I can see arguments for both, but commonly when people talk about magnitude, most people assume it's by the 10th powers. It might be necessary to specify if one wants to use the power of 2.
Very well, then. When a software geek is talking software and he says "we made that change and pages loaded several orders of magnitude faster", do you think 8x speed or 1000x speed?
The guy who said it was thinking 8x. I interpreted 8x without a hitch. A third listener said "wait, what?" And that prompted the question!
For technical folks, it may be clear, but I wouldn't want to let the marketers mislead people, most of whom would be right there with your third listener.
He needs to specify in order to not mislead, if the statement is without context.
This tells me that the third listener didn't have the context.
Which is not an indictment - At one point this morning I was forced to stop thinking in German to answer a question, and it took me a good fifteen seconds to formulate a comprehensible English sentence.
It does depend heavily on who you're talking to, then, because if I told my advisor boss that the device was getting a current three orders of magnitude higher, when I meant 8 times higher, there would have been much confusion and then mockery.
10 is just the default around physical sciences, I guess.
Not just that, but in the physical sciences, everything is in base 10. You measure Hz and KHz and MHz. You measure V and KV and uV. In the context of current, absolutely everything is in 10s.
Whereas everything is twos, in software. Even the 1000s are really 1024.
I suspect this would be unsatisfying. He'd stop in the middle to demand citations, complain that your penis is not presenting an NPOV, and argue with himself over the best technique mid-effort.
Software size; the common usage is base 10 as a rough approximation of the actual base 2 number. (1 gig is not exactly 1000kb, but is treated as such). In terms of response time, it depends on the extent. If you are measuring in flops, base 2. If you are measuring in fractions of a second, base 10.
I'm pretty sure it's only a "decimal" point in base 10. The baseless term is "radix point" I think; otherwise it would be a hexadecimal point or a ... binary point? I'm not sure on that last one.
Programmers who are trying to work with floating point numbers and need to maximize accuracy of operations need to understand binary fractions. If only to have it beaten into their heads why floating point arithmetic should always be viewed with suspicion.
I'm not sure you ever do actual math In The Real World, but when you're trying to optimize a system, or locate a source of inaccuracy, understanding how it works is vital.
Powers of 10, even though powers of 2 would seem more relevant. Calling powers of 2 "orders of magnitude" would let the marketing folks pull all kinds of shenanigans.
I'm hardwired to think powers of 10 if I hear order of magnitude, but I can see a good case for using powers of 2 for software. It would probably need to be specified, and I'd have to actually think about its meaning, though.
I would say base 10 off the top of my head, and I'm a programmer. ESPECIALLY when talking about time, which I refuse to do any kind of transformation on in non-decimal bases. Kiloseconds damnit, base 60 and base 24 are for people 6000+ years ago, not for modern civilized folks.
( there's a reason why I need Excel to do any time math more complicated than "Adding an hour" or similar. Base 60 makes my brain shut down in confusion )
Base 10 is where it's at in the Real World (tm), unless you're delving into things that are in the realm of assembly language.
In the world of Assembler (or assemblerish C/C++, like bit shifting), base 2, base 8 (octal) or base 16 (hexadecimal) might be appropriate, but that depends on the kind of processor you're dealing with.
Base 10 in most things (chemistry, physics), base 2 in computer speedy things. This was intuitive for me as soon as the N64 came out and I found out the previous systems were 8-bit, 16-bit, 32-bit... I was about 10, and I noticed the base 2 thing before I had any idea about logarithms.
So ya, I don't think, in the tech world it would be at all confusing to hear Orders of Magnitude referring to base 2. I might ask if you meant 2 or 10, but I wouldn't be all "whaaaaaaat?!" if you said 2.
(no subject)
Date: 2008-07-16 02:04 pm (UTC)(no subject)
Date: 2008-07-16 02:10 pm (UTC)(no subject)
Date: 2008-07-16 02:12 pm (UTC)(no subject)
Date: 2008-07-16 02:13 pm (UTC)(no subject)
Date: 2008-07-16 02:13 pm (UTC)(no subject)
Date: 2008-07-16 02:18 pm (UTC)(no subject)
Date: 2008-07-16 02:21 pm (UTC)(no subject)
Date: 2008-07-16 02:25 pm (UTC)I'm a base 10 fan myself.
