Carlyle and Kernighan
« previous post | next post »
Thomas Carlyle, "Sir Walter Scott":
There is a great discovery still to be made in Literature, that of paying literary men by the quantity they do not write. Nay, in sober truth, is not this actually the rule in all writing; and, moreover, in all conduct and acting? Not what stands above ground, but what lies unseen under it, as the root and subterrene element it sprang from and emblemed forth, determines the value. Under all speech that is good for anything there lies a silence that is better. Silence is deep as Eternity: speech is shallow as Time. Paradoxical does it seem? Woe for the age, woe for the man, quack-ridden, bespeeched, bespouted, blown about like barren Sahara, to whom this world-old truth were altogether strange!
This reminds of something I once heard from Brian Kernighan, 30-odd years ago, about his interaction with the managers responsible for developing million-line software systems for digital switches. Their measure of productivity was the number of lines of code added per programmer-day, and they were concerned that on average this number was pathetically small — say about ten. How can we increase productivity, they asked?
Brian's response was that they should evaluate a programmer's contribution not according to the number of lines of code added, but rather according to the number of lines of code eliminated. This suggestion was apparently not viewed as helpful.
Nor have many publishers taken up the idea of "paying literary men by the quantity they do not write".
The Carlyle quote was cited in a comment on today's Non Sequitur, which needs to be read in sequence with yesterday's:
Update — Elevated from the comments:
From Darwin's autobiography:
I remember a funny dinner at my brother's, where, amongst a few others, were Babbage and Lyell, both of whom liked to talk. Carlyle, however, silenced every one by haranguing through the whole dinner on the advantages of silence. After dinner Babbage, in his grimmest manner, thanked Carlyle for his very interesting lecture on silence.
From Carlyle's journal, for November 26, 1840:
Last night, greatly against wont, I went out to dine with Rogers, Milman, Babbage, Pickwick, Lyell the geologist, &c., with sundry indifferent-favoured women. A dull evening, not worth awakening for at four in the morning, with the dance of all the devils round you. Babbage continues eminently unpleasant to me, with his frog mouth and viper eyes, with his hide-bound, wooden irony, and the acridest egotism looking through it.
Jerry Friedman said,
March 6, 2018 @ 1:33 pm
A theory adopted by Hemingway. In "The Art of the Short Story", he wrote,
"If you leave out important things or events that you know about, the story is strengthened. If you leave or skip something because you do not know it, the story will be worthless. The test of any story is how very good the stuff that you, not your editors, omit."
(In that sentence he seems to have omitted an "is".)
David L said,
March 6, 2018 @ 1:47 pm
OTOH, Mr Carlyle is not exactly the poster-boy for the "leave it out" school of writing. I speak as someone who has read Sartor resartus all the way through. Well, I may have a skipped a word here and there.
[(myl) But we don't know how much of Sartor resartus he might have been paid to leave out. I admit that his publisher may well have run short of leave-it-out funds in the case of that work.]
Anthony said,
March 6, 2018 @ 4:58 pm
"…because I did not have the leisure to make it shorter."
Rubrick said,
March 6, 2018 @ 6:03 pm
In the spirit of this post, I am not leaving any comment whatsoever.
Allen Thrasher said,
March 6, 2018 @ 6:25 pm
Wise words, but not always applicable. Stendhal's Charterhouse of Parma has a rather abrupt and unsatisfactory ending, in spite of being for the most part tightly written and plotted, because its publisher thought it was too long to sell, and the ms of the original ending seems to be lost (or at least was unknown as of the date of publication of the introduction to the Penguin translation I read).
Also, I think in India the extreme length of some works, such as the Mahabharata or the Yogavasishtamaharamayana, is part of their aesthetic effect. But usually such works have abbreviated versions.
JPL said,
March 6, 2018 @ 8:34 pm
Poetry is much better, and not the long poems either.
rcalmy said,
March 6, 2018 @ 8:53 pm
From Sydney Padua's wonderful (and extensively annotated) graphic novel "The Thrilling Adventures of Lovelace and Babbage" I have learned the following wonderful true and relevant Carlyle anecdote, recounted by Charles Darwin:
"I remember a funny dinner at my brother's, where, amongst a few others, were Babbage and Lyell, both of whom liked to talk. Carlyle, however, silenced every one by haranguing through the whole dinner on the advantages of silence. After dinner Babbage, in his grimmest manner, thanked Carlyle for his very interesting lecture on silence."
