Talk:Algorithm/Archive 3

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1 Archive 2 Archive 3 Archive 4 Archive 5

Delisting article

Does not meet current Wikipedia GA standards, the main problem being only 3 inline citations for such a large article is not nearly enough. Judgesurreal777 (talk) 22:33, 1 February 2008 (UTC)Reply[reply]

There are far more inline citations than that; they just didn't happen to be using the <ref> syntax. I'm not sure why this suddenly seems to be mandatory; nevertheless, I've gone through adding <ref>s around them all, and there are now 48 entries in the Notes section. This seems quite sufficient. As no other objections have been made clear, I have listed the article at Wikipedia:Good article nominationsGurch 17:12, 2 February 2008 (UTC)Reply[reply]
The problem here, if I am seeing things correctly, is that the reviewer didn't read the article closely enough to notice the inline references. The solution to that is to point out to him or her that one must carefully read an article before assessing it.
I am very fond of Harvard referencing. The style guide discourages changing from one style to another, just as we don't change from American to British spelling. I restored the parenthetical references. I can argue in favor of this style at great length if asked. — Carl (CBM · talk) 20:34, 2 February 2008 (UTC)Reply[reply]
Complete waste of time using this archaic referencing style, just use the standard formatting that every other article uses. Judgesurreal777 (talk) 22:57, 2 February 2008 (UTC)Reply[reply]
There is no citation format that every other article uses. On the other hand, the use of Harvard references is explicitly permitted by WP:WIAGA. — Carl (CBM · talk) 23:28, 2 February 2008 (UTC)Reply[reply]
The article needs to have more references, and probably a copyedit as it is an older GA; I suppose you can put it into any style you like, but I still believe that it should be reviewed and improved. Judgesurreal777 (talk) 15:57, 3 February 2008 (UTC)Reply[reply]
While arguably good in printed text, I do not see the use of non-hyperlinked Harvard referencing in a hypertext. Apparently it has just the effect of having to manually scroll up and down the article a lot more... Goochelaar (talk) 17:22, 3 February 2008 (UTC)Reply[reply]

This article was delised from GA at least a year ago, maybe two. In response I added the history stuff appearing at the end, worked on the beginning, plus I hugely expanded the notions of "algorithm" in the secondary articles, added almost most of the references that you now see, etc, etc. So it has been copyedited within the past year or so. The only weakenesses in terms of references that I can see appears in the middle -- an area I felt unqualified to expand and footnote (if it were up to me alone I'd just refer everyone to Knuth) -- and the little note that the section about patents should be expanded (that note, by the way has been there for at least a year, if not longer). With regards to writing about the types and plethora of algorithms that exist, all that we can do is refer the reader to articles in wikipedia and Knuth's volumes (he's working on another, and possibly a revision of the existing set of three, but he works in Knuth-time, not in wikipedia time). Other than in these areas I can see no sense whatever in adding more footnotes. As to footnote style? I and another editor did convert the Hilbert article over from inline to the "ref" style but only after I filled it up with inline cites, came to agree with the other editor who had tried to discourage me, and then saw what an awful thing it had become with all the inline cites. Like Carl I'm a fan of inline cites, but I have kind of drifted toward the "ref" style because an editor can then expand the contents of the citations with more information, if necessary. It seems that wiki articles have drifted toward this "ref" style -- at the expense of a monstrously-long citation section. Bill Wvbailey (talk) 21:23, 3 February 2008 (UTC)Reply[reply]

Your work on adding references is extremely important and much valued. Referencing style only needs to be consistent and Harvard referencing is fine. I've made a couple of edits to the article to suggest how Template:Harv and Template:Citation can be used to make it easier for the reader to follow the Harvard references. I hope someone takes this up. Geometry guy 21:42, 3 February 2008 (UTC)Reply[reply]
I have it on my todo list, but if anyone gets to it before I do I won't be upset in the least. — Carl (CBM · talk) 03:42, 4 February 2008 (UTC)Reply[reply]

More 0s than 1s

I have moved this from the article to this talk page:

In the case of non-halting computation method (calculation procedure) success can no longer be defined in terms of halting with a meaningful output. Instead, terms of success that allow for unbounded output sequences must be defined. For example, an algorithm that verifies if there are more zeros than ones in an infinite random binary sequence must run forever to be effective. If it is implemented correctly, however, the algorithm's output will be useful: for as long as it examines the sequence, the algorithm will give a positive response while the number of examined zeros outnumber the ones, and a negative response otherwise. Success for this algorithm could then be defined as eventually outputting only positive responses if there are actually more zeros than ones in the sequence, and in any other case outputting any mixture of positive and negative responses.

First, it makes no sense to ask whether an infinite binary sequence has more 0s than 1s. Which of thise have more 0s than 1s?


Second, even if there was a meaningful way to read the problem, if the reading was nontrivial there is no algorithm that would solve the problem. For example, the halting problem is 1-reducible to the question of whether an infinite binary sequence has a finite or infinite number of 1s. So this is claiming that there is an algorithm to solve the halting problem? — Carl (CBM · talk) 15:00, 29 February 2008 (UTC)Reply[reply]