(no subject)
Date: 2008-07-16 02:36 pm (UTC)base 60
(no subject)
Date: 2008-07-16 02:31 pm (UTC)The guy who said it was thinking 8x. I interpreted 8x without a hitch. A third listener said "wait, what?"
And that prompted the question!
(no subject)
Date: 2008-07-16 03:12 pm (UTC)(no subject)
Date: 2008-07-16 03:38 pm (UTC)This tells me that the third listener didn't have the context.
Which is not an indictment - At one point this morning I was forced to stop thinking in German to answer a question, and it took me a good fifteen seconds to formulate a comprehensible English sentence.
(no subject)
Date: 2008-07-16 03:39 pm (UTC)(no subject)
Date: 2008-07-16 06:53 pm (UTC)advisorboss that the device was getting a current three orders of magnitude higher, when I meant 8 times higher, there would have been much confusion and then mockery.10 is just the default around physical sciences, I guess.
(no subject)
Date: 2008-07-16 06:59 pm (UTC)Whereas everything is twos, in software. Even the 1000s are really 1024.
(no subject)
Date: 2008-07-16 02:26 pm (UTC)(no subject)
Date: 2008-07-16 02:35 pm (UTC)(no subject)
Date: 2008-07-16 02:40 pm (UTC)(no subject)
Date: 2008-07-16 02:48 pm (UTC)Wikipedia Pete is a bit of a dick sometimes, though.
(no subject)
Date: 2008-07-16 03:41 pm (UTC)They attend the Church of our Saint Cliff Claven together.
(*Unless there really is a real person you're referring to and you're not merely coining an anthropomorphic synecdoche.)
(no subject)
Date: 2008-07-16 03:58 pm (UTC)There isn't and he is.
(no subject)
Date: 2008-07-16 05:27 pm (UTC)(no subject)
Date: 2008-07-16 07:04 pm (UTC)(no subject)
Date: 2008-07-16 02:52 pm (UTC)(no subject)
Date: 2008-07-16 02:53 pm (UTC)(no subject)
Date: 2008-07-16 04:22 pm (UTC)(no subject)
Date: 2008-07-16 07:08 pm (UTC)I mean, seriously. Yikes.
[1]: Who is not a computer
(no subject)
Date: 2008-07-16 08:06 pm (UTC)Programmers who are trying to work with floating point numbers and need to maximize accuracy of operations need to understand binary fractions. If only to have it beaten into their heads why floating point arithmetic should always be viewed with suspicion.
I'm not sure you ever do actual math In The Real World, but when you're trying to optimize a system, or locate a source of inaccuracy, understanding how it works is vital.
(no subject)
Date: 2008-07-16 10:10 pm (UTC)(no subject)
Date: 2008-07-17 06:08 am (UTC)standsit corrected.(no subject)
Date: 2008-07-16 03:09 pm (UTC)(no subject)
Date: 2008-07-16 03:38 pm (UTC)(no subject)
Date: 2008-07-16 03:41 pm (UTC)(no subject)
Date: 2008-07-16 04:04 pm (UTC)(no subject)
Date: 2008-07-16 04:13 pm (UTC)(no subject)
Date: 2008-07-16 04:25 pm (UTC)( there's a reason why I need Excel to do any time math more complicated than "Adding an hour" or similar. Base 60 makes my brain shut down in confusion )
(no subject)
Date: 2008-07-16 11:46 pm (UTC)In the world of Assembler (or assemblerish C/C++, like bit shifting), base 2, base 8 (octal) or base 16 (hexadecimal) might be appropriate, but that depends on the kind of processor you're dealing with.
(no subject)
Date: 2008-07-17 12:21 pm (UTC)And, in any case, the words "doubled" and "halved" are available and easy to use.
(no subject)
Date: 2008-07-18 05:30 am (UTC)So ya, I don't think, in the tech world it would be at all confusing to hear Orders of Magnitude referring to base 2. I might ask if you meant 2 or 10, but I wouldn't be all "whaaaaaaat?!" if you said 2.