Friday, May 26, 2006

previous entry | main | next entry | TrackBack (0)

Pen and paper rule!

Maplesoft -- "the leading provider of high-performance software tools for engineering, science and mathematics" -- commissioned a survey of "scientists, engineers, and researchers" to find out how they do their calculations. I think they found the results disturbing:

[A]ccording to an international survey, mathematical calculations in engineering and academia are still most often performed with pencil and paper. On a daily basis, respondents turn to scratchpads and calculators more frequently than any other tool for mathematical tasks. The same survey also revealed this community largely considers its field of work and study to be “fully modern” and “taking full advantage of modern tools and technology.”

These results are drawn from an extensive, international survey of scientists, engineers and researchers across a variety of markets, including aerospace, automotive, electronics, telecommunications, pharmaceutical, life sciences, finance and education. With more than 2000 participants, the survey offers unprecedented insight into the daily practices, experiences and perceptions of the technical user community.

When questioned about how frequently they used a range of tools and resources for design and analysis:

  • 52% indicated that they use “hand calculations (calculators) and paper” daily, with an additional 21% citing it as a weekly practice;

  • 47% of respondents indicated that the next most common resources used daily are “electronic references and tables (e.g. CD-ROM, Web),”)” with another 26% using them weekly;

  • 35% indicated that they use “print reference books and tables” daily, with another 31% using them weekly; and

  • 39% indicated daily use of spreadsheets, which remain the most common software tool used in analysis and design. Another 31% of users employ them weekly.
  • “It is startling to see such hard data revealing the continued reliance on tools and practices that require so much manual effort and leave so much room for error,” said Jim Cooper, CEO of Maplesoft. “This is a user base that is charged with driving innovation, exploring the cutting edge and bringing the best new products and services to market and yet, to a large extent, they are holding onto outdated and outmoded practices. So much of their important work will remain locked in their notebooks and lost to the layers of their spreadsheets rather than captured and carried forward with all of their logic and thinking documented.”

    Count me among the pen-and-paper crowd, sort of. There's no way in hell I'd start any theoretical modeling by typing it into a computer program. On the other hand, there's no way in hell I'd do any kind of statistical analysis or straight number-crunching by hand. Looking at the survey itself, it seems that engineers think of design in the same way that I think about theoretical modeling -- which makes intuitive sense to me.

    My question to readers: Is my use of pen-and-paper is simply an artifact of my age, and as people who have used computers since they were in diapers enter the scientific workforce, they will discard these ancient tools? Or is there something about the act of scribbling down initial thoughts about models or designs on paper that makes it work better than electronic entry?

    [You meant pencil and paper, right?--ed. I'm left-handed, and therefore stopped using pencils at the earliest moment possible.]

    posted by Dan on 05.26.06 at 10:09 AM


    > My question to readers: Is my use of
    > pen-and-paper is simply an artifact of
    > my age, and as people who have used
    > computers since they were in diapers
    > enter the scientific workforce, they
    > will discard these ancient tools?

    Based on my (limited) observations of fairly computer-intensive kids, I think they might. Certainly graphing calculators and Mathematica spring to their minds before back-of-the-literal-envelope Fermi approximations.

    Which scares me, in all honesty. The worst damage I have seen done in an idustrial environment has been by engineers and operators who have no feeling (intuition, submerged experience, whatever) for what they are doing but who blindly believe what the computer is telling them. Of course, the computer is only correct when it is correct...


    posted by: Cranky Observer on 05.26.06 at 10:09 AM [permalink]

    I'm a working electrical engineer. I was sitting in front of a computer before I entered grade school. And when I have to start a design, or figure out how a circuit should be operating, I reach for a pen (yes, pen) and my pad of quad ruled paper days before I bother to try getting it into the computer.

    The computer, while excellent at putting marks on paper in the ways that it wants, is generally pretty rotten at allowing you to put your own notes and markings down the way you want.

    posted by: Rob on 05.26.06 at 10:09 AM [permalink]

    Why no pencils if left-handed? My husband is left-handed, uses mechanical pencils exclusively. What's the difference?

    posted by: Stacy on 05.26.06 at 10:09 AM [permalink]


    I'm buying up blackboards.

    Oh, did I say chalk?