I have no idea how this crept in. As it has no attribution it would seem to be someone's notion. I suppose the point is (already made in the section), that an algorithm should always produce something to let the observer know it is "busy" and not locked in a loop. Example: many years ago for a course my little team and I designed a "puzzle box" -- about 6 inch mahogony cube with four brass knobs in the centers of the four sides and a little red LED on the top. The mission for the rest of the class was to figure out how to make the LED light up. The class eventually broke the nice box swinging it around and around, but they did solve the puzzle when they all clustered around the box, all of them, simultaneously touching all four knobs and waited. The waiting was the key, not touching the four knobs simultaneously. As I recall they had to wait about 4-5 seconds, and this just happened accidently for them the first time, and as they fiddled some more they figured it out. Our point -- that we humans demand that successful algorithms (machines) let us know what's going on "promptly", even if no "production" will necessarily be forthcoming. For example, I find vending machines that take my money and sit silently for a few seconds before they drop the drink very unnerving. Ditto for when Windows hangs up. Bill Wvbailey (talk) 16:50, 29 February 2008 (UTC)Reply[reply]
Sure, the idea that an algorithm can take arbitrarily long before returning is different than real life. But in this case, the "algorithm" would need to look at the entire infinite string before returning an answer, something no algorithm can do. But the more serious issue is that "more 0s than 1s" is not well defined in the first place; so we're talking about the existence of an algorithm to solve an ill-specified problem. — Carl (CBM · talk) 16:54, 29 February 2008 (UTC)Reply[reply]
I agree the paragraph should be stricken. 1) Infinity is not a number. 2) The word "more" is a comparative and requires a comparison. Whereas you can compare more with numbers, you cannot compare more with infinity. Therefore, you cannot verify that there are more zeros than ones in an infinite random binary sequence, and it will never be effective. Also, algorithms must end by definition, so non-halting concepts are not germane. Timhowardriley (talk) 18:04, 29 February 2008 (UTC)Reply[reply]
Any infinite string containing a finite number of zeros has more ones than zeros, and any infinite string containing a finite number of ones has more zeros than ones. Any infinite string having an infinite number of zeros and an infinite number of ones has a bijective mapping between the two sets (both are countable) and therefore has the same number of zeros and ones (in the cardinality sense). This isn't terribly relevant to computing though and I support the removal. Dcoetzee 18:14, 29 February 2008 (UTC)Reply[reply]
Hmm. The Ending of the Algorithm (as in The Silence of the Lambs) ... we engineers often design algorithms that don't end, they just loop around and around and around .... It's true that the code (usually) has an end (is finite ...), but the total state of the machine is (always) different from cycle to cycle (if it weren't different, the machine would be locked in a loop). I personally don't believe that algorithms-as-machines necessarily must end, nor, as I think about it, either do evolving algorithms-as-code, i.e. code that writes more code to execute as it "evolves" (grows both physically and mentally, etc). Interesting philosophic stuff, this ... Bill Wvbailey (talk) 22:59, 29 February 2008 (UTC)Reply[reply]
It's true that, in the most general sense, algorithms don't always halt. When we're interested in algorithms for number-theoretic functions, however, then halting becomes an important property.
Even in other contexts, most algorithms that don't halt in practice follow a pattern like this:
There is some current state initialized into some variables
Read or measure some input
Update the internal state variables, or leave them alone
Possibly, produce an output
Clearly such algorithms could be recast as algorithms that, presented with a starting internal state and an input, produce the ending internal state and output (that is, we could just cut the loop). Evolutionary algorithms, similarly, can be recast in this way, or they can have an extra parameter representing time added as another input so that the number of iterations can be specifically passed as an input. Again similarly, algorithms that produce an infinite binary string can be recast by adding an extra input telling which single bit of that infinite string to return and then halt. So there is no actual increase in generality or computing power, in cases like these, in considering algorithms that don't terminate. — Carl (CBM · talk) 23:23, 29 February 2008 (UTC)Reply[reply]
P.S. the viewpoint I am presenting here is the "mathematics" one, where we are particularly concerned with computing in theory. In the "computer science" view, the actual code and programming techniques may be of much more interest, and then of course it makes a difference exactly how the algorithm is presented. — Carl (CBM · talk) 23:35, 29 February 2008 (UTC)Reply[reply]

Edits 2008-3-12

I removed this from the article:

As mathematical and computer science literature demonstrates, there is no consensus on the concept of "algorithm". That is why conventional Turing machines do not satisfy some of these definitions, while inductive Turing machines (Burgin 2005), which are much more general than Turing machines, satisfy other definitions of algorithms.
For instance, Stephen C. Kleene defines algorithm as "a procedure, performable for each set of values of independent variables, which procedure necessarily terminates and in such manner that from the outcome we can read a definite answer, "yes" or "no," to the question, is the predicate value true?" (Kleene 1943: 273). Conventional Turing machines are "performable for each set of values of independent variables," but the corresponding procedure does not "necessarily terminates." It means that in a general case, conventional Turing machines do not satisfy this definition.
At the same time, Michael Sipser writes, "an algorithm is a collection of simple instructions for carrying out some task" (Sipser 1997: 142). Inductive Turing machines satisfy this definition of algorithm because they give simple rules (actually the rules of simple inductive Turing machines are the same as rules of Turing machines) for carrying out some task.

There is, in fact, a great deal of agreement in the literature about what an algorithm is, apart from Burgin, who has an idiosyncratic definition not adopted by any standard computability texts. Separate from that, it makes little sense to talk about Turing machines satisfying the definition of an algorithm - a Turing machine is not an algorithm, although each number theoretic function for which there is an algorithm is computable by a Turing machine. Basically, this text continues the confusion about definitions that Burgin promotes, but which is not present in the standard literature on computability. — Carl (CBM · talk) 01:34, 13 March 2008 (UTC)Reply[reply]

Of course I didn't put this in -- I agree that it is idosyncratic and Burgin probably belongs in Algorithm characterizations. One of these days I'll look at Burgin. Bill Wvbailey (talk) 16:37, 13 March 2008 (UTC)Reply[reply]
Burgin essentially relaxes the requirement that the output of an algorithm must be returned after a finite number of steps, permitting "algorithms" that never actually know if they have yet obtained the correct result. His book is a fine resource for information about hypercomputation, provided the reader knows enough already to detect when Burgin's interpretations, based on nonstandard definitions of common terms like algorithm, are not applicable to the ordinary meanings of the terms. I agree that the characterizations article is a good place to mention his work. — Carl (CBM · talk) 16:43, 13 March 2008 (UTC)Reply[reply]

The advice to take the eliminated text to the article "Algorithm characterizations" looks reasonable and I follow this advice. I completely agree with Carl that Turing machine is not an algorithm but a mathematical model of algorithms. That's why I did corresponding changes in the text. However, I can't agree that now there is a great deal of agreement in the literature about what an algorithm is. Definitions from different sources cited in the eliminated text vividly demonstrate this. Besides, I can't see how this text is related to Burgin’s definition of algorithm as all definitions that are cited are from sources that are in no way related to the theory of super-recursive algorithms. With respect, Multipundit (talk) 21:25, 14 March 2008 (UTC)Reply[reply]

I'd suggest that Burgin's POV should be entered on the Algorithm characterizations page, together with the reference/source. Also see the entry under Gurevich et. al.; my interpretation of his published reports is that he does believe that a suitably-programmed TM (with suitable data, etc) defines/is "an algorithm." Bill Wvbailey (talk) 21:48, 14 March 2008 (UTC)Reply[reply]

A paradoxical situation

Recently I have encountered a paradoxical situation. In the book of Lewis, H.R. and Papadimitriou, C.H. [1] on the page 246, the following Church-Turing thesis is formulated:

"We therefore propose to adopt the Turing machine that halts on all inputs as the precise formal notion corresponding to the intuitive notion of an "algorithm." Nothing will be considered as an algorithm if it cannot be rendered as a Turing machine that is guaranteed to halt on all inputs, and all such machines will be rightfully called algorithms. This principle is known as the Church-Turing thesis."

As the conventional Turing machine does not in general satisfy this Church-Turing thesis, we come to a paradoxical statement:

Turing machine refutes Church-Turing thesis.

Indeed, a universal Turing machine cannot be "rendered as a Turing machine that is guaranteed to halt on all inputs."

By the way, if we take the original Turing machine described in the classical paper of Alan Turing (1936), as mentioned, we'll have more problems with the Church-Turing thesis.

[1] Lewis, H.R. and Papadimitriou, C.H. Elements of the Theory of Computation, Prentice-Hall, Uppre Saddle River, N.J., 1998

With respect to all discussion participants, Multipundit (talk) 20:20, 19 March 2008 (UTC)Reply[reply]

There isn't any contradiction or paradox there. The "standard" Church Turing thesis states that a total number-theoretic function has an algorithm if and only if it is computable on a Turing machine. Lewis and Papadimitrou appear to be referring to that thesis. The thesis make no claims about partial functions or about Turing machines that don't always halt, so there is no paradox.
You shouldn't add this sort of novel theory to the article on the Church-Turing thesis. It is well known that by changing from the canonical statement of the Church-Turing thesis, it's easy to get a different thesis that may be trivially true or trivially false. That doesn't make the actual thesis paradoxical. — Carl (CBM · talk) 20:32, 19 March 2008 (UTC)Reply[reply]
Your argument is essentially that, under the definition above, there is no algorithm to simulate an arbitrary Turing machine and determine its result - and this is correct, since they are choosing to define "algorithm" as a process that always produces a result, and not all Turing machines produce results on all inputs. Dcoetzee 20:56, 19 March 2008 (UTC)Reply[reply]

Sorry for misunderstanding, but I tried to explain that the situation with the Church-Turing Thesis is paradoxical. The Thesis itself (at least, in its main versions) is very rigorous and I can agree with Carl that it contains no paradox. However, neither Lewis and Papadimitriou nor Kleene write about functions as Carl suggests. They write about algorithms and this makes the situation paradoxical. I think that this paradox has an easy solution, but computer scientists themselves have to solve this paradox and not to put it under the carpet. As this paradoxical situation continues to exist, I believe it's necessary to write about it in some article and I thought that the article on the Church-Turing thesis is the best place for this.

With respect to all discussion participants, Multipundit (talk) 01:58, 21 March 2008 (UTC)Reply[reply]

The use of the term 'algorithm' isn't paradoxical, either. If this was indeed a paradox, wouldn't one of the many bright recursion theorists have already noticed it? — Carl (CBM · talk) 02:24, 21 March 2008 (UTC)Reply[reply]
Here are the quotes I promised from Burgin's book.
"The domineering opinion is that the thesis is true. It is supported by numerous arguments and examples..." (p. 44)
"This brings us to a new vision of the Church-Turing thesis. While in general this thesis has been disproved by the invention of different classes of superrecursive algorithms, ..." (p. 46)
I didn't overlook the seeming contradiction there, or the curious word choice in the first quote. — Carl (CBM · talk) 18:49, 21 March 2008 (UTC)Reply[reply]

Multifaceted topics

There are multifaceted topics related to many articles. Very often these articles have different discussion participants. That's why it would be useful to discuss multifaceted topics on all related article pages. This will help to connect discussion participants related to different articles when topics of these articles are closely connected.

With respect to all discussion participants, Multipundit (talk) 01:58, 21 March 2008 (UTC)Reply[reply]

Grounded scientific arguments

In the discussion on the article “Turing machine”, I found some critic of the Selim Akl's work "The myth of universal computation.” However, when you read his paper, you come to the conclusion that this paper is very informative. Akl argues that there is no universal mathematical model of algorithm and no universal mathematical model of computation. My opinion is that Akl gives grounded arguments to support his assertion. Many forget that computation is not only an abstract construction about which they are reasoning but also a physical process. In addition, the situation in computer science, where many researchers have suggested a variety of models of algirthm and computation, supports Akl's assertion. Arguments of Akl are based on physical properties of true concurrency and it would be beneficial for the article "Algorithm", as well as for the article "Computation", to present this grounded point of view. May be it would be the best to ask Selim Akl to write about his approach.

With respect to all discussion participants, Multipundit (talk) 03:15, 21 March 2008 (UTC)Reply[reply]

It would not benefit this article to give much weight to Akl's claims, which like Burgin's rely on changing the definitions of the terms involved, and like Burgin's are not widely accepted in the broader community.
The classical study of algorithms and computability is concerned with functions: the input is recorded once and for all, and then the computational process is initiated. Akl argues that, if the input is later changed during computation, this might result in an incorrect output. The very notion that the "input" to a function can change during the computation of the value of the function differs from the standard understanding of "function". Akl's argument might be relevant to engineering or control theory, but his arguments don't address classical computability theory. — Carl (CBM · talk) 19:08, 21 March 2008 (UTC)Reply[reply]
I agree that Akl's arguments go beyond classical computability theory. However, it seems that the article "Algorithm" would benefit if it represents not only classical computability theory but also its recent development. With respect, Multipundit (talk) 20:37, 27 March 2008 (UTC)Reply[reply]

Harvard Style

I was unfamiliar with the Harvard Style of citations. I have now read the reference given to me, Thanks for the correction. Matthew Glennon (talk) 19:09, 27 March 2008 (UTC)Reply[reply]

Holographic Algorithms

Should holographic algorithms be added to the Classification by design paradigm section? Bender2k14 (talk) 05:18, 11 April 2008 (UTC)Reply[reply]

Iran/Persia issue -page protected

I protected the page for 7 days to give everyone a chance to discuss. One helpful strategy is to present references that have the description you prefer (Iranian or Persian). — Carl (CBM · talk) 17:30, 19 April 2008 (UTC)Reply[reply]

I don't particularly care one way or the other, but there have been lengthy discussions about this at Talk:Muhammad ibn Mūsā al-Khwārizmī, and I believe at Talk:Algebra and WP:WPM as well. I think an RfC might be in order, but someone who knows the sources and precedents would be in a better position than I to set it up. At any rate, I think wider input would be helpful to firm up Wikipedia's position on this. silly rabbit (talk) 17:53, 19 April 2008 (UTC)Reply[reply]

Is Persia different from Iran?

The term Persia was an international term to call the country which is now known as Iran. The people who lived in "Persia" were using Iran for calling their landduring all the history and there is no where in persian literatures that you can find term "Persia" as the name of the land. In contrast, the term "Iran" is used in many old persian literatures like Shahnameh as the name of the land. On 21 March 1935, the ruler of the country, Reza Shah Pahlavi, issued a decree asking foreign delegates to use the term Iran in formal correspondence in accordance with the fact that "Persia" was a term used for a country called "Iran" in Persian. To make it more clear you can think of the terms "Germany" and "Deutschland" both returns to the same land but only "Germany" is used as an international term. If some day Germans ask to call their country "Deutschland" and their nationality "Deutsch", you can not separate their history and say for example Albert Einstein was not Deutsch but was German and you can not say that Deutschland did not not exist. It is exactly the same about Iran and Persia. Another thing which should not mix up with this issue is the Persian Empire. The land which was called "internationally" Persia before 1935, was a part of the old Persian Empire and Muhammad ibn Mūsā al-Khwārizmī was not living in the time of Persian Empire. hope I could clarify this fact.( (talk) 19:06, 19 April 2008 (UTC))Reply[reply]

I think that the disagreement is not on what to call modern-day Iranians - everyone calls them Iranians. My impression is that the people on the opposite side of the issue are thinking of historical figures such as Charlemagne (who is not ordinarily called German or French) and Hammurabi (who is not ordinarily called Iraqi). I hope this helps to clarify the issue.
The best way to make your point that Muhammad ibn Mūsā al-Khwārizmī should be called Iranian is to give references to English-language scholarship that uses that term to describe him. I say English language because the issue here is not one of fact (everyone can agree on that, I think) but of language usage, and so the published usage of native speakers is particularly important. — Carl (CBM · talk) 18:49, 19 April 2008 (UTC)Reply[reply]

I am not discussing only about Muhammad ibn Mūsā al-Khwārizmī, and I am not trying to say he was Iranian but not Persian. I am just telling the term Persia was only an international term and of course in English-Language scholarship it was also used to call the land which people living inside it knew it as Iran and their are enough sources about this issue (Read Reza Shah). Lets go back to German example. If Germans ask to call them Deutschland you can not say no where in English-language scholarships Albert Einstein was called Deutsch so he is not Deutsch but German. Isn't it clear?( (talk) 19:04, 19 April 2008 (UTC))Reply[reply]

It isn't clear, because the dispute isn't about 20th century figures like Einstein. As an example, would you say it is correct to call Hammurabi Iraqi? — Carl (CBM · talk) 19:09, 19 April 2008 (UTC)Reply[reply]

