3 Some of the planned features include:
5 - Remove some of the old porting cruft. since we no longer support
6 FIXME-2011: in mid to late 2011, remove the following directories:
7 amiga, atari, msdos, nt, os2, vms
9 - Separate out the shell scripts like gzexe into a new little package;
10 these scripts are less used and less reliable and should be optional.
12 - Internationalize by using gettext and setlocale.
14 - Structure the sources so that the compression and decompression code
15 form a library usable by any program, and write both gzip and zip on
16 top of this library. This would ideally be a reentrant (thread safe)
17 library, but this would degrade performance. In the meantime, you can
18 look at the sample program zread.c.
20 The library should have one mode in which compressed data is sent
21 as soon as input is available, instead of waiting for complete
22 blocks. This can be useful for sending compressed data to/from interactive
25 - Make it convenient to define alternative user interfaces (in
26 particular for windowing environments).
28 - Support in-memory compression for arbitrarily large amounts of data
29 (zip currently supports in-memory compression only for a single buffer.)
31 - Map files in memory when possible, this is generally much faster
32 than read/write. (zip currently maps entire files at once, this
33 should be done in chunks to reduce memory usage.)
35 - Add a super-fast compression method, suitable for implementing
36 file systems with transparent compression. One problem is that the
37 best candidate (lzrw1) is patented twice (Waterworth 4,701,745
38 and Gibson & Graybill 5,049,881). The lzrw series of algorithms
39 are available by ftp in ftp.adelaide.edu.au:/pub/compression/lzrw*.
41 - Add a super-tight (but slow) compression method, suitable for long
42 term archives. One problem is that the best versions of arithmetic
43 coding are patented (4,286,256 4,295,125 4,463,342 4,467,317
44 4,633,490 4,652,856 4,891,643 4,905,297 4,935,882 4,973,961
47 Note: I will introduce new compression methods only if they are
48 significantly better in either speed or compression ratio than the
49 existing method(s). So the total number of different methods should
50 reasonably not exceed 3. (The current 9 compression levels are just
51 tuning parameters for a single method, deflation.)
53 - Add optional error correction. One problem is that the current version
54 of ecc cannot recover from inserted or missing bytes. It would be
55 nice to recover from the most common error (transfer of a binary
58 - Add a block size (-b) option to improve error recovery in case of
59 failure of a complete sector. Each block could be extracted
60 independently, but this reduces the compression ratio.
62 For one possible approach to this, please see:
64 http://www.samba.org/netfilter/diary/gzip.rsync.patch
66 - Use a larger window size to deal with some large redundant files that
67 'compress' currently handles better than gzip.
69 - Implement the -e (encrypt) option.
71 Send comments to <bug-gzip@gnu.org>.
74 ========================================================================
76 Copyright (C) 1999, 2001, 2006, 2009-2016 Free Software Foundation, Inc.
77 Copyright (C) 1992, 1993 Jean-loup Gailly
79 Permission is granted to copy, distribute and/or modify this document
80 under the terms of the GNU Free Documentation License, Version 1.3 or
81 any later version published by the Free Software Foundation; with no
82 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
83 Texts. A copy of the license is included in the ``GNU Free
84 Documentation License'' file as part of this distribution.