[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#163528: ITP: randomize-lines -- Randomize lines of input text



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mon, 7 Oct 2002, Junichi Uekawa wrote:

> I'm just pointing out to you that there exists an application that does
> what you describe, as Debian package. That's the duty of filing an ITP.

I think that Debian is big enough for two small tools that have
overlapping functionality (the're nothing wrong with more choice) so I
asked my sponsor to upload the new release (which has even more features
as described below).

> > Some differences (from browsing the manual page and toying with it):
> >
> > * rl can pick lines from the input (you can say give me 10 random
> >   lines from this file), this can be done with and without recurring
> >   lines (my math/statistics knowledge is mostly Dutch but it's like
> >   drawing cards with and without replacing them to the deck).
>
> head/tail, sort, uniq can do them.

The operation is subtly different in that lines from the input may be
picked multiple times. uniq can only pick out lines that have the same
content but does not distinguish between different lines.

Also when using the --count option operation is much more efficient
(memory usage) than randomizing the whole file and picking the first
couple of lines.

> > * I don't understart bogosort's default mode of operation
> >   (it just permutates until the file is sorted? at least it takes more
> >   time than I am willing to wait for it), with --nosort it behaves
> >   just like rl by default
>
> It works like "sort" with a less optimal algorithm.

This is something that randomize-lines does not do. (unless you just call
rl until you think the file is sorted)

> > * bogosort has some command-lines features like --output and --verbose

randomize-lines now also has a --output option.

> > * rl seems to be about three times as fast (just running
> >   'time rl 100000lines > /dev/null' and
> >   'time bogosort -n 100000lines > /dev/null a couple of times, no real
> >   statistics involved here)
>
> It could be due to the way bogosort generates randomness.

- -- arthur - arthur@tiefighter.et.tudelft.nl - http://tiefighter.et.tudelft.nl/~arthur --
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)

iD8DBQE9vCVjVYan35+NCKcRAhzAAJ0R1t4tYGeDv3uDEG1TlEQxRaT25wCeLleu
4vrzFMi61CZ7kDn0vCIJ0Sw=
=pD9u
-----END PGP SIGNATURE-----




Reply to: