Jump to content

Wikipedia:Articles for deletion/Super-recursive algorithm

Add links
From Wikipedia, the free encyclopedia
< Wikipedia:Articles for deletion
This is an old revision of this page, as edited by Yakushima (talk | contribs) at 07:12, 27 May 2008 (formatting). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

Super-recursive algorithm

Super-recursive algorithm (edit | talk | history | protect | delete | links | watch | logs | views) (delete) – (View log)

See below for reasons, I am just trying to correct the form of this nomination Hans Adler (talk) 17:15, 25 May 2008 (UTC)[reply]

PROPOSED: delete this article

I don't believe the topic clears the bar for notability in computer science. Perhaps it merits a brief mention, but without undue weight, in algorithmics-related articles where a single-purpose account who wrote most of the original version here has already contributed.

EVIDENCE

Here is where I have looked for the obvious evidence:

Book searches yield only brief mentions so far

Google book search: Super-recursive algorithms are very briefly mentioned in a few books that (so far, in my searches) show little evidence of actually exploring the topic under that name. These mentions seem to be confined to the kind of kitchen-sink listing of vaguely related work that a serious author might only bother with up-front in order to preempt being bombarded much later by people asking why their work wasn't mentioned.

Peer review scant, at best

Peer reviewed literature: Super-recursive algorithms are discussed at length in papers written by Mark Burgin, who appears to have coined the term. A few of these papers have a co-author. These articles are referenced in other papers by Mark Burgin, but otherwise do not seem to be significantly cited.

Book is not peer reviewed

A monograph by Mark Burgin, Super-recursive algorithms is available from Springer. However, it appears not to have received the benefit of copy-editing by a native English speaker; furthermore, Springer monographs are not peer-reviewed. Amazon.com offers two very brief reviews of this book. One of them is by D.V. Feldman, a mathematician at the University of New Hampshire who, from cursory web searches, seems to contribute quite a few very brief reviews of books on topics outside his specialties. This review says that Super-recursive functions "synthesizes all isolated heresies from the journal literature". The same review also claims that the book is "important"; however, Amazon lists it as about #1,700,000 in sales rank, after over 3 years in print. The other Amazon review is by Vilmar Trevisan. This researcher has a record of publication in areas relating to the design of efficient algorithms for specific purposes (e.g., polynomial factorization), but has not published anything clearly related to the theory of computation per se. His review mentions only that Burgin's book "serves to develop a new paradigm", but mentions no particular groundbreaking results.

The most specific and authoritative review is also the most dismissive

In the discussion of this article, the only review mentioned as discussing Super-recursive algorithms at any length was written by Martin Davis, a mathematician who is a recognized authority in the theory of computation. As noted by computer scientist Vaughan Pratt and others in the discussion with some mathematical sophistication, this review's withering sarcasm is, at best, thinly veiled. The main author and defender of this article, Multipundit, might be forgiven for not detecting just how negative Davis' review is, since (by some odd coincidence) Multipundit's grasp of English seems little better than Mark Burgis' in Super-recursive functions.

Personal comment: A computer science undergrad is perplexed at its publishability

My personal opinion might seem out of place here, but I have studied some computing theory, and for those who haven't, my perspective might help you understand why establishing notability in this case is likely to be difficult, if not impossible. I have read some of Super-recursive algorithms. Frankly, when I see a definition of super-recursive algorithm as an algorithm capable of computing what Turing machines can't, the next thing I expect to see (in a real computing theory book, anyway) is a rigorous proof that there exists at least one such thing. Extraordinary claims call for extraordinary evidence, and this is a very extraordinary claim. But does Burgin then do the math? No. he appears more likely to refer to obsolete fault-tolerant commercial systems for his existence proofs. I admit I am not an expert in computing theory. I have read a few textbooks on it, and a handful of papers; I took a few courses in it at U.C. Berkeley, and graded homework for those courses a few times. And even this experience was almost three decades ago. However, the style of rigorous mathematical argument in this mathematical specialty is not something one soon forgets, and where Burgin discusses super-recursive algorithms, what little rigor I see is superficial at best.