[(myl) That is quoted from a passage in Darwin's autobiography, which contains other examples of his opinion of Carlyle.
From Carlyle's journal, for November 26, 1840:
Last night, greatly against wont, I went out to dine with Rogers, Milman, Babbage, Pickwick, Lyell the geologist, &c., with sundry indifferent-favoured women. A dull evening, not worth awakening for at four in the morning, with the dance of all the devils round you. Babbage continues eminently unpleasant to me, with his frog mouth and viper eyes, with his hide-bound, wooden irony, and the acridest egotism looking through it.
]
D.O. said,
March 6, 2018 @ 11:12 pm
People are definitely getting payed for not writing. Not exactly in a way Carlyle imagined though. Stormy Daniels did quite well for herself in this department.
Breffni said,
March 7, 2018 @ 6:20 am
More on the maximalist aesthetic in coding: Steve Ballmer on IBM and KLOCs.
[(myl) In the same general spirit is The Mythical Man-Month.]
J.W. Brewer said,
March 7, 2018 @ 9:28 am
There are some genres of not-quite-literary writing, e.g. in journalism and law, where the final product as published/filed must adhere to a strict/inflexible size limit (whether defined in word-count or page-count or column-inch count or what have you). Some writers in those genres do first drafts that may end up being 20% or 30% too long, just to get all the thoughts they have on the topic out on the table, and then spend time editing down to the required size. Others have the knack of doing a first draft that is already the right size and thus requires less subsequent editing time. The second approach is probably more efficient and will ceteris paribus lead to a more remunerative career, so they are in a sense being paid for not writing excess words at the first-draft stage.
F said,
March 7, 2018 @ 10:47 am
.
Topher Cooper said,
March 7, 2018 @ 4:35 pm
In re. Lines of code.
This came out longer than I anticipated, so anyone with no interest in programming languages or the history of software methodology, you can skip this.
There had been some studies, seemingly good ones, that had found (as a matter of observation, rather than a matter of policy) by studying the progress of many commercial development projects, that the number of lines of production quality code produced — including, in addition to just the initial code, designing, commenting, documenting, debugging, clean-up (what would now be called refactoring) testing and bug-fixing — had a mean of 10 lines with only a moderate variance. Essentially the methodology looked at entire projects (whether they were developing a new product or expanding an existing one) and examining the number of lines of code that the resulted being added (ignoring lines of code dropped or modified — the projects were chosen to minimize the impact of this) and divided by the number of work days that it took. There were many replications of varying levels of formality and scale.
This was widely misinterpreted as referring only to the coding itself, allowing developers to be very sure that their coding speed, and therefore ability, was considerably faster than "normal programmers". Managers could also use this misinterpretation to take credit for the superiority of the his (back then it was overwhelmingly usually "his") teams coders. I suspect that this was also the basis of the above quoted standard.
I was rather intimately involved using these studies because I was one of the developers of the BLISS language at Digital Equipment Corporation. DEC had promulgated a policy that 1) any software development project had to use a higher level language instead of assembly language, even in systems programming, or do a full formal review of the decision not to; and 2) if that language was not BLISS there should be a justification within the planning documents for making such a choice (the BLISS team actually felt that the latter was probably a poor policy — except for systems programming, the SS stood for System Software). However, the manager of the VMS development project (Dave Cutler, later of Microsoft, where he was responsible for the conception, design and development of Windows NT — he was still there last I heard) believed passionately that "Real programmers only use assembly" and had set a project standard of coding everything possible in assembly language — "let me worry about the stupid policy". The BLISS team were "point" in arguing against his insistence that everybody working for him use Assembly language, and this was one of our prime arguments that higher level was better (along with studies showing that in practice, assembly language code was less efficient than higher level code, especially if the language had an optimizing compiler — and BLISS was then quite literally the standard for compiler optimization; and assorted other relevant programming productivity studies).
When I was developing an early production quality Object Oriented language (Trellis/Owl) we noticed that the ten lines/day rule didn't apply. It was already well known that the rule didn't apply to what some computer language theorists called "scripting and prototyping" languages, including Smalltalk-80, the progenitor of all subsequent OO languages. We found that use of what would later be called an IDE, a broad built-in support library, support for code-sharing and ability to deal in high-level abstractions completely disrupted the conditions under which that odd constant-lines-of-code rule worked. Since then the development of lots of other tools, both software and procedural/conceptual has buried that rule even deeper.
Rhona Fenwick said,
March 7, 2018 @ 10:11 pm
@F:
.
!