Errata for PNG spec (tRNS chunk)

[Glenn Randers-Pehrson]

Chris,

In clause 11.3.2.1 (tRNS chunk), the table entry for colour type 2 has
the green and blue samples in the wrong order.

It should read:

     Colour type 2

     Red sample value      2 bytes
     Green sample value  2 bytes
     Blue sample value     2 bytes

Glenn

------------------------------------------------------------------------------
_______________________________________________
png-mng-misc mailing list
png-mng-misc@...
https://lists.sourceforge.net/lists/listinfo/png-mng-misc

[Chris Lilley]

On Thursday, June 25, 2009, 1:20:51 AM, Glenn wrote:

GRP> Chris,

GRP> In clause 11.3.2.1 (tRNS chunk), the table entry for colour type 2 has
GRP> the green and blue samples in the wrong order.

GRP> It should read:

GRP>      Colour type 2

GRP>      Red sample value      2 bytes
GRP>      Green sample value  2 bytes
GRP>      Blue sample value     2 bytes

Glenn,

Thanks for the report.

The W3C process for errata management is described at
http://www.w3.org/2005/10/Process-20051014/tr.html#rec-modify

In accordance with that document I have updated the errata document
(this is the first reported erratum).
http://www.w3.org/2003/11/REC-PNG-20031110-errata.html

Since the original PNG specification had the correct order, and since
the order Red, Blue, Green is used nowhere else in the specification,
it looks as if implementations have silently corrected for the obvious
typo. Thus, this erratum does not alter conformance and I have
classified it accordingly.

Is anyone aware of any implementation which followed the letter of PNG
second edition here (and would thus become nonconformant)?

Please review the proposed erratum, even though the error is simple
and obvious.

I wonder if a simple test case, say a PNG image with two colors, one
of which is transparent if the erratum (or PNG first edition) is
followed and the other of which is transparent if PNG second edition
had been followed, would be a useful addition?

Since this specification was published as both a W3C Recommendation
and as an ISO Standard, I have copied David Duce since the proposed
correction (despite being trivial) needs to be reported to ISO. I
assume that they will issue a Technical Corrigendum, but welcome
guidance from David here.

--
 Chris Lilley                    mailto:chris@...
 Technical Director, Interaction Domain
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG


------------------------------------------------------------------------------
_______________________________________________
png-mng-misc mailing list
png-mng-misc@...
https://lists.sourceforge.net/lists/listinfo/png-mng-misc

[David Duce]

Chris, Glenn,

I've had an email exchange with Jean Stride at BSI. Jean also runs some
of the key ISO Committees.

It turns out that the ISO Committees are meeting in London in 10 days
time. ISO has a process for Technical Corrigenda, which involve ballots.
Jean is looking into whether this item could be considered at the London
meetings. If it could it is likely to save considerable effort. But we
will have to move very quickly if they agree to consider it.

Jean has contacted the chair of the meeting, Dick Puk, to see if he can
help. You may recall that Dick participated in some of our meetings.

If I get a positive response, has enough process been completed on the
W3C side for me to effectively say to Dick that W3C has approved this
erratum through its processes and hence we now invite ISO/IEC to do
likewise?

I've looked back through some old files. This error must have creapt in
very early on. Certainly the PNG 1.1 and 1.2 drafts I have use the RGB
order. The error is in the first Committee Draft sent to ISO/IEC and the
corresponding BSI Draft for Public Comment dated 98-12-01. But I have at
least one document which doesn't contain the error - that's an internal
document where we'd started the conversion to ISO/IEC format.

Best regards

David


Chris Lilley wrote:

> On Thursday, June 25, 2009, 1:20:51 AM, Glenn wrote:
>
> GRP> Chris,
>
> GRP> In clause 11.3.2.1 (tRNS chunk), the table entry for colour type 2 has
> GRP> the green and blue samples in the wrong order.
>
> GRP> It should read:
>
> GRP>      Colour type 2
>
> GRP>      Red sample value      2 bytes
> GRP>      Green sample value  2 bytes
> GRP>      Blue sample value     2 bytes
>
> Glenn,
>
> Thanks for the report.
>
> The W3C process for errata management is described at
> http://www.w3.org/2005/10/Process-20051014/tr.html#rec-modify
>
> In accordance with that document I have updated the errata document
> (this is the first reported erratum).
> http://www.w3.org/2003/11/REC-PNG-20031110-errata.html
>
> Since the original PNG specification had the correct order, and since
> the order Red, Blue, Green is used nowhere else in the specification,
> it looks as if implementations have silently corrected for the obvious
> typo. Thus, this erratum does not alter conformance and I have
> classified it accordingly.
>
> Is anyone aware of any implementation which followed the letter of PNG
> second edition here (and would thus become nonconformant)?
>
> Please review the proposed erratum, even though the error is simple
> and obvious.
>
> I wonder if a simple test case, say a PNG image with two colors, one
> of which is transparent if the erratum (or PNG first edition) is
> followed and the other of which is transparent if PNG second edition
> had been followed, would be a useful addition?
>
> Since this specification was published as both a W3C Recommendation
> and as an ISO Standard, I have copied David Duce since the proposed
> correction (despite being trivial) needs to be reported to ISO. I
> assume that they will issue a Technical Corrigendum, but welcome
> guidance from David here.
>

--
Professor David Duce

School of Technology
Oxford Brookes University
Wheatley Campus
OXFORD
OX33 1HX
UK

email: daduce@...
tel: [phone]
fax: [phone]

------------------------------------------------------------------------------
_______________________________________________
png-mng-misc mailing list
png-mng-misc@...
https://lists.sourceforge.net/lists/listinfo/png-mng-misc

[glennrp-2]

> has enough process been completed on the W3C side for me
> to effectively say to Dick that W3C has approved this erratum
> through its processes and hence we now invite ISO/IEC to do
> likewise?

I've looked back through some old files. This error must have creapt in
very early on. Certainly the PNG 1.1 and 1.2 drafts I have use the RGB
order.


    I believe the PNG-1.2 draft establishes our position, so I don't
    think a formal vote is necessary. Our voting process takes a
    minimum of 28 days.  I recall that in the past we have used
    a Call for Unanimous Consent or Call for Objection, which took
    14 days.  But in this instance I don't think even that is required.


    Glenn

------------------------------------------------------------------------------
_______________________________________________
png-mng-misc mailing list
png-mng-misc@...
https://lists.sourceforge.net/lists/listinfo/png-mng-misc