If "not even wrong", not notably not-even-wrong

Wolfgang Pauli once said of a particularly shoddy piece of physics work, "it's not even wrong." From what I can see, Burgin is not even wrong in what he claims about super-recursive functions. And others in a better position than I to judge Burgin's super-recursive functions appear to have -- with one scathing exception -- also agreed this stuff is not even wrong, with their resounding silence: there just isn't a whole lot to say about it. Note that "wrong" doesn't make anything "not notable"; far from it. I could (and have) argued that Lotfi Zadeh was wrong, that Fuzzy Logic was inferior to Bayesian approaches to reasoning under uncertainty. But Fuzzy Logic did become notable, whatever its faults, and from a certain point of view, maybe it's good that it did -- reasoning under uncertainty ("is there any other kind?" someone once quipped) needed a push, and Zadeh gave it that push. (Also, to his credit, he didn't push past any reasonable point, he began yielding gracefully to Bayesianism, if anything.) What has Burgis achieved, except to claim he has some umbrella concept that he can't rigorously describe?

Conclusion: DELETE

Burgin's super-recursive algorithms have not achieved notability in computing theory, even though they purportedly comprise fuzzy logic systems somehow. Nor have they achieved notability anywhere else, apparently. It's not that Burgin is wrong. It's not even that he's not even wrong. It's that this supposed theory of super-recursive functions is not notably not even wrong. Therefore, even in the narrow and rather obscure discipline of computing theory (which I would contextualize here by noting that Hartley Rogers' lovely classic text is ... well, not even as high as #400,000 in Amazon sales rank), I don't see that we have Wikipedia notability here.

So I say delete. Yakushima (talk) 10:39, 25 May 2008 (UTC)[reply]

Postscript

I will continue work on the article, trying to get it into shape, trying (probably failing) to be objective and neutral, investigating notability. Maybe I'll change my mind. Probably not. Yakushima (talk) 07:12, 27 May 2008 (UTC)[reply]

demus wiesbaden KEEP

I want to keep the arcticle and to see views and arguments (s. talk page) getting incorporated. --demus wiesbaden (talk) 17:16, 25 May 2008 (UTC)[reply]
If you want to keep it, help make a solid case for notability Yakushima (talk) 04:42, 26 May 2008 (UTC)[reply]

Hans Adler WEAK DELETE

Weak delete. Initially I thought there couldn't be notability problems with a Springer book by a UCLA professor. But now I know more, especially after several altercations involving editor Multipundit from UCLA who, I still hope (because of Multipundit's general cluelessness in what should be Burgin's area of expertise), is just one of Burgin's undergraduate students and not Burgin himself. It seems that "super-recursive algorithm" is just a fuzzy buzzword, designed to mean everything and nothing. Given that, the negative review by Martin Davis (which seems to be essentially the only real response by mainstream science), and the reaction of Vaughan Pratt to this article and its author, I think deletion of this article as non-notable fringe science is probably justified. "Weak" delete because I am not entirely sure my delete !vote isn't in part due to the wish to get rid of the ridiculous conflict with Multipundit, who either has a severe conflict of interest or a severe obsession with the topic of the article. I will probably make up my mind and change my vote after I have seen other people's comments. --Hans Adler (talk) 17:37, 25 May 2008 (UTC)[reply]

"... book by a UCLA professor". Atually, it's a monograph, and by a UCLA visiting scholar, not a professor. If the subject of hyper-recursive algorithms has a claim to fame, I think it's mainly because of a special issue of Theoretical Computer Science (journal) on "Super-recursive algorithms and hypercomputation"[1]. However, that special issue was apparently guest-edited by Burgis and Klinger; I don't think any article in that special issue treats of super-recursive algorithms per se except for the one by Burgis and Klinger. If an article in a guest-edited journal is written by the guest editors, is it necessarily peer-reviewed?

