[BBLISA] Compression Algorithms

[David Allan]

On Tue, 29 Jun 2010, Daniel Feenberg wrote:

>
>
> On Tue, 29 Jun 2010, Edward Ned Harvey wrote:
>
>>
>> Ah yes.  They're using the same compression algorithm as 7zip, but they're
>> in-line-filterable, and behave just like a normal "gzip" command etc.
>>
>> AKA, there is never any reason to use bzip2.  Using lzma --fast is much
>> faster, and better compression than bzip2.  It's superior in every way.
>
> I think that if you are sending files to another person, it is much more
> important that he/she have the decompression software than that the file
> be small. At least that is what I feel when I receive a file. It is part
> of "be conservative in what you send", etc.

I agree with this logic, and I'd extend it a bit to say use the algorithm 
that makes sense for the task.  If I'm just compressing something trivial 
to mail to someone, I use gzip--it's fast, everybody pretty much has it 
now, and the compression ratio is acceptable.  If I'm archiving something 
huge that I really don't intend to decompress more than once or twice, 
then I'm going to use the maximum ratio I can find, with the caveat that I 
only use FOSS for this stuff partly because I'm a FOSS guy, but mostly 
because I don't want to run the risk of having the tool unavailable when I 
want the data back.  Or maybe it's data that I'd need to get back in a 
hurry because a need to recover would only happen during system downtime 
so decompression time is relevant...lots of parameters...

Good to have the data, that's very helpful in making these decisions.

Dave