2005.04.19 18:55 "Re: [Tiff] libtiff ,is any Unicode version available?", by Andrey Kiselev

Hi Dirk,

This problem was first concerned by Dirk.Haun. and today i downloaded the newest verison of libtiff ,which is marked as 3.7.2. and it still doesn't support unicode build.

my concern is: although, tiff file format is born without contain "unicode text". all information (string ,text),it contain as a ANSI text.

I was thinking some idea as follows:

In Unicode version:(probably)
text can be transformed to unicode when it was retrieved .(by read)
text can be transformed to ANSI when it was saved by client application which is unicode build.(by write)

(this implies that all the text to be read or saved in tiff file is ANSI compatible. that is, such as english(Roman/Latin based language )). but text in other language such as CJK,Korea Language,Chinese...will not allowed.

Does anyone do some research on this area? and i since more and more softwares are word-wide oriented, why libtiff still doesn't contain unicode build? I think it was a such important demand.

I do not think that libtiff should have any support for Unicode unless the TIFF specification will define the multibyte character types. For now all strings should be stored in the 7-bit ASCII encoding. Of course, you can use 8-bit Unicode encoding here (UTF-8), but that will be application dependent and incompatible with other apps. And, of course, you do not need any special support from libtiff to store the UTF-8 data instead of plain ASCII. All decoding/encoding work should be done by the top level application. Recent Adobe specifications (DGN) contains UTF-8 along with ASCII as an encoding for new string fields. Also there is new data type 14 reserved by Adobe to handle Unicode data, but there is no description for that new type and I do not know what encoding these new strings should use.

I'm very interested in adding Unicode types to TIFF specification. It is very disappointing that I can't add comments to TIFF file in my native language. On the first look simple replacing ASCII strings with the UTF-8 ones shouldn't have any side effects, but I may be wrong.

Best regards,
Andrey

Andrey V. Kiselev
Home phone: +7 812 5970603 ICQ# 26871517