Non-programmer friendly

« previous post | next post »

Brad DeLong linked to a paywalled Financial Times article by Lisa Pollack about problems with spreadsheet usage, and observed that

[C]onsiderations like these make me extremely hesitant when I think of asking my students in Econ 1 next spring to do problems sets in Excel. Shouldn’t I be asking them to do it in R via R Studio or R Commander instead? Audit trails are very valuable. Debuggability is very valuable. Excel ain’t got it…

The first comment, from "Captaindomestic":

I'm biased as a MathWorks employee, but you may want to look into MATLAB. It is really strong in the kinds of data analysis and plotting that econ students need to do. MATLAB has a pretty non-programmer friendly editor and model that helps new users.

At first I misinterpreted the scope of non-, and thought that MATLAB's editor was being described as not friendly to programmers, rather than friendly to non-programmers. Apparently X-friendly is more of a thing than non-X is, at least for me.

On the substantive point, I'd tend to agree with Brad in favoring R.

One trouble with Matlab is that it's expensive, at least as research software goes: $2,150 for an individual license, $500 for a academic license, $100 for a student license, plus significantly more for each of the many required toolboxes (e.g. $200 for an academic license to the Statistics and Machine Learning Toolbox, $200 for the Optimization Toolbox, $200 for the Financial Toolbox (which requires the Statistics Toolbox and the Optimization Toolbox), $200 for the Econometrics Toolbox (which requires the Financial Toolbox), etc. etc.

I used to use Matlab for the programming exercises in the DSP course that I teach. Now I recommend that students use the (free) version Octave, though I keep Matlab running on one of my computers because there are some interesting applications that are still not compatible with Octave. MATLAB/Octave is still the best choice for signal processing, in my opinion — R is not really very well adapted for that sort of thing.

But for general data analysis, statistics, and graphics, R is a better choice anyhow, in addition to being free.

A vaguely-related linguistic point: MathWorks advertises on NPR (though NPR pretends that this is "support" rather than "advertising"…), but one of the NPR voices reading a MathWorks ads yesterday apparently thought that the "Works" part of the name is a verb.

And yes, Math does work, if you do it right, but that's not what they mean.

The mistake is a plausible one, because the use of X-works to mean "place where (heavy industrial stuff of type) X is made or used" — ironworks, steelworks, dyeworks — is limited and now pretty rare.



  1. Peter said,

    August 21, 2015 @ 6:39 am

    I also parsed it as non-(programmer friendly).  I think I read non- as attaching to adjectives, or adjectival phrases, much more readily than to nouns.

  2. Stan Carey said,

    August 21, 2015 @ 6:49 am

    On the minor point about hyphen-related ambiguity: I'd have had to stop and think about "a pretty non-programmer friendly editor" because of the widespread aversion among writers (and even some editors) to multiple hyphenation: hence the likes of anti-social justice websites (i.e., anti-social-justice websites), self-driving car fantasists, anti-police brutality movement, etc.

  3. MattF said,

    August 21, 2015 @ 7:05 am

    I'm a MATLAB user, and I have to agree that it's far better than Excel, particularly for displaying your results. Another candidate for not-Excel is 'SciPy', a software stack based on the Python language that does pretty much what MATLAB does and is free.

    [(myl) SciPy is certainly what all the cool kids are using these days. An added benefit (and simultaneously a drawback) is learning a widely-used general-purpose programming language.]

  4. Bathrobe said,

    August 21, 2015 @ 8:49 am

    I also understood it as "not friendly to programmers". I interpreted it as meaning that because it's not friendly to programmers it must be friendly to ordinary people!

  5. Andrew (not the same one) said,

    August 21, 2015 @ 8:50 am

    Stan Carey: an example of that which is particularly likely to cause confusion is 'black cab-driver'.

  6. A MathWorker said,

    August 21, 2015 @ 9:51 am

    Two comments, as another MathWorks employee:

    * Prof. Liberman's summary of our pricing model excludes the home-use license. Base MATLAB starts at $149 and the toolboxes are considerably less as well. This does not meet everyone's needs, as it is not for academic, research, or other organizational but it should be mentioned, as it might be of interest to many readers. Also many universities already have licenses to MATLAB and many of its toolboxes, so students and professors can use it without even buying the student or academic versions.

    * My understanding of the company name is that it is an intentional pun, meant BOTH in the industrial "works" noun sense and in the verb sense. If you look at the old, pre-CamelCase logo of the company, it quite clearly reads The Math Works.

    Still, fun to see us mentioned here, on a blog I've been reading and admiring since years before I ever became a MathWorks employee!

