2013/4/3 Simon Marchi <span dir="ltr"><<a href="mailto:simon.marchi@polymtl.ca" target="_blank">simon.marchi@polymtl.ca</a>></span><br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div>> Normally I'd be the first to advocate k, M, G as powers of 10 and Ki, Mi, Gi as powers of two (ISO/IEC 80000-13:2008) but the kernel parameters use exclusively K, M, and G in the powers of two sense, and I believe just about every other part of Linux does also. We should thus, somewhat unfortunately, stick to powers of two. Just be sure to explicitly give the multipliers' values in the help. Oh, and accept just the upper-case letters, like the kernel parameters do.<br>
<br>
</div>Yeah.. we need to accept capital K for kilo (kibi) because everyone<br>
else does. Actually I was thinking of mimicking the behaviour of dd.<br>
For kilo (kibi), you can use upper and lower case (probably reeks of<br>
when there was no such things as lower case letters in computing. I<br>
look at you FORTRAN). But for mega (mibi) and giga (gibi), it's upper<br>
case only.<br>
<br>
The more I think about it, the more I think it doesn't matter if we<br>
accept upper and lower case for all... I don't know anybody who will<br>
request a buffer of 32 millibytes.</blockquote><div><br></div><div>I guess we could follow the "dd" way of work :</div><div><span style="font-size:13px;font-family:arial,sans-serif">BLOCKS and BYTES may be followed by the following multiplicative suffixes: c =1, w =2, b =512, kB =1000, K =1024, MB =1000*1000, M =1024*1024, xM =M GB =1000*1000*1000, G =1024*1024*1024, and so on for T, P, E, Z, Y.</span> </div>
</div>