I can share the files. Just let me know where to send them.
BTW, I was able to narrow down the problem.
If I create the initial XMP sidecar file with IMatch, PM will not read it.
I was going to simply blame IMatch, but before doing that, I tested with Photoshop/Bridge CS2 and the Adobe apps can read the XMP file without problems. CS3 as well. I also used the RDF Validator at w3.org and it passes, so I'm not convinced the problem is IMatch.
Anyway, I have the XMP that causes the problem. Let me know where to send it.
iMatch is not encoding its text properly.
For instance, notice the copyright symbol:
<rdf:li xml:lang='x-default'>© Some guy</rdf:li>
It is not properly UTF-8 encoded. It should be like this:
<rdf:li xml:lang='x-default'>© Some guy</rdf:li>
You'll note that both the CS2 file and the PM file have the proper encoding. In UTF-8 any characters that have a value higher than 127 have to be encoded to at least two bytes, and up to 6 bytes may be used to encode some of the less common unicode characters.
The iMatch application is not properly converting characters to UTF-8, and it is likely that they are not converting their text at all, so any characters, like accents, the euro symbol, etc. will cause encoding issues.
The reason it works in Photoshop is likely because they have decided to relax their version of the XMP parser a bit. We are very strict when interpreting XMP files.
Thanks for the samples.
-Kirk