    Trust me on this one.

    posted by: Michael on 05.26.06 at 10:09 AM [permalink]

    I'm currently getting a law degree and masters of electrical engineering at the same time (i'm a glutton for punishment). While I type all of my law notes on computer, I still use pen and paper for anything engineering. I, too, can't stand to use pencils, though I don't know why.

    All my calculations are done by hand or with a calculator. I can't speak for others "in my generation" - I'm pretty anti-social.

    posted by: Mark Payne on 05.26.06 at 10:09 AM [permalink]

    Though i don't do quant as a political scientist, i was a math major as an undergrad.
    I always reach for pen and paper (or even my wipe-board) when I have a new idea to sketch out.

    Echoing Rob, computers aren't creative and don't deal well with creativity. Computers aren't intuitive and can't recognize things unless you tell it to look for them. Computers can't mix forms--text notes, equations, diagrams, lines and arrows, and doodles are all part of a good idea-sketch. what computer program lets you do all that, and what computer program can make sense of any of that in just one glance?

    posted by: peter howard on 05.26.06 at 10:09 AM [permalink]

    The thing is, that the act of writing things down is part of both the organizational part of problem solving and the visualization of it. This is what we were taught in eighth grade algebra (e.g. Let x = ...).

    I just think better with a pen in my hand.

    I do use Excel in lieu of a calculator, for two reasons:

    1. I don't own a calculator.
    2. Visualization -- you get to see the numbers you've entered without having to use a paper tape.

    posted by: Norman Rogers on 05.26.06 at 10:09 AM [permalink]

    I'm not an engineer or anything so smart, but I did major in music in college, so from a music composition perspective, I agree. The fact is, computers simply aren't flexible enough--even the best music writing software won't let me scribble sixteeth notes fast enough to keep up with the melody in my brain, and the computer screen isn't big enough for me to layout several sheets of paper (windows), plus copies of other works that I may be taking inspiration from. So for me, the software is what I use to create a final, legible-to-others product. In that, it's very good and useful.

    posted by: crusader coyote on 05.26.06 at 10:09 AM [permalink]

    I am not aware of a user interface which can come anywhere close to competing with pen and paper for getting the equations down in the first place. Moreover, often times when you are writing equations for research projects you may be improvising the notation to reflect new concepts. So even sidestepping the difficult of something which could read your handwriting as an input device, it wouldn't have the semantics to deal with new notations.

    posted by: Robert Bell on 05.26.06 at 10:09 AM [permalink]

    This organic chemist uses pen-and-paper (chalk-blackboard, dry erase-whiteboard, too) before he touches a computer. Gotta say though, the computer can be a lot neater than me but dumb like Rob says.

    posted by: Klug on 05.26.06 at 10:09 AM [permalink]

    I'm a chemical engineer, and I've always used a combination of techniques: pencil and paper, and Excel. For a first run-through nothing beats pencil, paper, and my HP calculator. As I do the hand calculation, an efficient spreadsheet practically writes itself. Then I use the spreadsheet to test sensitivity or "what-ifs."

    The problem with left-handers and writing tools is that the heel of the hand tends to drag across fresh writing and smudge what is already nearly illegible writing (in my case). I use a Pentel Rolling Writer™ to avoid smudges.

    I have used Maplesoft products and have a lot of respect for them. Many people need to develop optimization schemes or plot Bessel functions in three dimensions. More power (and money) to them. For simple stuff like troubleshooting a sick piece of equipment it's hard to beat pencil and paper.

    posted by: Ernie Gudath on 05.26.06 at 10:09 AM [permalink]

    Pen and paper and hand calculators are perfect valid technology, as well as cheap and not prone to crashes. I enjoy writing in longhand, but when I write at work (and I'm an attorney), I almost always draft in a very simple text editor like vi or pine, then I have my secretaries format in Word. I find all the formating required by Word (and all word processing software, as well as database programs) to be very distracting, particularly when I'm drafting complex memos or briefs that require me to think, rather than cut and paste from templetes.

