This is a mirror of some of the files from Stuart Inglis's web page. In particular, the TIC (tic98) image compressor written as part of his PhD thesis. This compresses scanned documents very well, better than any other compressor (as of March 1999). The copies of the files here will be needed only if the main page disappears (again).
Because of limited disk space I have recompressed the files with bzip2, and only the A4-size version of the PhD thesis is kept. Also I have not kept the Windows binaries (but I could probably compile some myself if they're needed and the main site is down).
When you compress a file with tic98 you have to remember the same options to correctly decompress it. The wrappers entic98 and untic98 give a handy way to compress and decompress PBM files with a reasonable set of compression options. They need the netpbm tools installed.
If you want to keep scanned documents compressed and serve them up as PDFs you might want tic98topdf. It needs the untic98 wrapper mentioned above and pip.
To serve the PDFs through a web interface you might like the topdf CGI script. It takes a single CGI argument 'file' and serves it to the browser as a PDF. It understands tic98, PostScript, bzip2'd PostScript, PDF and bzip2'd PDF and does the right conversion depending on the filename. Of course, decompressing and converting to PDF every time a document is requested can be expensive. I've often found, however, that disk space is scarcer than CPU time.
While tic98 itself is GPL'd, you can consider these shell and Perl scripts to be in the public domain.