Colonel Warden KEEP/MERGE

Comment: Look a little more closely at your "enough citations" results, Colonel. Does this count, for example? I'd say it's more like a Springer advertisement. How about mere listings in the bibliographies of master's theses? Or how about this, not even published in a peer-reviewed journal, just available on an academic website, and only asking, at the end, whether it's possible that the result could be obtained by an "inductive Turing machine"? There's a lot of chaff here, of the kind that can be created by energetically pressing for mentions rather than by doing substantial theoretical work. Once you've cleared away mentions by authors other than Burgin that aren't significant (and notability guidelines say that more than a mere mention is necessary), the only researchers who seem to be persistently using the term "super-recursive algorithm" are Mark Burgin and the occasional co-author. (And in the case of co-authored papers, I have yet to look closely to see if the term gets more than a mere mention.) In one book, a 70-year retrospective on the Church-Turing thesis, Burgin gets a laugh-out-loud quote in one paper that dismisses hypercomputation as ultimately reliant on infinite computing power. The only other paper to mention him defends him stridently, but elsewhere says that calculus, and other parts of mathematics, would "disappear" if the set-theoretic foundations of mathematics were sufficiently eroded. (Well, that's odd -- calculus preceded set theory, IIRC, and I've met people who got quite fluent in calculus who didn't have much, if any, exposure to set theory.) Who takes super-recursive algorithms seriously, and are they actualy computer scientists who have done, and are doing, serious theoretical work on them under that name? Yakushima (talk) 03:04, 26 May 2008 (UTC)[reply]
* You might be mistaking "actually trying" for "reaching". Do you actually know the subject area at all? Yakushima (talk) 14:09, 26 May 2008 (UTC)[reply]

Itub WEAK KEEP

  • Weak keep. It may be "fringey" and not very notable, but still notable enough IMO. There are citations by third parties, and the fact that Springer decided to publish the monograph suggests to me that at least the editors there decided it was notable enough to print. As far as I know, Springer is not a vanity press but a reasonably respectable scientific publisher. --Itub (talk) 08:27, 26 May 2008 (UTC)[reply]
Comment: I encourage you to go to Amazon "search inside" for this book and try reading some of it. There are clearly longish passages that no editor has bothered reading for grammar or sense. (Maybe the manuscript got a spell-check pass?)
Yes, Springer is a reasonably respectable scientific publisher, but that doesn't exclude elements of "vanity press" in its business model. For what Burgis has on offer, you won't find a sucker born every minute -- it is, after all, a computer science title with mathematical symbols in it. However, in view of the Amazon figures for how many copies are new and used for the rather high price of around $30 (given what rubbish this is), I'd guess there is a sucker of the required type born perhaps once a week.
Get your own taste of the drivel, here. The question isn't "How can it be so bad if Springer will publish it?" Rather, it's "What's happening at Springer that they would even bother to read 10 pages of something like this, much less print it?" I'd say that what's happening at Springer is that they (like many publishers) now have ways to get something into print with very low overhead, compared to the bad old days when you had to pay a union wage for a typesetter skilled enough to set mathematical type accurately. Yakushima (talk) 14:09, 26 May 2008 (UTC)[reply]
Oops, sorry, ltub -- I should have looked at your user page before commenting. You don't have a computer science background, do you? I guess if I were coming from chemistry, as you do, and read ten pages of Super-recursive algorithms, I might not notice anything amiss except that I didn't really understand much.
By the way, for future reference, "citations by third parties" is not enough for notability. The subject must have been discussed significantly, not merely cited, by third parties, and in reliable sources. "Reliable" in a scientific context means "peer-reviewed"; Google Scholar is pretty cool, but it's not yet smart enough to tell whether a source is peer-reviewed or not. For example, is Peter Kugel's "It's Time to Think Outside the Computational Box" peer-reviewed? I'm sure an editor or two looked at it, and thought it would amusing for CACM readers. But if you took Kugel's name off it, and tried to run as a research contribution through the gauntlet of theoretical computer science peer-review, it wouldn't pass muster. Kugel's case in point of "super-recursive algorithms" is Programming by example. There are no algorithms in the field of PBE that can't also be run on a Turing machine. Kugel offers up Burgis' bogus "proof" that Turing machines can solve the halting problem. As someone with a computer science education, my first lip-curling reaction is "Who the hell is this guy? He can't have had a proper education in computing theory!" And, in fact, there is nothing in Kugel's publications to suggest that he's ever even taken a course in the subject, much less taught one. It looks to me like he got tenure a long time ago, before the CS field had a well-formed curriculum, and kicked back for a career of writing mildly controversial op-eds in the AI field and musing about computers in education.
Challenge to everyone here: give me one peer-reviewed publication on super-recursive algorithms. Just one. Yakushima (talk) 14:16, 26 May 2008 (UTC)[reply]