    My firm does a significant IP practice, and we've found many engineers and designers start with pen or pencil and paper for the same reasons as stated above: the formating requirements demanded by word processing, database and design software are too distracting. It is much easy to focus intital thoughts with pen and paper. The Microsofts and maplesofts of the world are just going to have to deal with that.

    posted by: Shine on 05.26.06 at 10:09 AM [permalink]

    Personally, I find the computer to clumsy for calculations. For those I use my trusty HP-12C (acquired about 22 years ago). Spreadsheets are invaluable for data gathering and analysis, but setting up formulas to calculations (esp time value) that the 12C does natively is too difficult.

    My daughter is a math major and is now taking advanced undergraduate math courses. She tells me that she does her homework with pencil and paper, and then copies it over by hand for neatness. I asked her about typing it on the computer, but she said that the software is too difficult to use and keyboards are too clumsy.

    posted by: Robert Schwartz on 05.26.06 at 10:09 AM [permalink]

    I rely heavily on Matlab for numerical engineering problems, and even use its command line in lieu of a calculator since I can verify and recall the figures I've entered. And I use Matlab for many things other people would use a spreadsheet for. But for symbolic problems, i.e., actually deriving or manipulating equations, it's mostly pencil and paper. I've tried (and hope continue learning) Mathematica for symbolic math, but it has an extremely steep learning curve and I always end up wrestling with it more than my original problem. Also, I agree with other commenters that the ability to invent notation and write ambiguous or ill-defined equations is important for fleshing out concepts.

    Pet peeve: you cannot buy a notebook of blank writing paper. They always have to put ruling or grids or some formatting on it. What's the matter with starting out with a blank sheet of paper?

    posted by: ArtD0dger on 05.26.06 at 10:09 AM [permalink]

    "Why no pencils if left-handed?" Because most schools (still!) don't know how to teach lefties (not the political type) how to write cursive properly. If you teach a lefty to obtain the same forward slant in his cursive as a righty has, he has to hook his hand over the line of writing. Of course, the left hand then immediately drags across what was just written, smearing the pencil writing and the edge of the hand.

    I used to come home every day from elementary school with the outside of my hand black from pencil writing. My mother was annoyed enough that she did some research, and had me re-learn my cursive writing with a back slant so I could keep my hand under the line like a righty does. Still, though, I have to this day a gut-level sense of pencils being very messy and I virtually never use them, even when I am writing out my equations longhand on paper (I'm with most of the other commenters here in that regard.)

    posted by: Curt on 05.26.06 at 10:09 AM [permalink]

    I'm an economist and I use Mathematica/Matlab much more often than I use pen/paper or a calculator (or STATA for any statistical computation). The main reason it is near impossible to solve a number of problems using pen and paper that need to be solved - for instance, many optimization problems. Further, the software can log all of my work.

    I also use software to calculate integrals and do matrix math. I passed those math exams years ago - there's no sense spending 3 minutes solving an integral when the computer will do it instantaneously.

    posted by: cure on 05.26.06 at 10:09 AM [permalink]

    The computer, while excellent at putting marks on paper in the ways that it wants, is generally pretty rotten at allowing you to put your own notes and markings down the way you want.

    and this
    Computers can't mix forms--text notes, equations, diagrams, lines and arrows, and doodles are all part of a good idea-sketch.

    and this
    I am not aware of a user interface which can come anywhere close to competing with pen and paper for getting the equations down in the first place. Moreover, often times when you are writing equations for research projects you may be improvising the notation to reflect new concepts.

    all hit the mark directly on the head. At the present time, there are no affordable display technologies which can match the flexibility of pen and paper. I have thought often about the characteristics that a touch-sensitive display would have to have to replace (mostly) my use of paper for the creative aspects of a job. My minimum estimate is 11x17", so you can have two 8.5x11" pages at once (make adjustments if you live in an A4 country), 300 dpi resolution, 4-bit gray scale. The touch-sensitive part has to be good enough to mimic the differences between, say, a fine point and medium point pen. And it needs to be thin enough -- an inch is too thick for me to feel comfortable writing close to the right-hand or bottom edges where my hand is "falling off". And it needs to be rugged enough to survive the occasional liquid accident that happens. If you have affordable technology that can deliver this, there's money to be made.

    posted by: Michael Cain on 05.26.06 at 10:09 AM [permalink]

    Excellent thread, Dan.

    posted by: Tom Holsinger on 05.26.06 at 10:09 AM [permalink]

    I'd add that pen and paper:
    1) is more portable
    2) rarely crashes

    ...and I don't have to wait a minute for my calculator to boot up.