  7. GH said,

    August 21, 2015 @ 10:11 am

    I seem to remember at least one style guide (possibly something to do with LATEX?) that recommended using different-length hyphens in multi-word compounds, to indicate the order of precedence: a regular hyphen to join the most closely linked compound, and then (unspaced) n-dashes for additional hyphenated bits.



    (The distinction is barely noticeable in this font.)

  8. Mara K said,

    August 21, 2015 @ 10:47 am

    Stan: Please explain what antisocial justice (as opposed to anti social-justice) is. I would like to introduce it to Tumblr.

  9. Dick Margulis said,

    August 21, 2015 @ 10:51 am

    All of these ambiguity problems can be solved by recasting the problem passage a little more verbosely. As much as I think, pace Geoff Pullum, that Strunk & White offers valuable advice to befuddled non-writers tasked with producing comprehensible prose, I think there's too much attention paid to brevity at the expense of clarity.

  10. ThomasH said,

    August 21, 2015 @ 10:56 am

    In a context in which some economics students might be programmers and other non-programmers, I did not see room for any confusion. Exactly how does Mr Lieberman think the term should have been punctuated?

  11. Quodlibet said,

    August 21, 2015 @ 11:01 am

    Can someone introduce me to this "pretty non-programmer friendly editor and model"?

  12. Dick Margulis said,

    August 21, 2015 @ 11:23 am

    GH: The most widely used style guide for academic nonfiction is the Chicago Manual of Style. Because a great many editors follow it more slavishly than is warranted, the solution you describe has been deprecated of late. The en dash (a hyphen with a college education, as one editor described it) is used to join an open compound to an affix: post–industrial revolution (after the industrial revolution, with an en dash) vs. post-industrial revolution (a revolution following the industrial age, with a hyphen); pre–Vietnam war (with an en dash, preceding the Vietnam war) vs. pre-Vietnam war (with a hyphen, a war that predates the existence of Vietnam).

    This places an awful lot of semantic weight on a choice between two glyphs that represent the same punctuation mark (hyphen is a punctuation mark as well as the name of one of the two glyphs that represent it; en dash is a slightly longer glyph that also represents the hyphen). Skilled readers may subliminally be aware of the distinction. Otherwise, it's a nicety noticed only by editors and compositors.

    In any case, Chicago (along with most other style guides) frowns on mixing hyphens and en dashes in one compound, on aesthetic grounds, I believe, preferring all hyphens or, better yet, recasting, as I noted above.

  13. carlageek said,

    August 21, 2015 @ 12:57 pm

    @Thomas H, to my editor eye, clarifying the phrase it is not a matter of punctuation but of rewriting: "an editor that is pretty friendly to non-programmers."

  14. carlageek said,

    August 21, 2015 @ 12:59 pm

    And my editor eye must be half-closed today, as it left a typo in the above. So, let me try again:

    @Thomas H, to my editor eye, clarifying the phrase is not a matter of punctuation but of rewriting: "an editor that is pretty friendly to non-programmers."

  15. Y said,

    August 21, 2015 @ 1:30 pm

    I vote for 'nonprogrammer-friendly' as the clearest typographic compromise, and for Quodlibet's comment as the best commentary on this whole thing.

  16. Brett said,

    August 21, 2015 @ 1:47 pm

    I may have mentioned this here before, but I am quite bemused that the phonetic "en dash" seems to predominate so strongly over the etymological "N dash."

  17. Rubrick said,

    August 21, 2015 @ 2:40 pm

    Since The Elements has already been brought up: As a youth, back when I found that book inspiring, I was always disappointed in the section about hyphen ambiguity (with the example of the Chattanooga News-Free Press), since it described an actual problem (and unlike much of the rest of the book's advice, I still believe that to be so), but no solution whatsoever; it just said "Don't expect the hyphen's power to extend to non-adjacent words" or somesuch.

    My preferred sort-of-solution is to put spaces around the hyphen (or virgule, which has the same problem) when it applies to multi-word phrases: "Chattanooga News – Free Press", "young male / older female". I don't know if any style guides sanction this.

  18. Dick Margulis said,

    August 21, 2015 @ 2:45 pm

    Brett, the en, half the width of an em, has been a typographical measure for a long time. An en dash is simply a dash that is one en wide. An S-curve is shaped like an S. A T-shirt is shaped like a T. A U-turn resembles the shape of a U. An N dash would be shaped like an N, I should think. So I'm not sure where your bemusement is coming from.

