Re: 8.1 system info / admin functions - Mailing list pgsql-hackers

From Andreas Pflug
Subject Re: 8.1 system info / admin functions
Date
Msg-id 4328385D.9060803@pse-consulting.de
Whole thread Raw
In response to Re: 8.1 system info / admin functions  (Stephen Frost <sfrost@snowman.net>)
List pgsql-hackers
Stephen Frost wrote:
> * Tom Lane (tgl@sss.pgh.pa.us) wrote:
> 
>>Neil Conway <neilc@samurai.com> writes:
>>
>>>While we're on the subject, the units used by pg_size_pretty() are 
>>>incorrect, at least according to the IEC: for example, "MB" is 
>>>strictly-speaking one million bytes, not 1024^2 bytes. 1024^2 bytes is 1 
>>>MiB (similarly for KiB, GiB, and TiB). I'll take a look at fixing this 
>>>as well, barring any objections.
>>
>>[ itch... ] The IEC may think they get to define what's correct, but
>>I don't think that squares with common usage.  The only people who
>>think MB is measured in decimal are disk-manufacturer marketroids.
> 
> 
> That isn't entirely accurate, unfortunately.  Telecom also generally
> uses decimal.  I'm as unhappy with the current situation as anyone but
> it's not just disk makers & marketing folks, even amoung the computing
> industry.

Telcos don't tend to follow standard computer industry thinking...
Kilo, Mega and so on are used as decimal exponent multiples to SI units, 
like meter, gram, .... There's no such thing as a byte as physical unit, 
it's completely artificial.

pg_size_pretty follows the standard of "df -h" or "du -h" and so on, 
which use the well-known 1024-multiples, as every sane user of bytes will.

Regards,
Andreas


pgsql-hackers by date:

Previous
From: "Merlin Moncure"
Date:
Subject: parameterized fetch
Next
From: Josh Berkus
Date:
Subject: Re: About method of PostgreSQL's Optimizer