You are confusing Persia with Persian Empire, as I explained above they are two different issues and Muhammad ibn Mūsā al-Khwārizmī was not living in the time of Persian Empire. We can not call Hammurabi Iraqi because even he did not know himself as Iraqi which did not exit that time. But the time Muhammad ibn Mūsā al-Khwārizmī was living in Iran, he knew himself as Iranian. The mistake is here that you think "Iran did not exist that time". Iran existed that time and was called exactly "Iran" by people who living inside it. But no body was calling himself "Iraqi" at the time of Hemurabi. ( (talk) 19:22, 19 April 2008 (UTC)}Reply[reply]

Thanks, I follow what you're saying and I see your point. Can you provide scholarly English-language sources that identify Muhammad ibn Mūsā al-Khwārizmī as Iranian? If so, I think that would be a very strong argument for using the term Iranian here. If, on the other hand, English scholars always call him Persian, that is a strong argument against it. My impression is that this is entirely an issue of word choice, rather than an issue of "correctness". Maybe some other participants on this page will have their own thoughts to add. — Carl (CBM · talk) 19:25, 19 April 2008 (UTC)Reply[reply]

You are going back again to English-language scholarships which as I said are based on the old name of Iran. if Deutschland is called Germany in english scholarships, it does not mean that "Deutschland" does not exist. If some day Germans ask us to call them "Deutsch", then we should call Einstein a Deutsch. We can not say he was not Deutsch because in English scholarships he was called German. ( (talk) 19:32, 19 April 2008 (UTC))Reply[reply]

I follow waht you're saying. On the other hand, in English wikipedia, we don't say that Einstein was Deutsch, we say he was German (or German-born at least). This is why I think it's a matter of word choice, and that we should follow the choices made by contemporary English-language scholarship in the area. — Carl (CBM · talk) 19:36, 19 April 2008 (UTC)Reply[reply]

You'll find that most scholarly sources call him Persian (see references mentioned in the article on him), although at least Toomer calls him "a man or Iranian descent". However, people arguing for calling him Iranian are doing so based on the argument that the terms Persian and Iranian are equivalent. One can in a non-contradictory way argue that the terms are not equivalent. "Persian" would be correct in both interpretations and is therefore the obvious choice to use. —Ruud 19:37, 19 April 2008 (UTC)Reply[reply]

As I said, internationally and therefore in English scholarly the land of Iran was called Persia before 1935. What I say is what is internationally meant by Persian is exactly Iranian. Again I say I am not talking about Persian Empire which was much larger than what was called Persia(and called Iran now days) after that. ( (talk) 19:43, 19 April 2008 (UTC))Reply[reply]

Also we call Einstein German because it is the international word which is currently using for "Deutsch". ( (talk) 19:45, 19 April 2008 (UTC))Reply[reply]

I did not say he was not Persian. I just added Iranian in parenthesis after Persian to make it complete. I think it is a good compromise. ( (talk) 19:58, 19 April 2008 (UTC))Reply[reply]

To Ruud: I really would like to see your non-contradictory argument that Persian is the "obvious choice". —Preceding unsigned comment added by (talkcontribs)

The term "Iranian" is used in two different senses:
  1. Relating to the modern country of Iran.
  2. Relating to the Iranian peoples, a collection of ethnic groups defined along linguistic lines as speaking Iranian languages.
To describe Al-Khwārizmī as Iranian in the first sense is simply wrong, as the modern country of Iran did not exist when he was alive, and in any case his place of birth is not in modern Iran, but is in modern Uzbekistan. To decribe him as Persian Iranian in the second sense is tautology since the Persian people are by definition an Iranian people. It is like describing Ethelred the Unready as Anglo-Saxon Germanic or Gordon Brown as Scottish British. It is not strictly incorrect, but it is unnecessary and confusing. Gandalf61 (talk) 11:15, 20 April 2008 (UTC)Reply[reply]

To Gandal: If you read more about history of Iran, you will see that the country of Iran existed at the time of Al-Khwārizm and his birth place was inside the Iran borders at that time. As explained above, there are many literatures from his time which use the term "Iran" as the name of the country. The modern Iran is of course smaller than what was at his time. The modern Iran and the old Iran, both were called internationally "Persia" before 1935, but the people living inside "Persia" knew themselves as "Iranian" during all the history which we are talking about. I know that because of the political situations in recent years there are some intentions to avoid everything positive from term "Iran", but we are in the Wikipedia and not in the White House. —Preceding unsigned comment added by (talkcontribs)

Indeed we are in Wikipedia, and one of the policies of Wikipedia is verifiability. And yet, despite several requests, you have still not produced any references to reliable sources that support you assertions. Unless you do so, no-one is going to take your unsupported claims seriously. Gandalf61 (talk) 12:40, 20 April 2008 (UTC)Reply[reply]

I hope you have read the article which I mentioned above about Reza Shah. Anyhow in another article in Wikipedia, you can read:

The name "Persia" until 1935 was the "official" name of Iran in the Western world, but Persian people inside their country since the Sassanid period (226–651 A.D.) have called it "Iran" meaning "the land of Aryans".