Firefly322 DELETE

DGG KEEP

  • keep I do not really want to rely on an argument for deletion based upon the assertion that the subject is not important because "Cybernetics is MIT's version of pseudo-science." I suppose the cybernetics and MIT articles will be proposed for deletion next, along with Norbert Wiener & John von Neumann. I rely on common sense in judging the comments of expert editors. DGG (talk) 22:35, 26 May 2008 (UTC)[reply]
  • Comment I'm just trying to be honest. Obviously, cybernetics and Norbert Wiener are well established topics that aren't someone's pet publication topic like Super-recursive algorithm. To suggest that I would propose MIT for AfD is not so cool and totally absurd, which makes me believe that your vote and your comment here is a knee-jerk reaction. --Firefly322 (talk) 23:10, 26 May 2008 (UTC)[reply]
yes, a slight exaggeration, but we should not delete based upon that the work is not actually of high quality fundamentally. Not the role of Wikipedia, to decide on the academic quality of work in a subject, if there are good references to standard peer-reviewed journals. There's a lot of stuff i personally thing over-exaggerated narrowly-focused studies in Wikipedia that I wouldnt accept as a peer-reviewer, but the question is whether here are references show that people consider it important. This isnt academic peer-review. DGG (talk) 00:27, 27 May 2008 (UTC)[reply]
Comment: Actually, DGG, what you offered wasn't a "slight exaggeration", it was simply wrong. Firefly322 made a comment that you apparently interpreted as being of the form "I don't like topic X, topic Y is no better than topic X, therefore topic Y should not have a Wikipedia article." Firefly322 might fairly be accused of a not-very-useful editorial digression. However, AgF requires that I view Firefly322's submission as being on topic for this page if his/her conclusion supports that view; this discussion is about whether there should be an article on super-recursive algorithms; Firefly322's conclusion was that he/she agrees with what I've written, and what I've written here is a case (with its own editorial digressions, admittedly) fundamentally based on the claim that the topic hasn't achieved notability in the theory of computation because there isn't be any peer-reviewed work on it. There may appear to be peer-reviewed work, but so far, I haven't seen any.
DGG, although you have no apparent computer science credentials, you are a librarian. With skills like yours, you could be more useful than any of the rest of us here, on a specific question very relevant to the deletion issue: is there an independent, peer-reviewed publication on the topic of super-recursive algorithms? As a librarian, you must be aware of the distinctions involved.
If you're game, let me help you get started. Last I checked, CACM is peer-reviewed; however, it consists largely of what passes for light topical reading among computer scientists, and not all articles in it are necessarily vetted by experts in the article's topic. Thus, Peter Kugel's defense of super-recursive algorithms in an issue of CACM doesn't, in itself, make super-recursive algorithms a formally recognized topic in computer science; from what I can see, it is little more than an off-the-cuff comment (one of many in a career apparently consisting of little else) from a computer scientist who is more of a gadfly than a serious researcher; moreover it is an off-the-cuff comment with at least one serious and glaring technical error in it. Nor would 10 such articles by 10 such authors necessarily establish the topic as legitimate within theoretical computer science. And without such support from within the field itself, you either have to look for significant notability (i.e., more than just a mention) somewhere in the popular press instead, or take very seriously the proposition that the topic is, at best, better covered as fringe theory, to be discussed in some other Wikipedia article, but not at a length that suggests it is being given undue weight. Yakushima (talk) 03:29, 27 May 2008 (UTC)[reply]

Vaughan Pratt DELETE

  • Delete The article is too unfocused to judge what it is about. It begins with a long list of examples of "algorithms that are more powerful than Turing machines." That topic is already covered by the article on hypercomputation. The rest of the article focuses on a single class of such examples, namely inductive Turing machines, which the article distinguishes from an ordinary Turing machine as not having to stop after producing its result. While this vague notion could mean various things, judging by the article's reference to Gold it most likely refers to the concept treated in Language identification in the limit discovered by Gold in 1967 and rediscovered by Burgin in the 1980s. Burgin does not make clear whether he wants "super-recursive algorithm" to be the new name for hypercomputation, language identification in the limit, or something in between. In any event no one but Burgin and the two anonymous Wikipedia editors of the article have found any use for the term, making WP:FRINGE the applicable guideline. The controversy is not notable outside Wikipedia (there is nothing controversial about Martin Davis's negative review) whence my recommendation to delete. --Vaughan Pratt (talk) 03:08, 27 May 2008 (UTC)[reply]
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia:Articles_for_deletion/Super-recursive_algorithm&oldid=215222549"