Talk:The Art of Computer Programming

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Misc[edit]

The first paragraph mentions upcoming chapters 9 and 10 but gives them different topics then they are given in the table of contents. — Preceding unsigned comment added by 162.83.142.110 (talk) 18:30, 7 May 2004 (UTC)[reply]

maybe on the site of Dr. Knuth you can find the correct information, as this is a wiki, your are allowed to change the wrong entry.
Ppareit 19:44, 7 May 2004 (UTC)[reply]

Vol 4, fasc 4.[edit]

Addison-Wesley says that Vol 4, fasc 4 was published on 2/6/06. Amazon says 2/7, B&N says 2/10, but both of them say that it isn't available yet. Bubba73 (talk), 22:10, 10 February 2006 (UTC)[reply]

83 pages[edit]

I checked and 83 is now the correct number of pages for the Boolean Basics fascile. Sorry for thinking it was vandalism. Vandals do like to change one digit. Bubba73 (talk), 02:35, 18 February 2006 (UTC)[reply]

Ambiguous re. typeset[edit]

'In 1976 Knuth prepared a second edition of Volume 2, requiring it to be type set again....'

No inidcation is given to the relevance of the typeset, for instance, why couldn't Knuth use the same type-set as everybody else? Some explanation should be given. —Preceding unsigned comment added by 86.129.218.204 (talkcontribs)

I'm not sure enough of this to state in the article, but I think it had to do with the technology of the time - thirty years ago. A new edition of the book had to be typeset again. An old technology was called hot type, where they poured molten lead into a form to make a plate that was used to print. Later they used photo offset which produces a low qualty product. Since the book was going to have to be typeset again, Knuth decided to work on typesetting by computer. (He thought it would take six months, it took 10 years.)
I'm not sure enough of these details to put them in the article. Perhaps someone can verify or correct them. Bubba73 (talk), 00:21, 7 October 2006 (UTC)[reply]
You'll find them repeated and cited in the TeX article. RossPatterson 22:28, 7 June 2007 (UTC)[reply]

New publishments[edit]

The third edition of volume I and volume II have already been published! —Preceding unsigned comment added by 205.209.103.230 (talkcontribs)

That happened years ago, and it is noted under "current editions". Bubba73 (talk), 15:15, 18 November 2006 (UTC)[reply]

How many books were planned?[edit]

I read here that, "Knuth began the project, which was originally planned to be one book, in 1962". Yet on the page about Knuth it says, "...where he became a professor and began work on The Art of Computer Programming, originally planned as a seven-volume series". Can somebody resolve if it was planned as one book or a seven-volume series?

That is confusing. It was originally planned to be one book. When he started turning in the manuscript, it was clear that one book wouldn't do it, so it was planned to be seven books. Now volume 4 itself will be at least three books. Bubba73 (talk), 02:31, 8 February 2007 (UTC)[reply]
I ran across a reference at work today (on a Scientific American capsule biography, 1977) that the work was at that time "to be seven volumes" - fair enough - and, more amusingly, that the first five were expected to be completed "by 1980"... Shimgray | talk | 17:27, 9 October 2007 (UTC)[reply]
Yes, and even in 1980 he said that volume 4 would be done in a year.  :-( Bubba73 (talk), 18:54, 9 October 2007 (UTC)[reply]

Regarding the paragraph on Knuth's use of assembly language:

It is not clear to me why the author thinks that readers of these volumes would consider translating the MIXAL implementations to a high-level language. The algorithms themselves are the best source of the logic to be implemented. They are written in English so have no innate preference for the idiosyncracies of any particular HLL (or low-level language).

Beginning in 2000, I regularly taught a senior course in data structures using Volumes 1 and 3 of TAoCP. The implementations were in IBM assembler. I couldn't imagine having the students translate the MIXAL examples to IBM assembler; on the other hand, the algorithms themselves are readily coded frequently with one machine instruction for each logic statement.

Kcats 14:12, 13 March 2007 (UTC)[reply]

More trivia[edit]

There is a reference to The Art of Computer Programming in a book called Tea with the black dragon by R. A. MacAvoy. Not sure if this is really worth mentioning though, has anyone else heard of this book? Well, it does have an article. m.e. 02:46, 18 June 2007 (UTC)[reply]

Well, I've read it (because of a review on lostbooks.org), but I must've missed that reference completely. --Gwern (contribs) 03:20 18 June 2007 (GMT)

Fair use rationale for Image:Art of Computer Programming - Cover.gif[edit]

Image:Art of Computer Programming - Cover.gif is being used on this article. I notice the image page specifies that the image is being used under fair use but there is no explanation or rationale as to why its use in this Wikipedia article constitutes fair use. In addition to the boilerplate fair use template, you must also write out on the image description page a specific explanation or rationale for why using this image in each article is consistent with fair use.