I think it is clear enough for someone who wants to know the truth. ( (talk) 12:50, 20 April 2008 (UTC))Reply[reply]

From the same article: "In 1959 Mohammad Reza Shah announced that both "Persia" and "Iran" could officially be used interchangeably. Now both terms are common; "Persia" mostly for historical and cultural texts, "Iran" mostly for political texts." As I said, "Persian Iranian" is a tautology. I am done here. Gandalf61 (talk) 12:59, 20 April 2008 (UTC)Reply[reply]

What you mention is not in contrast with what I asked. I just added Iranian in brackets after Persian to clear that they are the same. Using the term Persia for culture and history, and term Iran for political issues is just a political trick to separate Iran from its history and culture. Unfortunately this political intension is overcoming and supported in Wikipedia as well. ( (talk) 13:04, 20 April 2008 (UTC))Reply[reply]

I don't think it has much to do with politics. "Persia" has been the correct English-language term for hundreds of years. When residents of the region indicated that they preferred to be referred to as Iranians, English terminology was changed for them, but the old term "Persian" was kept for historical references. It's impossible to go back and change hundreds of years of pre-existing usage, so, for better or worse, the old usage was kept. Klausness (talk) 13:18, 20 April 2008 (UTC)Reply[reply]

You are completely right, but we are talking about Wikipedia, which is a quite new source and could be written based on new world terminology. I still believe that the best compromise is adding Iranian in brackets after Persian. As we can see in this discussion, many people does not know that they are referring to the same part of the world ( someone told Chwarizmi was living in the time of Persian Empire!!). The goal if Wikipedia is increasing the knowledge of his readers, not following the political intentions, at least it is what I think.( (talk) 13:36, 20 April 2008 (UTC))Reply[reply]

I think you were the first person to mention the Person Empire. I would think that adding "(modern-day Iran)" after the word Persian might be a compromise. But I also think about the fact that the link to Persian already explains what the word means.
So far, you haven't proposed any sources that describe him as Iranian, and others have claimed that the bulk of sources simply call him Persian. The issue seems to be that although Persian and Iranian mean the same thing in another language, they are used in English to mean different things. At leas that's the impression I am getting from the comments above. — Carl (CBM · talk) 14:01, 20 April 2008 (UTC)Reply[reply]
Also, as noted above, he wasn't Persian in the sense of modern day Iran. He was from a region near the Aral Sea in what is today known as Uzbekistan. silly rabbit (talk) 14:21, 20 April 2008 (UTC)Reply[reply]

To Carl: If you go to the history of the article you will see that "Silly Rabbit" said:

"He lived in the Persian empire, and sources identify him as Persian. This happened over 1000 years before Iran as a nation existed."

As you see some people have wrong knowledge about these terms and they are not reading all the articles which are linked to in one article (I do not think even you have time do that). About the sources calling him Iranian, we have discussed above and I do not want to repeat them here again. As I said before it is not only about him, it is about the terms "Persia" and "Iran". Moreover at least Ruud mentioned one source which explicitly called him "Iranian", but anyhow even without this source, as we discussed above, he knew himself as "Iranian". But anyhow we are telling the same thing: adding "(modern-day Iran)" after the word Persian might be a compromise.( (talk) 14:26, 20 April 2008 (UTC))Reply[reply]

Except see my comment above. Modern day Iran is definitely geographically incorrect, and identifying him as Iranian is misleading. Perhaps we should call him Uzbek? silly rabbit (talk) 14:45, 20 April 2008 (UTC)Reply[reply]

To SR: I agree with you, he was not in Modern-day Iran. But he was Iranian, this is the point which I try to explain. ( (talk) 14:49, 20 April 2008 (UTC)).Reply[reply]

About calling him Uzbek, just read the discussion above with Carl about calling Hammurabi Iraqi. You will get your answer.( (talk) 16:09, 20 April 2008 (UTC))Reply[reply]

I have read the thread. I am merely correcting you that saying "modern-day Iran" is wrong. No sources assert this. I'm sorry. silly rabbit (talk) 17:19, 20 April 2008 (UTC)Reply[reply]
Thanks, actually Carl was the one who used "modern Iran", and I just repeated his sentences word by word. But I agree that the modern Iran is different from Iran which Khwārizmī was living in.( (talk) 17:28, 20 April 2008 (UTC))Reply[reply]
I was confused there. Are there contemporary English sources which describe people from where Khwārizmī lived as Iranian? — Carl (CBM · talk) 18:04, 20 April 2008 (UTC)Reply[reply]
As Ruud said above, there is at least one source which called him Iranian. I do not know about it, hopefully Ruud will explain more. If you go to this article you will see that the people inside "Persia" called themselves "Iranian" since 226 AD, and Khwārizmī was of course living after 226 AD, so he knew himself as Iranian.( (talk) 18:25, 20 April 2008 (UTC))Reply[reply]

Donald Knuth writing in the late 1960's refers to a "famous Persian textbook author, Abu Ja'far [etc], as a native of Khowârizm, " the small Soviet city of Khiva." Does anyone know where this place might be? Is this info still accurate? It clearly cannot be located in Iran. I have a cc of Gibbon's 3 volumes The Decline and Fall of the Roman Empire, where the Persians are described at length -- the Romans and Persians fought for 700 years and the Romans feared and loathed them [from Chapter XLVI (570-642 A.D.): "An experience of seven hundred years might convince the rival nations of the impossibility of maintaining their conquests beyond the fatal limits of the Tigris and Euphrates (Vol. 2 p. 759)]. My recollection is that Alexander as well as other "occidentals" suffered severely at the hand of the Persians, a people who dominated the entire lands of what we now think of as as Iraq and Iran. But then the "Arabs" conquered the entire area in their quest to spread Islam by the sword [e.g. Chapt LI. The Conquest of Persia, Syria, Egypt, Africa, and Spain, by the Arabs or Saracens ... (632-718 A.D.)" (Volume 3 page v)]. I shall have to research this in detail, since I read this stuff some 20 years ago. Bill Wvbailey (talk) 18:54, 20 April 2008 (UTC)Reply[reply]