    “It is startling to see such hard data revealing the continued reliance on tools and practices that require so much manual effort and leave so much room for error”

    That is a weird quote. Certainly typing requires manual effort. And as far as pen & paper leaving "so much room for error", well, garbage in garbage out. With purely mental calculations certainly there's more room for error, but the study seems to lump that in with use of calculators, which aren't exactly stone age technology. (Scientific calculators can be quite fancy.) If they're using calculators with pen & paper, the room for error won't be that different.

    posted by: dorkafork on 05.26.06 at 10:09 AM [permalink]

    I have an MS in computer science and spend a lot of time writing code. Any time I have to produce a script of any significant length, I work out the program with pencil and paper first. I wasn't really taught to do it this way, it's been painful trial and error which has made me realize that it is much faster in the long run (although perhaps if I was a better programmer, that wouldn't be true to the degree it is).

    Having seen the weird-ass errors that commercial software can sometimes produce, too, I find Jim Cooper's comments to be a little high-handed. Not to mention that most software requires the user to adjust to it in some way or another; that adjustment can close down avenues of thought sometimes unconsciously, and the software isn't going to help.

    posted by: JakeBCool on 05.26.06 at 10:09 AM [permalink]

    Obviously paper-and-pen(cil) is better for hashing things out, while the computer is better for performing many technical actions quickly and precisely.

    As for the left-hander issue: Remember the erasable pen? That was a creation of the handist bigots. The blue ink stained worse than gray pencil lead.

    Lefties of the world unite! You have nothing to lose but your stained pinky finger!

    posted by: Andrew Steele on 05.26.06 at 10:09 AM [permalink]

    I sure wouldn't depend on Maple to design anything critical, it is too state dependent and there is no easy way to be sure the results aren't an artifact. That said, for complex computations nothing but a computer will do. Need to grind through millions of numbers, plot complex functions, find roots, or find an optimum solution to a problem depending on hundreds or thousands of variables? Use a computer. But for higher level understanding it is hard to beat rumination, doodling, and the occasional resort to a calculator.

    posted by: chuck on 05.26.06 at 10:09 AM [permalink]

    I see I'm not the first musician to chime in. As a composer, I first do lots of sketching of ideas with pencil and manuscript before using computer notation software for the final score. I encourage my students to do the same, but, for the most part, they go straight to the computer. In my opinion, this forces them to make too many concrete decisions too early in the creative process. When they get stuck, I encourage them to explore a number of possible continuations using any type of graphical, pseudo musical notation. Though this typically helps them break through the creative wall, alas, they usually start the next piece in exactly the same way.

    posted by: cliff on 05.26.06 at 10:09 AM [permalink]

    Umm...from the theoretical physicist side of things...Jim Cooper is crazy. Maple (and Mathematica for that matter) are wonderful for some things and very insufficient for others. Though I use these programs if I really need to multiply 1.61803 times 3.14159 I rarely need to do such things. More often, I need to do things such as find out how express an exponential function as a sum of spherical bessel functions times legendre polynomials which is quite difficult to do in Maple unless you already know the answer. We don't use scratchpads 'cause we're ignorant. We need them 'cause sometimes they're still just better than Maple.

    posted by: A on 05.26.06 at 10:09 AM [permalink]

    I came back to this thread for a look around, and I would like to tell a couple of stories that may shed some light on the subject at hand.

    I took a course in fluid mechanics under T. K. Sherwood at MIT. Professor Sherwood's announcement that Friday's quiz would be closed book was met with a chorus of groans. Professor Sherwood looked at the class and said, "It may be that some day the best engineering you will ever do will be on the back of an old envelope while you're standing on the corner of a wind-blown piece of equipment in the middle of nowhere. And it will be closed book."

    Another time, the final question on the Friday quiz was: "UNLIMITED EXTRA CREDIT List as many was as you can to measure the flow of water."

