AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

1994.02.10 19:32 "LibTIFF working on MS-Windows 3.1.", by Soren Pingel Dalsgaard
1994.02.11 09:39 "Re: LibTIFF working on MS-Windows 3.1.", by Bjorn P. Brox
1994.02.11 01:09 "Re: LibTIFF working on MS-Windows 3.1.", by Richard Minner
1994.02.11 14:40 "Re: LibTIFF working on MS-Windows 3.1.", by Jie Yao
1994.02.11 01:16 "Re: LibTIFF working on MS-Windows 3.1.", by Richard Minner
1994.02.11 10:17 "Re: LIBTIFFW working on MS-Windows 3.1.", by Niles Ritter
1994.02.14 11:03 "Error/warning handlers in libtiff", by Erik Corry
1994.02.16 22:38 "Re: Error/warning handlers in libtiff", by Sam Leffler
1994.02.11 18:27 "Re: LibTIFF working on MS-Windows 3.1.", by Francois Gauthier
1994.02.11 21:29 "Re: LibTIFF working on MS-Windows 3.1.", by Sam Leffler

1994.02.11 21:29 "Re: LibTIFF working on MS-Windows 3.1.", by Sam Leffler

Good News! (for some, anyway :-)

I have been toying with LibTIFF for a long time,

  • Ability to write both II and MM type tiff images. Yes that's right you can write little and big endian images using the library.

I think this feature may prove useful when an application seeks to append images to an existing TIFF file on a machine that has different byte ordering than the original. This is not normally possible since the byte ordering is file wide.

Providing this feature in libtiff would in my view increase it's machine independence. It could make the append mode of TIFFOpen() more robust.

This is the only reason that I'd be willing to do byte-swapping when writing a file. As was mentioned previously, the TIFF spec has required readers to be able to handle both byte orders for MANY YEARS. As best as I can determine the majority of the implementations that do not support both big- and little-endian byte orders are on PC's. A good version of my library for these machines would undoubtedly go a long way toward eliminating this problem; much in the way that the free availability of the software has helped insure interoperability on other platforms.

The bottom line is that I am adamantly opposed to adding support to the library that encourages the continued use of bogus TIFF implementations. My preferred solution is to have available software that can be freely incorporated into products so that these incompatibilities can be avoided.

BTW, support for byte-swapped TIFF is *extremely* simple.

        Sam