See Khiva. silly rabbit (talk) 19:05, 20 April 2008 (UTC)Reply[reply]
Khiva is located in Uzbekistan now and was a part of Persia at the time of Khwārizmī. So as explained above as someone who was living in "Persia" he knew himself as Iranian.( (talk) 19:07, 20 April 2008 (UTC))Reply[reply]
I doubt any sources exist as to what, specifically, he regarded himself. Did his personal writings on the matter survive? We only have the statements of reliable sources to guide us, and many quite reliable sources describe him as Persian. Is there a comparable number saying he was Iranian? I don't know. But engaging in speculation is not going to get us anywhere. silly rabbit (talk) 19:21, 20 April 2008 (UTC)Reply[reply]

you are going again to the beginning of the discussion. If we accept he was Persian, and if we accept Persian called themselves "Iranian" at that time which he was living (and both are cited in Wikipedia), there is no doubt that he knew himself as Iranian. i do not know what is not clear about these two simple facts. ( (talk) 19:30, 20 April 2008 (UTC))Reply[reply]
If it's so uncontroversial, then why does it seem that most scholars say he was Persian instead of Iranian? We go with the sources here, and we don't accept on faith that he would have called himself Iranian. If I'm going back to the beginning of the discussion, it's probably because the discussion should have ended there. See WP:STICK. silly rabbit (talk) 19:29, 20 April 2008 (UTC)Reply[reply]
I suggest you to read again all discussed above, you might have missed some parts. I explained above that after 1935 Persians decided and asked the world to call then with the term "Iranian" which has been used since 226 AD by people lived inside Persia(Iran). ( (talk) 19:36, 20 April 2008 (UTC))Reply[reply]

I see nothing in Gibbon about "Iranians", just Persians. I do see "Irak" mentioned. In fact, reading Gibbon gives me the impression that the "Iranians", at that time, probably thought of themselves as of Persian descent but, as not being Zoroastrian, would have been pretty cautious about claiming Persian desent. Gibbon describes the extent of the Persian empire under the Sassanidan kings (Artaxerxes A.D. 226 to Yazdejerd III A.D. 632 and Adeser A.D. 628) as the following: "His [Artaxerxes'] kingdom, nearly equal in extent to modern Persia, was, on every side, bounded by the sea, or by great rivers; by the Euphrates, the Tigris, the Araxes, the Oxus, and the Indus, by the Caspian sea, and the Gulf of Persia. (Vol. I, p. 177 in the Modern Library Edition). As best as I can tell, this region would have included the location of Khiva. Then along came the Arabs, and in a space of 100 years they conquered the Persians first: "One hundred years after his flight from Mecca the arms and the reign of his successors extended from India to the Atlantic Ocean, over the various and distant provinces which many be comprised under the names of, I. Persia; II. Syria; III. Egypt; IC. Africa; and V. Spain." (p. 135) Gibbon describes in detail the defeat of the Persians and the destruction of Zoroastrianism: As the Persian king flees north from the Arabs, he finally encounters some help from the Chinese emperor Taitsong who had his western-most outposts in the vicinity of the northern reaches of the Persian empire ("His last garrisons of Cashgar and Knoten maintained a frequent intercourse with their neighbours of the Jaxartes and Oxus; a recent colony of Persians had introduced into China the astronomy of the Magi; and Taitsong might be alarmed by the rapid progress and dangerous vacinity of the Arabs" (Vol. III, p. 143). But the Persian king was betrayed by his servant and killed and his death ended the reign of the Sassanidan kings -- and their kingdom of Persia. The two daughters of Yezdegerd "married Hassan, the son of Ali, and Mohammed, the son of Aubeker..." (footnote 40, p. 143, loc cit). The point is: the Arabs stopped just short of the Chinese but subdued everyone south of the Jaxartes (cf Vol. III, p. 144) and this included Khiva. As to what al-K. would have called himself, one is left to wonder if -- to get ahead in his now Arabic-speaking world -- he would have admitted to being of Zoroastrian "Persian" stock; and maybe he wasn't -- clearly, the Arabs and Persians intermarried. Unless his lineage is well-defined and accurate, we cannot say whether or not he was all "Persian", part "Persian", part "Arab", part "Chinese" or "Tartar" or what....

Edward Gibbon (1737-1794), The Decline and Fall of the Roman Empire,, The Modern Library, New York. Volumes I, II, and III. No ISBN, no Library of Congress catalog muber. With emendations by Oliphant Smeaton.

Bill Wvbailey (talk) 22:02, 20 April 2008 (UTC)Reply[reply]

Thanks Bill for his effort to clarify this case. To make the discussion understandable for those who did not follow from the beginning, I summarize all the arguments and claims discussed above:

- Is there any sources that call Khwārizmī as Iranian? Yes, as Ruud said, at least in one source, Toomer called him Iranian.

- So why in the most of literatures he is called “Persian” and not “Iranian”, or generally why the terms Iran and Iranian are not used in the historical sources? Because the term Persia and Persian were “internationally” used till 1935 to call the land and people of Iran.

-What happened in 1935? Reza Shah, the ruler of Iran at that time asked the world to use the term “Iran” and “Iranian” instead of “Persia” and “Persian”.

- So is Iran a new country which did not exit before 1935 and different from Persia at 1934? No, the people who were living in the county which was referred as “Persia” in western sources were calling their country “Iran” from 226 AD. So "Iran" as a country existed from 226AD.

- Was the Persia at the time of Khwārizmī the same as Persia at 1934? No, the Persia (Iran) at 1934 was much smaller than the Persia at his time.

-So if he was from a part of Persia which is not located in Iran. So why should we call him Iranian? He was living in internationally called “Persia” but locally called Iran. So he was as living as an Iranian,

-Why Gibbon talks about Persians and not Iranians? Because he was living before 1935.

