2006.08.24 17:45 "Re: [Tiff] IPTC tag", by Phil Harvey

2006.08.25 18:30 "RE: [Tiff] IPTC tag", by Chris Cox

Unfortunately, the TIFF readers that have the bug are in widespread use in the newspaper and magazine industries.

I haven't had a lot of luck getting them to fix even major bugs -- so we constantly try to work around them.

And, sadly, TIFF errors are far too common -- most TIFF readers have a long list of exceptions and corrections (Photoshop's has a LONG list).

As always, it pays to be lenient in what you read, and strict in what you write. But sometimes you have to write it in a way that other people/applications expect rather than the most correct form.


-----Original Message-----

From: tiff-bounces@lists.maptools.org on behalf of Phil Harvey

The one thing we should be able to agree on is that regardless of the data type, IPTC data should NOT be byte-swapped when written. If writing with the current version of libTIFF, this means that should be written as undefined. I tried this strategy (writing IPTC as undefined) with my software and found users who complained of incompatibilities, so I changed to writing long as Adobe does. For me it was an easy change, although I agree that it serves to perpetuate a historical 'bug' in the way IPTC is written. I also agree that it would be best if this bug could be eventually squashed, but it looks like here I am part of the problem, not the solution.

Along the same lines, I also consider it a bug that IFD offsets are written as long (value of 4). They should be written as IFD type (value of 13), otherwise unknown IFD's can not be parsed. If we're trying to fix historical bugs, maybe we could fix this as well. While you may never have been stung by this bug, I have repeatedly because of the non-standard TIFF-format information that my software has to deal with.