Please go to the image description page and edit it to include a fair use rationale. Using one of the templates at Wikipedia:Fair use rationale guideline is an easy way to insure that your image is in compliance with Wikipedia policy, but remember that you must complete the template. Do not simply insert a blank template on an image page.

If there is other fair use media, consider checking that you have specified the fair use rationale on the other images used on this page. Note that any fair use images lacking such an explanation can be deleted one week after being tagged, as described on criteria for speedy deletion. If you have any questions please ask them at the Media copyright questions page. Thank you.

BetacommandBot (talk) 07:57, 2 January 2008 (UTC)[reply]

AMSCI link[edit]

American scientist link is broken. did they purge their old pages or did their syntax just change? 65.30.177.9 (talk) 02:38, 20 June 2008 (UTC)[reply]


Chapter corrections[edit]

Knuth's homepage show sections 7.2 and 7.3 in volume 4A. So I will take the liberty of correcting that in the article. 187.78.144.170 (talk) 19:11, 7 November 2009 (UTC)[reply]

Anyone got edition 3 of Volume 2? I have edition 2. On page 26 of edition 2, it offers a method of random-number generation: X[n] = X[n-24]+X[n-55] mod m - equation (7) but says on the next page "The only reason it is difficult to recommend sequence (7) wholeheartedly is that there is still very little theory [about it]"

As it happens, I've discovered some theory and I want to write to DEK about it but only if no-one else has already thought of it. Ergo, my questions: 1) Does edition 3 onwards say that there is any relevant theory not mentioned previously? 2) Does anyone know of any theory, other than what's in edition 3? — Preceding unsigned comment added by 188.220.70.97 (talk) 22:35, 10 February 2013 (UTC)[reply]

The third edition does not have that line. He goes on to state the exact length of the period. Bubba73 You talkin' to me? 02:54, 11 February 2013 (UTC)[reply]

Optimization in chapter 4[edit]

Is "optimization" in chapter four about Optimization (mathematics) or Program optimization? Bubba73 You talkin' to me? 21:28, 6 May 2011 (UTC)[reply]

 Fixed Bubba73 You talkin' to me? 05:03, 7 May 2011 (UTC)[reply]

Thanks for noticing the fix!  :)
 Kiefer.Wolfowitz 14:14, 7 May 2011 (UTC)[reply]
I thought that was the case, but I was busy making a bunch of edits. Bubba73 You talkin' to me? 14:47, 7 May 2011 (UTC)[reply]
Well done!
:-)
Now optimization is redirected to mathematics|mathematical optimization, which now lists a DAB page listing other uses of "optimization".  Kiefer.Wolfowitz 00:52, 8 May 2011 (UTC)[reply]
Thanks. "Optimization" linked to "program optimization" for I don't know how long (I didn't do it). I thought it was probably mathematical optimization, based on the previous sections, but recursion followed it, so I thought it might be program optimization. Bubba73 You talkin' to me? 01:08, 8 May 2011 (UTC)[reply]

Good job guys[edit]

This is one of the few programmer-centric pages on Wikipedia that isn't horrible. Good job on staying focused and to the point on this one. —Preceding unsigned comment added by 74.176.210.68 (talk) 03:01, 16 May 2011 (UTC)[reply]

Rated problems[edit]

The TAOCP page has DEK's checks, but it lacks Don's descriptions of problem ratings from 1..50. A friend (now deceased) who was introduced to Don before the friend died similarly rated his problem exercises using his version of the 1..50 scale. So Don inspired some small number of other authors to rate their problems as well. 143.232.210.38 (talk) 19:54, 3 May 2013 (UTC) P.S. Holder of 2 checks, 1 decimal, other other hex.[reply]

The problems are rated with difficulty 0 to 50, 0 being "extremely easy", 50 requires "research", so for example Fermat's Last Theorem originally had a rating of 50 and now is at 45. 194.207.86.26 (talk) 07:39, 11 May 2019 (UTC)[reply]
A gentle reminder that the scale is logarithmic and subjective. 2A01:CB0C:CD:D800:F45E:2108:85C6:F28A (talk) 09:11, 8 July 2021 (UTC)[reply]

Volume 1, Fascicle 1[edit]

https://en.wikipedia.org/wiki/The_Art_of_Computer_Programming#Current_editions says: « Volume 1, Fascicle 1: MMIX … (will be in the fourth edition of volume 1) »