    After the papers had been passed in, he said, "I asked that last question for a purpose. Show of hands. How many people said 'stop watch and a bucket'?"

    posted by: Ernie Gudath on 05.26.06 at 10:09 AM [permalink]

    As a theoretical physicist I can count on the fingers of no hands the number of times I have found software to be a better way of working than pencil and paper. I regularly go through a couple of reams of blank paper a month and simply cannot imagine how a computer package would be better suited to fleshing out an idea or getting a feel for how a computation works. The conclusions drawn from this survey sound rather suspiciously biased to my ears.

    posted by: dam on 05.26.06 at 10:09 AM [permalink]

    As a theoretical physicist I can count on the fingers of no hands the number of times I have found software to be a better way of working than pencil and paper.

    Reminds me of a student's observation that abstract mathematicians used chalk on the blackboard while applied mathematicians had power point slides. There is a lot of truth to that.

    posted by: chuck on 05.26.06 at 10:09 AM [permalink]

    It depends on the sort of computations you do. I work with images, so I need ways of poking around with millions of data points. This means Matlab, IDL, et cetera-- Filtering, histogramming, pixel-by-pixel boolean operations are necessities for me, not conveniences. There's simply no way to do it by hand.

    posted by: Matt on 05.26.06 at 10:09 AM [permalink]

    Why do people insist on believing that a new technology will displace an old technology. It has never happened!

    Radio isn't gone. Television isn't gone. Newspapers aren't gone. Books aren't gone. Why in the hell would calculators or tables be gone? It works, it's convenient, it's easy to use, it doesn't cost a fortune or make money for some company, but...

    Oh, I see.

    posted by: Grokodile on 05.26.06 at 10:09 AM [permalink]

    Slide rules are.

    posted by: Klug on 05.26.06 at 10:09 AM [permalink]

    I use Excel most of the time for big number crunching, and SAS or Matlab if I can get one (using Excel for number crunching is like Professor Sherwood's back of the envelope in a windy field, let us not go into the things I've made Microsoft Paint do).

    But for setting up problems, I use pen and paper. Writing maths syntax on a computer is too distracting from actually doing the maths. Plus paper is easy to carry around and I can litter papers around my desk. If I had 7 or 8 computer screens they'd start to match the information I can spread out around me on pieces of paper.

    And I love my calculator. It keeps equations in memory so I can go back and edit them.

    Computers are at least as error-prone as paper or calculators in my experience. As well as errors in inputting information, you can make plenty of errors in the code.

    posted by: Tracy W on 05.26.06 at 10:09 AM [permalink]

    When doing the calculations for my condensed matter thesis I did most of it out the long way (pen and paper) until I got to the really nasty stuff (eigenvalues of a disgustingly complex matrix) at which point moved over to Mathematica. Mainly because I had tried doing out by hand a few times and kept worrying about dropping terms.

    And the Monte Carlo simulations were done either on a Cray or a CM5. I seem to remember outlining the structure of the program using block diagrams on paper, then went directly to the screen for coding individual modules. (Moral: never, ever try to vectorize a 4-D matrix with circular boundary conditions. Debugging your mapping is a bitch.)

    At present, writing on a particular topic I can do directly into Word. Diagramming or organizing--use paper and pen and magic markers and a lot of doodling.

    posted by: tzs on 05.26.06 at 10:09 AM [permalink]

    As an Architect, I frequently do my own structural 'engineering' on spiral-bound grid-paper with the formulas and constants written-down in long-hand. A simple engineering calculator, a special "dimensioning" calculator and a good pencil being my only tools. One reason is that it's not the actual calculation that's complicated, it's the alegrbraic-manipulations of the formulae that needs writing-down; then break-out the calculators to solve for 'x'. The other reason is that I'm required by-law to be able to present my calculations to the Building Department, or atleast have them in my files for review. A computer spread-sheet leaves no paper-trail, and it's too easy to have GIGO when short-cuts or assumption are made...and you can't recreate your steps on the PC.

    When I'm working on structural calcs', I have reference books and 'cheat-sheets' scattered all over my large drafting table while I synthesise the result. There are a few rote' calculations that I do have computer programs for, and they are designed to tracks all the constants and design assumptions for future reference. These I print-out as I go. It's much easier to sift through a stack of neatly-labeled papers than through endless seemingly-identical spreadsheet screens.

    I even have a few slide-rules that I use regularily.

    posted by: Ted B. (Charging Rhino) on 05.26.06 at 10:09 AM [permalink]

    Post a Comment:


    Email Address:



    Remember your info?