View Single Post
Old 02-11-2008, 03:12 PM   #221
igorsk
Wizard
igorsk ought to be getting tired of karma fortunes by now.igorsk ought to be getting tired of karma fortunes by now.igorsk ought to be getting tired of karma fortunes by now.igorsk ought to be getting tired of karma fortunes by now.igorsk ought to be getting tired of karma fortunes by now.igorsk ought to be getting tired of karma fortunes by now.igorsk ought to be getting tired of karma fortunes by now.igorsk ought to be getting tired of karma fortunes by now.igorsk ought to be getting tired of karma fortunes by now.igorsk ought to be getting tired of karma fortunes by now.igorsk ought to be getting tired of karma fortunes by now.
 
Posts: 3,442
Karma: 300001
Join Date: Sep 2006
Location: Belgium
Device: PRS-500/505/700, Kindle, Cybook Gen3, Words Gear
Quote:
Originally Posted by DMcCunney View Post
Correct. They've stated that they couldn't simply use something like the Zip "Deflate" algorithm because they wanted to be able to start decompressing at any specified point in the file (like the dictionary definition for a particular word), and Zip always started at the beginning.
What they did was just compress source HTML in chunks so that each chunk after compression fits inside a PDB record. They also added an index for mapping an uncompressed text position to record number so that only necessary records need to be decompressed. None of this actually requires Huffman, it can be done with deflate or lzw or lzma or whatnot.

Last edited by igorsk; 02-11-2008 at 03:16 PM.
igorsk is offline   Reply With Quote