category (a), still prevents
other people from taking the arguably distinct ID.

Category (b) is a match like (a) and also a mismatch like (c).

Pat LaVarre





More information about the T10 mailing list