I just emailed Pearson to ask: « When will the 4th edition of TAOCP vol.1 be printed? I can't wait to have an edition with MMIX in the main text. »

The Editor-in-Chief's reply: « I assume you're talking about incorporating the MMIX fascicle into volume 1? No plans to do that at the moment. »

98.207.155.81 (talk) 21:35, 15 April 2015 (UTC)[reply]

Contradiction[edit]

"In June 1965, Knuth finished the first draft of what was originally planned to be a single volume of twelve chapters. His hand-written first-draft manuscript (completed in 1966)..." Can someone with a source reconcile this? Zeke, the Mad Horrorist (Speak quickly) (Follow my trail) 06:10, 9 March 2019 (UTC)[reply]

4B next month (or not!)[edit]

I got a notice from Amazon that volume 4B is to be released May 17, 2019 (ISBN 978-0201038064). Bubba73 You talkin' to me? 02:07, 10 April 2019 (UTC)[reply]

I undid your edit. Amazon's date is almost certainly in error, possibly the result of a date that was estimated years ago.Volume 4B is expected to include fascicles 5,6,7. Fascicle 5 is estimated for availability in November, 2019. Fascicle 7 is currently nowhere to be found72.211.204.66 (talk) 23:29, 22 April 2019 (UTC)[reply]
Check the reference I added to the end of the introduction. Bubba73 You talkin' to me? 23:36, 22 April 2019 (UTC)[reply]
Amazon shows that Volume 4C will be available in 2005, ISBN 0201038072. Don't be fooled!72.211.204.66 (talk) 23:51, 22 April 2019 (UTC)[reply]
Also announced by the publisher "ISBN-10: 0201038064 • ISBN-13: 9780201038064 ©2019 • Addison-Wesley • Cloth, 640 pp Estimated Availability: 28 Jun 2019" http://catalogue.pearsoned.co.uk/educator/product/Art-of-Computer-Programming-Volume-4B-The-Combinatorial-Algorithms/9780201038064.page (Although Wikipedia says that primary sources are not allowed either!) 194.207.86.26 (talk) 22:00, 28 April 2019 (UTC)[reply]
Knuth is very methodical. He releases online pre-fascicles, collects feedback, then releases paperback fascicles, collects more feedback, then and only then releases hardback volumes. Fascicle 5 is currently estimated for November 2019, and Fascicles 5 & 6 are expected to comprise only the first two thirds of Volume 4B. So don't rely on self-contradictory near-term publisher estimated release dates for Volume 4B.72.211.204.66 (talk) 21:38, 1 May 2019 (UTC)[reply]
Amazon.com appears to have retracted their estimated release date. Volume 4B doesn't even appear now on Amazon.com using title or ISBN search, although google ISBN search can find it, but with no estimated release date. 72.211.204.66 (talk) 17:29, 8 May 2019 (UTC)[reply]
Volume 4B has been released and article has been updated with relevant details. Devpro1981 (talk) 06:28, 5 November 2022 (UTC)[reply]
That's true. On April 9, Amazon sent me the following message, and I assumed that it was about to ship:

We have good news! One of your pre-ordered items is now eligible for release date delivery and has been upgraded at no additional charge. Your new delivery estimate is:

Knuth, Donald E. "The Art of Computer Programming, Volume 4B: Combinatorial Algorithms (Series in Computer Science & Information Processing) (v. 6)" Estimated arrival date: May 17, 2019

Bubba73 You talkin' to me? 17:34, 8 May 2019 (UTC)[reply]

It looks like Pearson just changed one of their estimates from 5/7/2019 to 5/7/2020. 72.211.204.66 (talk) 17:48, 8 May 2019 (UTC)[reply]

need this sentence?[edit]

At the end of the lead " Near-term publisher estimates put the release date at May or June of 2019, which proved to be incorrect." Do we need this? I'd rather remove it. Bubba73 You talkin' to me? 06:37, 25 July 2019 (UTC)[reply]

The problem of a serious discrepancy in publisher estimated release dates for Vol 4B hasn't gone away. Reference #3 still shows an imminent release date, recently bumped from June 2019 to July 2019, but still not based in reality. Musicman103 (talk) 18:18, 3 August 2019 (UTC)[reply]
I'm in favor of taking that out. Bubba73 You talkin' to me? 18:51, 3 August 2019 (UTC)[reply]
I prefer to keep this sentence. The release date of volume 4B is of

interest to computer nerds. This article also offers a better overview of the state of TAOCP than Knuth's own website, so that it may be used as a handy reference. — Preceding unsigned comment added by 5.150.92.130 (talk) 18:17, 10 September 2019 (UTC)[reply]

