I should have pointed out that the name field will now be required for a
custom size name, as well as standardized names, whereas in Draft D0.9 the
name field was optional in custom names.
There are two advantages to requiring a name field in a custom size name:
1. The syntax for custom and standard names becomes the same, simplifying
implementation (since both will require a name).
2. It is good practice for an administrator when defining a custom size to
invent some sort of a name to suggest the size's use to its users.
About your second question, I see I wasn't clear about whether or not the
ABNF would appear in the standard or only in the PWG FTP site. I think that
the first ABNF (current with version 1.0 of the standard) should be in the
standard, including the commenting, so that implementations can see what to
expect in case they choose to fetch the current (updated) ABNF from the PWG
When we post the final approved standard, we should also post the ABNF flat
file which will be identical to what is in the standard. Only when we
register a new size name that needs a new class name, will we need to update
the ABNF flat file on the PWG server. Periodically, but not more frequently
than, say, once a year, we will update the standard with new size names and
any new class names that have been registered in the meantime.
Another reason to include the ABNF in the standard is that fetching the
current ABNF from the web site isn't required (for an implementation or an
From: Bergman, Ron [mailto:Ron.Bergman@Hitachi-hkis.com]
Sent: Monday, July 09, 2001 11:23
To: 'Hastings, Tom N'; ipp (E-mail)
Subject: RE: IPP> MED - Proposed final ABNF for Media Size Self
It should be pointed out that this does change the current syntax for custom
size names, in that the "size-name" will be mandatory. This part is
optional in the current draft.
Also, I would like a clarification. Does this proposal completely remove
the size ABNF from the document? Or, is the ABNF included with a note
indicating where to get the latest? Although it may be somewhat confusing,
it would be cleaner to only include the reference to the ABNF.
From: Hastings, Tom N [mailto:email@example.com]
Sent: Monday, July 09, 2001 10:57 AM
To: ipp (E-mail)
Subject: IPP> MED - Proposed final ABNF for Media Size Self Describing
Ron and I want to publish a final version of the Media Standardized Name
standard for Last Call. There are three standards that require its use, so
we need to complete the Media standard. One is UPnP Basic Printing
Template. Another is IPP FAX.
Please send any comments on this proposed ABNF BEFORE we publish the next
draft. Silence will be interpreted as agreeing with the ABNF.
Please send any comments by Tuesday, July 17, 2001.
There has been the suggestion on the mailing list that the ABNF should be a
single production, including the custom part and the future registered class
name part as is common practice.
We are also agreeing to retain the explicit list of class names in the ABNF.
When a new size name is registered which needs a new class name, then the
ABNF will be updated to include the new class name. The updated ABNF will
be available on the PWG site as a flat text file in a published and stable
Advantages of this approach:
1. All of the ABNF for the Media Size Self Describing Name is specified as a
single ABNF production, so that code can be written that will be stable,
even as new media size names are registered with new class names. This is
the common practice for other standards that use ABNF.
2. Which class names go with which units is part of the ABNF, so that it is
clear that using the wrong set of units with a class name violates the ABNF
and permits a parser to detect such mal-formed ABNF.
3. Using a single "custom" class name (rather than one for in and one for
mm) simplifies the parsing for custom names because the parser needs to look
for one special prefix ("custom"), as with Media Type and Media Color,
rather than for, say, "custom-in" and "custom-mm" class names. Since
administrators are defining the custom sizes, not vendors, whether they
choose in or mm units will depend on their users.
Disadvantages of this approach:
1. When new class names are registered, the existing parsers won't be able
to validate that the units are correct, unless they access the PWG web site
for the current ABNF. (The standard will NOT require implementations to
access the PWG web site to get the current ABNF, nor does it require that
implementations detect corrupt names that use the wrong units).
The ABNF flat file will have comments at the beginning to identify the date
and version of the standard and the date and version of the ABNF. The ABNF
date and minor version number will be updated every time it is updated. The
ABNF follows RFC 2234.
Here is the proposed final ABNF as it will appear in the flat file on the
PWG site in:
; ABNF for the Media Size Self-Describing Name
; Part of the IEEE/ISTO 5101.1 Media Standardized Names standard
; For use with the IEEE/ISTO 5101.1 Standard Version 1.0, July 9, 2001
; ABNF version 1.0, July 9, 2001
; This ABNF is updated whenever a new media size class name is registered
; according to the procedures of IEEE/ISTO 5101.1.
( class-in "_" size-name "_" short-dim "x" long-dim "in" ) |
( class-mm "_" size-name "_" short-dim "x" long-dim "mm" )
class-in = "na" | "asme" | "oe" | "custom" | reg-class-in-name
class-mm = "iso" | "jis" | "jpn" | "prc" | "roc" | "om" | "custom" |
reg-class-in-name = ( lowalpha | digit ) *( lowalpha | digit | "." )
reg-class-mm-name = ( lowalpha | digit ) *( lowalpha | digit | "." )
size-name = ( lowalpha | digit ) *( lowalpha | digit | "-" )
short-dim = dim
long-dim = dim
dim = integer-part [fraction-part] | "0" fraction-part
integer-part = non-zero-digit *digit
fraction-part = "." *digit non-zero-digit
lowalpha = "a" | "b" | "c" | "d" | "e" | "f" | "g" | "h" | "i" |
"j" | "k" | "l" | "m" | "n" | "o" | "p" | "q" | "r" |
"s" | "t" | "u" | "v" | "w" | "x" | "y" | "z"
non-zero-digit = "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9"
digit = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9"
This archive was generated by hypermail 2b29 : Mon Jul 09 2001 - 17:47:45 EDT