Since Vista doesn't have compatibility for .RAR files, what should I use?
I tried WinRAR but it didn't really work.
Uh, the question here was about decompression. And windows 7zip does that for a load of formats, and reasonably well.
>>13-16
bzip2 works better on zip files with no compression that it does on tar files. uncompressed zip + bzip2 is better than 7z.
>>18, that makes no sense. Uncompressed zip and tar are both just collections of files. If this is true, it could only be because you're talking about a zip/tar of lots of tiny files where the file info takes up a substantial percentage of the space.
For a directory with about 30 files in it (all scripts, in fact), I compressed three files. The tar is a tar, the format is never compressed. The zip and 7z were created with compression turned off. Results:
-rw-r--r-- 1 me me 501760 2007-12-14 19:38 test.tar
-rw-r--r-- 1 me me 472411 2007-12-14 19:38 test.zip
-rw------- 1 me me 469089 2007-12-14 19:51 test.7z
Now I will bzip2 the tar and zip files, and then additionally create a zip and 7z file. For all commands I specified compression level 9 (best). Results:
-rw-r--r-- 1 me me 267195 2007-12-14 19:38 test.tar.bz2
-rw-r--r-- 1 me me 267668 2007-12-14 19:38 test.zip.bz2
-rw-r--r-- 1 me me 266118 2007-12-14 19:53 test.zip
-rw------- 1 me me 233809 2007-12-14 19:51 test.7z
Interesting things to note:
Incidentally I tried creating a solid archive as well (in this situation 7z groups the files together and then compresses the whole set in blocks) as I expected this to be smaller. But it turned out to be exactly the same size, so my input files mustn't be small enough to pay off for that optimisation.
>>21 Oh wow, the difference of 1,077 bytes between .zip and .tar.bz2 is HUGE :/
bzip2 is often worse or on par with deflate (gzip / zip) for very small files, and I guess that's what happened here.
All you guys have been talking about are plain text files.
Take binary files (even different kinds) into account and you can start your discussion from the beginning again.
Not really. The bottom line is that LZMA decompresses much faster than bzip2, compresses better, and is slightly slower than bzip2 to compress (or on par / faster in certain cases).
LZMA obsoletes bzip2, in my view.
Text or binary is irrelevant. What's relevant is entropy.
> > (1) Free -- download it at no cost.
> > (2) Open -- source code and format specifications available.
>
> the same can be said about zip and bzip2.
This is somewhat true. There are free implementations of the ZIP format, such as InfoZIP ( http://www.info-zip.org IIRC ), but I'm not sure what the patent situation is.
BZIP2 is a open-source royalty-free compression algorithm, so in that regard, yes, it has the same benefits as 7z.
> > (3) Decent -- 7z does pretty good compression, better than ZIP.
>
> uncompressed zip + bzip2 does better, and zip and bzip2 are far more common than 7z.
7z and ZIP are simply different ways of bundling files for compression by an algorithm such as BZIP2 or LZMA. As such, the container format shouldn't make a great deal of difference unless there's significant overheads. File ordering might play a part here too.
> my input files mustn't be small enough to pay off for that optimisation.
that's probably why the .zip and .tar.bz2 are smaller than the .zip.bz2 also.
>>26
zip normally uses the same deflate algorithm as gzip. it's not patented.
> BZIP2 is a open-source royalty-free compression algorithm, so in that regard, yes, it has the same benefits as 7z.
Yes and no. bzip2 can only do compression, 7z is a whole archive format, and not strictly a form of compression (the compression it uses is LZMA, I believe.)
>>23 fails to understand that 99% of what needs compressing is text in the first place. For graphics, video and audio we already have other ways to do it. Remaining files which are not graphics, video, audio nor text, are not the reason we are running out of disk space and thus do not need compressing.
On the other hand mail archives (text) get big. Log files (text) get big.