I'm in favor of taking that out.. We only know that Fascicle 5 will be released this November. That's almost certain, because Knuth has already sent his work to the publisher (he stated that in summer and in Spetember he wrote it's in the printer's hands). Volume 4B will not be released soon. He still needs to write remaining 1/3 of it, what will become firstly Fascicle 7 and only after that Fascicles 5+6+7 will be released together as Volume 4B. Daevid (talk) 01:19, 1 October 2019 (UTC)[reply]

Update: The release of fascicle 5 is estimated no earlier than 2020-02-20 according to https://www.informit.com/store/art-of-computer-programming-volume-4-fascicle-5-mathematical-9780134671826 — Preceding unsigned comment added by 5.150.92.130 (talk) 14:44, 30 November 2019 (UTC)[reply]

You are looking at the ebook publication date. The hardcopy publication date is shown as 22 November 2019. https://www.informit.com/store/art-of-computer-programming-volume-4-fascicle-5-mathematical-9780134671796 Musicman103 (talk) 07:14, 2 December 2019 (UTC)[reply]

If fascicle 5 indeed already released, the article needs to be updated in several places... — Preceding unsigned comment added by 5.150.92.130 (talk) 14:38, 2 December 2019 (UTC)[reply]

Why associate Chaper 22 of "Selected Papers on Design of Algorithms" with "Recursion" ?[edit]

The current chapter outline says: "Chapter 8 – Recursion (chapter 22 of "Selected Papers on Analysis of Algorithms")"

I'm looking right now at Chapter 22 of "Selected Papers on Design of Algorithms", which is a reprint of "Semi-Optimal Bases for Linear Dependencies", originally published in Linear and Multilinear Algebra 17 (1985). It is a four page paper, with the following abstract (https://www.tandfonline.com/doi/abs/10.1080/03081088508817636?journalCode=glma20):

> Let A be an m × n matrix of real or complex numbers, and let μ. be a given constant ≥ 1. If A has rank m, it is possible to choose m columns of A such that, if B is the m × m matrix formed by these m columns, all entries of B −1 A are less than or equal to μ in absolute value. Moreover, if μ > 1, it is possible to find m such columns in a number of steps that is polynomial in m and n and inversely proportional to log μ.

The word "recursion" does not occur in the text of chapter 22. The book *does* list five entries for "recursion" in the index, but none of them are associated with chapter 22.

Therefore: Was this change in error? https://en.wikipedia.org/w/index.php?title=The_Art_of_Computer_Programming&diff=prev&oldid=875169161

Pnkfelix (talk) 13:59, 9 November 2022 (UTC)[reply]

Note that the "Selected Papers on Design of Algorithms" that you're reading (https://cs.stanford.edu/~knuth/da.html, the seventh of eight collected-papers volumes) is different from "Selected Papers on Analysis of Algorithms" (https://cs.stanford.edu/~knuth/aa.html, the fourth of the eight volumes), in which Chapter 22 is called "Textbook Examples of Recursion" (and is a slightly updated version of https://arxiv.org/abs/cs/9301113). So at least whoever made the change wasn't making an error, though personally I think that article has only mild relevance to be mentioned in the chapter outline here. Shreevatsa (talk) 20:40, 9 November 2022 (UTC)[reply]
\facepalm that was embarrassing. I even managed to transcribe my title without noticing the difference from the line above. Never mind my question then. Pnkfelix (talk) 02:43, 10 November 2022 (UTC)[reply]

When Knuth began the project in 1962 ...[edit]

So, TAOCP has been a work-in-progress monograph for over sixty years! Can it earn a place in the Guinness Book of World Records on that basis? Toddcs (talk) 14:09, 24 May 2023 (UTC)[reply]

Perhaps, but it has organically transformed into the project as we see it now; with hindsight. The project he embarked upon was much more limited in scope. Or let me try to explain that in more detail as one who has in fact written several highly regarded textbooks. You think: Oh, it would be nice to bring all this stuff together in, oh, about a dozen chapters in, tops, 500 pages. But then you notice that things that in your own (specialist's) mind appear to be easy, simple, monolithic, unitary, are actually the opposite of all that: they are subtle and complicated, and if you want to do right by your neophyte reader, they require a lot more careful explanation. And before you know it, the project balloons into a dozen volumes instead of a dozen chapters. All academic editors can tell you tales about trying to reign in their authors. On top of all this, Knuth was infected with the bug of comprehensivitis. (Don't get me wrong, I absolutely love TAoCP.)2A01:CB0C:CD:D800:A8AF:DF88:4C28:2656 (talk) 08:11, 27 October 2023 (UTC)[reply]