[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Lzip-bug] Maybe replace "bits/byte" with "ratio"?
From: |
Antonio Diaz Diaz |
Subject: |
Re: [Lzip-bug] Maybe replace "bits/byte" with "ratio"? |
Date: |
Mon, 11 Sep 2017 17:10:54 +0200 |
User-agent: |
Mozilla/5.0 (X11; U; Linux i586; en-US; rv:1.9.1.19) Gecko/20110420 SeaMonkey/2.0.14 |
Hi Timothy. Thanks for your comments.
Timothy Beryl Grahek wrote:
I personally haven't found "bits/byte" useful at all; I actually, even
now, don't understand what "bits/byte" actually means.
You can find a definition in the excellent book "Data Compression
Explained"[1], by Matt Mahoney:
"Compression ratio is often measured by the size of the compressed
output file, or in bits per character (bpc) meaning compressed bits per
uncompressed byte."
[1] http://mattmahoney.net/dc/dce.html#Section_2
"bits/byte" is more useful in measuring the compression of text.
but I can definitely say that the % saved part has been the most
useful and #:# moderately useful. Showing the inverse compression ratio
would be quite favorable to me, and would make Lzip output more tasteful
to me as a whole. This, of course, is my humble opinion. :)
It is also my opinion, and as nobody has objected, I think I'll make
lzip show the inverse compression ratio.
vv52_new.log.lz: 15.329:1, 6.52% ratio, 93.48% saved. ok
Having both the saved and inverse values expressed as percentages seems
like the better of the two options when considering decimal versus
percentage.
Agreed. Two things I like of showing the inverse compression ratio as a
percentage are that:
1) It gives more precision because the integer part is almost always 0.
2) It makes it easier to see that ratio + saved = 1 = 100%
Thank you for asking for user opinion on this matter.
I find it the most natural to ask users about the user interface. :-)
Best regards,
Antonio.