Friday Fun Poll: Logarithms

What does "log" mean to you?

View Results

I’ve got a couple of columns in a spreadsheet that shouldn’t be an order of magnitude apart. I don’t think this is the source of the problem, but I still advocate the explicit use of ln and log10 over log. Sigh.


  1. Cosma wrote:

    I am sure I speak for all members of the Binary American community when I say I am deeply saddened to see this otherwise progressive blog reinforce hegemonic decimo-normativity.

  2. yami wrote:

    We here at Green Gabbro aim to deconstruct the false dichotomies enshrined in normative electrical and computer engineering practice, not legitimize them in mathematical notation. We may be pomo but our morally relative permissiveness does not extend to people whose disciplines fail to intersect much with our own.

  3. Florine wrote:

    My rule of thumb is (or was when I still encountered logs): if it’s written by a mathematician, log means ln, if by a physicist, log means 10log.
    Maybe the third option is: if by a computer scientist, log means 2log?

  4. Alejandro wrote:

    I once had a Matlab script cause me endless problems because I had no idea that “log” in Matlab was base e. That never made sense to me. log_e should just be ln. log should just mean log_10. Everything else should be explicit. Of course, this is my bias as a scientist, after having lost many days of debugging to this nomenclature problem.

  5. Andrew Ironwood wrote:

    Back when I was programming in SAS, this used to give me no small measure of mental grief, since the function call in that language for the natural log was ‘log’ and ‘log10’ for the base 10 (and my natural impulse was *always* to invoke ‘log’ whenever I really wanted ‘log10’…)

  6. Lab Lemming wrote:

    I choose a variation of answer 5, a smartass reply for field scientists.

    Although I should point out that sedimentology uses log2 for grainsizes.

  7. Angela wrote:

    I use to think of log as base 10 (when not specified), until my first undergrad math class, now I think it is base e unless it is something I know would be base 10 like various science stuff, but usually I write ln if I mean base e, and a subscript for anything else, to avoid confusing myself or others.

  8. Bill Tozier wrote:

    I am sure I speak for all Kricfalusi aficionados when I point out that the correct answer, “It’s better than bad, it’s good,” is not represented as a major response in your so-called poll. I would expect Cosma to say as much too, but he’s a heathen Luddite when it comes to television.

    We who participate — if only occasionally and in a retro, chichi, ironic way — in the popular culture are becoming increasing ashamed of our know-nothing academic peers.

    And dibs on Godwin’s Law, when we reach that stage.

  9. yami wrote:

    I didn’t want the smartass response to be more appealing than the real ones.

    Does it count to be participating in a retro ironic way, if the pop culture phenomenon in which you’re indulging was actually designed with your hipster sensibilities in mind?

  10. Lab Lemming wrote:

    One last serious comment before I classically derail this discussion.

    The reason that math people can deal with log = ln while scientists need log = log 10 is this:

    Scientists use math to process data, and data are recorded in decimal numbers.

    using ln to relate academic equations is fine, but when I log scale my electron multiplier output, what I’m looking for are the order of magnitude- in base 10- of my signal.

    Now, that derailment.

    Re: Bill’s dibs:

    Hitler thought that integer logs were mathemetical constructs of the lesser races.

  11. C. Birkbeck wrote:

    Back in high school, my math teacher told me that log = log10, loge = ln and logx = logx.

    I don’t see anything wrong with this scheme.

  12. Rew wrote:

    Smartass answer:

    It rolls down stairs and over chairs and over your neighbor’s dog…

  13. Peter wrote:

    My measurements do not occupy that many orders of magnitude. I’m more likely to be doing log-likelihood ratio tests or thinking about time constants; natural is the log of choice.

    Information-wise, bits are the province of people who think too much about computers. Decibans and nats are clearly more elite.

  14. Peter Turney wrote:

    (A) There are two kind of people: those who believe that log without a subscript implies base 2 and those who don’t.
    (B) There are ten kinds of people: those who understand binary and those who don’t. Clearly, log10 and log2 are the same.
    (C) There are three kinds of people: those who believe that log without a subscript implies base e and those who aren’t good at math.

  15. Juar Vandaful wrote:

    first of all there are two kinds of people: those who believe that there are two kinds of people.

    Considering this it just seems natural that log is log e.

    On the other hand it is evident that log is log 10 in most cases, just as 10 is 1010 base 2 and not 2 base 10.

    The interpretation of log as log 2 can only be legimitated by context. Otherwise it strongly indicates a case of missinterpretation of the fact that there are two kinds of people.

Post a Comment

Your email is never published nor shared. Required fields are marked *