Closed markus1189 closed 9 years ago
I'm trying to think of a specific name for it, as the mebibyte format isn't really used anywhere. In the stack
case, it doesn't matter much because you're telling people how many mebibytes there are, so they don't have to be familiar with it.
An alternative could be to add two versions, bytesDecimal and bytesBinary, inspired by https://en.m.wikipedia.org/wiki/Binary_prefix On Jul 13, 2015 10:56 AM, "Chris Done" notifications@github.com wrote:
I'm trying to think of a specific name for it, as the mebibyte format isn't really used anywhere. In the stack case, it doesn't matter much because you're telling people how many mebibytes there are, so they don't have to be familiar with it.
— Reply to this email directly or view it on GitHub https://github.com/chrisdone/formatting/pull/17#issuecomment-120856399.
Yeah, bytesDecimal
works for me.
Applied as bytesDecimal
and bumped as 6.2.1. Thanks!
Uhm, hang on there, bytesDecimal
would be the version that uses 1000
as the base not 1024
as did the version I used, KB = 1000 bytes, KiB = 1024 bytes, so the version you currently named bytesDecimal
is wrong. What about #18 instead?
Okay, this is confusing. ls
and nautilus and whatnot report a file with 1024 bytes in it as "1K" and "1.0 kB (1,024 bytes)" respectively. This is the typical notion of a kilobyte that I'm used to since forever. I've never seen anywhere use 1K or 1KB to mean 1000 bytes. Your use of KiB — a notation I've never seen in the wild, lead me to believe it was somehow different, referring to 1000 bytes. After looking again I see it just means what we're used to (KB=1024, MB=1,048,576 bytes, etc.), but with an additional i
in there. I'm not sure how I feel about this.
I'm just going to put this in as bytes
without the i
inside like everybody else and call it good.
Yeah it is a confusing. AFAIK most consider KB = 1024 but the kilo prefix in other units is supposed to mean 1000. So per standard kilobytes = 1000 bytes not 1024, KiB (kibibytes) explicitly specifies that it uses 1024, while with KB everybody kind of does what he wants. Now that you're primed you will be surprised how often you can actually see KiB etc. instead of KB, just keep looking.
However whatever works for you works for me too :), so I closed the other pr
@markus1189 @chrisdone I'm kind of disappointed that the most illogical decision was made here. We have a standard of prefixes, and I would have expected that this standard was followed. A lot of Linux tools have standard output of KiB instead of KB. 1 KiB = 1024 Bytes 1 KB = 1000 Bytes
@BrunoChevalier
We have a standard of prefixes, and I would have expected that this standard was followed.
I think "standard" is a matter of use, not prescription.
A lot of Linux tools have standard output of KiB instead of KB.
At the time of writing my previous message, Nautilus displayed as "1KB". OS X displays a 4*1024*1024
file as 4 KB
:
Chriss-MacBook-Pro:~ chris$ stack exec ghci
GHCi, version 8.0.1: http://www.haskell.org/ghc/ :? for help
Loaded GHCi configuration from /Users/chris/.ghci
Prelude> writeFile "x.txt" (replicate (4*1024) 'a')
If you want to present to users a familiar size then I think the normal (aka standard) notation is more legible than something unfamiliar with an unfamiliar meaning.
I'm kind of disappointed that the most illogical decision was made here.
Calling things you don't agree with illogical is not a good way to get what you want.
Extracted from commercialhaskell/stack#571
I wasn't sure under which category I should place it, in the end I went with
Bases
, though maybeUnits
would be better, I felt like opening a whole section was overboard, but I leave that to you ;)