  19. DWalker said,

    August 21, 2015 @ 2:57 pm

    "pretty non-programmer friendly editor and model"

    What does "pretty" attach to? Does it mean that the editor is pretty, or that the editor is "pretty non-programmer friendly"?

    Or that the editor is friendly to a pretty non-programmer? If you are a non-pretty programmer, then this is not for you.

  20. Brett said,

    August 21, 2015 @ 3:41 pm

    @Dick Margulis: An "em" is the width of a capital M in whatever font you are using. It's spelling it "em" that I find funny. The original (presumably spoken) jargon meaning was interpretable as "M dash."

  21. Jonathon Owen said,

    August 21, 2015 @ 3:46 pm

    @Dick Margulis:

    An en dash is a dash that is one en wide, but an en was originally the width of the character. (I was taught that it was the width of an N, but Wikipedia says it was the width of an n. But an em dash was the width of an M character.) So that's why Brett is asking why it's not an N dash.

    Also, not every compound formed with an initial letter represents the letter's shape. An A bomb is not a bomb shaped like an A, a U boat isn't shaped like a U, and an X-ray isn't shaped like an X.

  22. Jonathon Owen said,

    August 21, 2015 @ 3:47 pm

    Brett apparently posted while I was still writing that.

  23. Dick Margulis said,

    August 21, 2015 @ 4:03 pm

    Brett, M is a common abbreviation for an em, and N is a common abbreviation for an N. An em was long ago (okay, not that long ago, probably with the movement toward mechanical typesetting in the latter half of the nineteenth century) standardized as a square on the body size (point size), because traditional fonts had M sorts that were very close to that dimension. When punches were recut for the purpose of punching Linotype matrices, the M and N were held to the em and en widths on purpose in standard text fonts. (Condensed and expanded fonts had M and N mats that varied from the the em and en widths.) So an em is not the width of an M in all fonts.

    As for the dashes, the diaskeuast's and proofreader's mark for an em dash is 1/M (written as a stacked fraction), meaning one per em; the en dash is indicated as 1/N. But that's understood as shorthand for the measure, which is always written em or en when spelled out, as the case may be. The same notational convention is used for dot leaders. So the glyph mistakenly called a horizontal ellipsis (it isn't that) in Unicode and in SGML-based encodings (…) is traditionally called a 3/M dot leader, the sort used in tabular matter such as a table of contents. There were also 1/M and 2/M dot leaders, but those have mostly fallen out of use in the computer age. A true ellipsis is built up of dots and nonbreaking fixed spaces (varying from compositor to compositor but often 5/M or 6/M thins).

  24. Quodlibet said,

    August 21, 2015 @ 5:00 pm

    @DWalker The editor is pretty, of course – after all, she's also a model. Working those two jobs probably keeps her too busy to learn programming. But she's friendly and also helps new users.

  25. Garrett Wollman said,

    August 23, 2015 @ 10:59 am

    The term of art for the advertising on noncommercial educational ("NCE") broadcast stations is "underwriting". The legal distinction is that there are certain kinds of language that may not be used in such announcements: no "calls to action" (imperative clauses, essentially), no mentions of prices, no comparisons with or disparagement of competitors — those are the most important rules. You'll notice that on those PBS shows where they sell books and DVDs related to the show after the closing credits there's always some fine print that says "offer made by _____" — this is a legal fiction (we all know the offer is really being made by some telemarketing fulfullment house) to make the sale "self-support activity" (i.e., fundraising, which is allowed) rather than advertising.

    A related restriction for programs that are not produced by a public broadcaster is that the producer/distributor can't pay the stations to air a show, even if it's ostensibly noncommercial in content.

  26. ohwilleke said,

    August 24, 2015 @ 8:16 pm

    On the merits, Xcel is the way to go. The fact of the matter is that outside math and science majors, solid working knowledge of Xcel exceeds solid working knowledge of anything else by 100-1 or more, few people are willing to learn a new software program for tangential use in a single course, and Xcel is used in incredibly sophisticated applied business economics applications by investment bankers, major budget decision makers in government, and the like precisely because they aren't willing to invest in learning another software program either.

  27. DWalker said,

    August 27, 2015 @ 11:09 am


    "@DWalker The editor is pretty, of course – after all, she's also a model. Working those two jobs probably keeps her too busy to learn programming. But she's friendly and also helps new users."

    Well, then the sentence should say "MATLAB has a pretty non-programmer friendly editor and model WHO helps new users", rather than "MATLAB has a pretty non-programmer friendly editor and model that helps new users". Right? :-)

RSS feed for comments on this post