2004.06.09 14:41 "[Tiff] TIFF validation", by Fernando Loygorri

2004.06.09 16:15 "Re: [Tiff] TIFF validation", by Joris Van Damme

does this mean that the note on page 16 has no bearing on the validity of a TIFF file and it just exhorts writers of viewing software to try to do the best they can when faced with a malformed TIFF file?

I guess so.

I think it's justified to generalize that statement, too. Writers should be built to do the exact right thing. Readers should be built to cope with as many wrong things as possible, without breaking in any way other then emitting a warning message, if even remotely possible.

You don't even need to spec to specify that. Consider a file that starts with a header signature that indicates little endian, followed by a 42 word that indicates big endian. A simple reader will consider second value completely incorrect and give up. A more sofisticated reader may consider second value a 'common mistake' that is 'tolerated', and move on, going from the assumption that first value is correct. A very sofisticated reader might actually spawn two further decoding attempts, trying out both endians, and going with the one that is (most) successfull. It's not up to the spec to tell us what to do in such a case, the spec only tells us how the values should have been, and it's up to us to build the most robust and sofisticated and tolerant possible reader.

Joris Van Damme
Download your free TIFF tag viewer for windows here: