[lkml]   [1996]   [Dec]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
Patch in this message
SubjectRe: GB vs. MB (End this thread!)

The problem here is STANDARDS and CONVENTIONS.
Even though many of us may think that the defacto industry convention
(1GB = 1000 MB) is wrong, and should be "1GB = 1024 MB", we are pissing into
the wind.
Let's try to stay with the mainstream.

A wise man told me once, "Pick your battles. If you persist on spending too
much effort on little things that aren't really that important, you won't
have the strength/credibility for the things that ARE important."

My $.02.
From: on Wed, Dec 4, 1996 7:50
Subject: z-Re: GB vs. MB (End this thread!)

two comments on this thread from me who originated it (much to my surprise;)
to get it finally stopped again (hopefully):

> Please people, end this thread! Those who want to make changes to the way
> the kernel does things are invited to make patches. When the kernel
> behaves exactly the way you think it should, send the patch to
> linux-kernel, Linus, and/or the maintainer of the driver you wish changed,
> with a full explanation of why you believe your patch does the
> RightThing(tm). If the driver maintainer or Linus also believes that your
> patch does the right thing, it will be included in the kernel.
> If you want to change stuff, don't talk about it. Send a patch.
> End of discussion. End of thread. Followups without patches > /dev/null

don't know if you saved the original mail for this thread which I've written.

this mail was exactly a one-line-patch to fix a problem (but not really that
which the now discussed; the original problme was 1000*2^20 vs. 1024*1^20)

with a two sentences of explanation.

> PLEASE : every time I got a message in this thread, I
> thought "will they stop ?", and I'm starting to post
> too... so the question is now : what do the concerned
> guy think about this, considered what has been "discussed"
> these last days (yes, the one who maintain this code)

I've to say hat I'm very suprised (and a bit disapointed) about the
in this thread.

I've now saved 50 mails for this thread (maybe I've missed a few) and
absolutely noone really commented on my patch or the real problem which
was *not* at all powers of 2 versus powers of 10 (should I send the original
mail/patch again? probably no good idea to get this discussion down;)

when I realized after a few followups that there is some misunderstanding
about the real topic of my patch I've send a 2nd short note the next day
trying to make things clearer but this didn't help at all.

I guess everyone just read something about 2^20 vs. 10^6 and thought:
ha, I've to give my $0.02 too (which sums up to exactly $1 right now ;-)

I think this thread was (since now it's stopped;) a horrible example
about the signal/noise ratio in this list in general :-(


PS & ob-linux: oops, this mail got longer than I thougt first, so here is
patch again (please read and think about this and every mail before
replying and commenting!) :

--- /soft/linux/drivers/scsi/sd.c Wed Apr 17 10:51:59 1996
+++ linux/drivers/scsi/sd.c Thu Apr 18 10:16:45 1996
@@ -1170,7 +1170,7 @@
mb = rscsi_disks[i].capacity / 1024 * hard_sector / 1024;
/* sz = div(m/100, 10); this seems to not be in the libr */
- m = (mb + 50) / 100;
+ m = (mb*1000 + 51200) / 1024 / 100 ;
sz_quot = m / 10;
sz_rem = m - (10 * sz_quot);
printk ("SCSI device sd%c: hdwr sector= %d bytes."

All SCSI disks will from now on ___ _____
be required to send an email notice 0--,| /OOOOOOO\
24 hours prior to complete hardware failure! <_/ / /OOOOOOOOOOO\
Harald Koenig, \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik // / \\ \ ^^^^^ ^^^^^

 \ /
  Last update: 2005-03-22 13:38    [W:0.020 / U:2.624 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site