AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2002.03.22 14:28 "TIFF image tiles from the UK Ordnance Survey Errors out.", by Roger Bedell
2002.03.22 16:05 "Re: TIFF image tiles from the UK Ordnance Survey Errors out.", by Phillip Crews
2002.03.22 17:12 "RE: TIFF image tiles from the UK Ordnance Survey Errors out.", by Roger Bedell
2002.03.22 17:25 "RE: TIFF image tiles from the UK Ordnance Survey Errors out.", by Phillip Crews
2002.03.22 17:46 "RE: TIFF image tiles from the UK Ordnance Survey Errors out.", by Roger Bedell
2002.03.22 16:46 "RE: TIFF image tiles from the UK Ordnance Survey Errors out.", by Laurence R Moore
2002.03.22 17:18 "RE: TIFF image tiles from the UK Ordnance Survey Errors out.", by Roger Bedell
2002.03.22 17:07 "Re: TIFF image tiles from the UK Ordnance Survey Errors out.", by Jason Summers
2002.03.22 22:31 "Re: TIFF image tiles from the UK Ordnance Survey Errors out.", by Chris Cox
2002.03.25 20:13 "UK Ordnance Survey Maps were fixed in 1999", by Roger Bedell

2002.03.22 16:46 "RE: TIFF image tiles from the UK Ordnance Survey Errors out.", by Laurence R Moore

We have quite a bit of experience with scanned maps in TIFF format at the USGS. I looked at the IFD of your file using od (Unix hex dump utility), and it looks like the error messages you are getting are correct diagnostics of an illegal dataset. The RowsPerStrip tag really does have a count of zero, and the DateTime tag has a count of 21. Both are clearly violations of the TIFF standard (page 31, DateTime N=20, page 39, RowsPerStrip N=1)).

I think it unlikely that the file was accidentally corrupted; the overall structure of the IFD is coherent and the content looks mostly legal. My recommendation is to contact the data producer. Before doing that, I suggest you confirm with your customer that the Ordinance Survey is the real source of the data; we have a lot of trouble with complaints from people who believe they have USGS data when in fact they have a picture of a USGS map that was scanned by someone else. On the other hand, we also produce bad data now and then, and we like to know about it when someone discovers the error.

----------
Larry Moore
U.S. Geological Survey
Mid-Continent Mapping Center
lmoore@usgs.gov


Editor's note: This mail was not originally archived, and has been reconstructed from quotes.