-Was he “Persian” or “Iranian”? He was both. He was an Iranian who was living in a country “internationally” known as Persia before 1935.

-Why should we change something which has been used for hundreds of years? Because hundreds years ago, the “Persian” was the internationally correct term for calling “Iranian”. As have seen in this discussion many even educated people do not know this fact. We do need to replace the term “Persian” by “Iranian”, but we can put “Iranian” in brackets after “Persian” to help people have a better understanding of these terms.

I hope I could cover all the discussed topics; however if I have missed something, please help me to complete it. I am a scientist and I start a discussion only to find the truth and to replace my belief with knowledge and science. I feel the winner of a discussion if I can find the truth after that, even if it is against what I believed before that. I thank those who are helping to discover the truth together. I am sorry if anybody got angry when their mistakes were corrected during the discussion, nobody felt as a winner by correcting their mistakes, so they do not need to feel that they have lost. ( (talk) 09:50, 21 April 2008 (UTC))Reply[reply]

Ruud says that one source describes Khwārizmī as being "of Iranian descent", which is different than calling him Iranian. If, as you say, most sources call Khwārizmī Persian, I think we should follow their lead. I don't think it really matters what announcements the Shah made. I also don't think it matters what Khwārizmī called himself, since English wasn't yet formed at the time and we are writing an English document here.
What would be helpful for me is a list of references that shows what terminology each one uses. Right now we have several people who say that most sources use "Persian" and we have Ruud's statement that one source says "of Iranian descent."— Carl (CBM · talk) 13:21, 21 April 2008 (UTC)Reply[reply]
Do you think it is wrong to use both Persian and Iranian for him?( (talk) 13:27, 21 April 2008 (UTC))Reply[reply]
Moreover I explained why he is called Persian and not Iranian in the old literatures. Another source which called him Iranian and published after 1935 is "The Muslim contribution to mathematics. London: Croom Helm, . ISBN 0-85664-464-1". To make it clear for you, just assume that you are living in 1934 (before 1935) and know nothing about him, and someone tells you that he was Persian, what could it mean to you?( (talk) 13:36, 21 April 2008 (UTC))Reply[reply]
My personal opinion isn't established yet - I'm not in any way an expert on this sort of historical terminology. I'm going to try to find some more references at the library. Can you give me the page number for the book you just cited? — Carl (CBM · talk) 13:41, 21 April 2008 (UTC)Reply[reply]
Unfortunately I do not have the book an cannot give you the exact page number, but will try to find it.( (talk) 13:43, 21 April 2008 (UTC))Reply[reply]

I just found something quite interesting. I had a look to Khwārizmī pages in Wikipedia in some different languages. In Persian Wikipedia it is written he was from "Iran", in Arabic Wikipedia it is written he was from "Iraq" in Chinese Wikipedia it is written he was an Iranian-Arabian mathematicians.( (talk) 14:46, 21 April 2008 (UTC))Reply[reply]


I looked at Farsi and al-Khwārizmī and got an idea: the trick will be to avoid the words "Iran" and "Iranian" and just say al-K. was born in the region known historically as "Persia": " Persian (local names: فارسی [fɒrˈsi], Fārsi or پارسی [pɒrˈsi], Pārsi; see Nomenclature) is an Indo-European language spoken in Iran (Persia), Afghanistan, Tajikistan, Uzbekistan, and the Persian Gulf states. It is derived from the language of the ancient Persian people."

We could expland the sentence to say that "Most evidence (in particular his name) points to al-Khwārizmī's birthplace as Khiva, a city located in what is known historically as ancient Persia, now as modern Uzbekistan." We could just skip his genetics (was he Arab? "Iranian/Persian/Farsi-speaking"? Of mixed blood?). BillWvbailey (talk) 14:59, 21 April 2008 (UTC)Reply[reply]

I found another interesting source. As we know he was working in House of wisdom. I found something about House of Wisdom:

as the place where poetic accounts of Iranian history, warfare, and romance were transcribed and preserved…” He continues: “We have no reason to doubt that in the early ‘Abbasid administration it retained this function since its adoption was effected by individuals who were carriers of Sasanian culture and under the mandates of a caliphal policy to project Sasanian imperial ideology. Its function, in other words, was to transcribe and preserve books on Iranian national history, warfare, and romance.” (Dimitri Gutas, Greek Thought, Arabic Culture: The Graeco-Arabic Translation Movement in Baghdad and Early ‘Abbasid Society (2nd - 4th/8th - 10th centuries. London: Routledge, 1998.)

( (talk) 15:09, 21 April 2008 (UTC))Reply[reply]

I think the most important thing we're missing here is that al-Khwārizmī's nationality is completely immaterial to the topic at hand. If this topic is addressed in detail, it should be at Muhammad ibn Mūsā al-Khwārizmī (which currently labels him a Persian and discusses his hometown). In this article, we could omit it entirely, or give a brief label like "Persian" followed by "see Muhammad ibn Mūsā al-Khwārizmī for more discussion of al-Khwārizmī's nationality." This article is not the place to expound upon it, because it's not relevant. Dcoetzee 17:09, 22 April 2008 (UTC)Reply[reply]
This does not solve the problem. We have two choices: 1. continue and conclude this discussion and apply the results to his page as well or 2. Stop here and move this discussion to his page which somehow means start from beginning. ( (talk) 19:12, 22 April 2008 (UTC))Reply[reply]
There is a certain purity in Dcoetzee's advice: "When in doubt, cut it (references to his nationality and religion) out." If you want to keep the discussion going, just copy this and paste it to the al-K. talk page. Or, cut this whole discussion from this talk page, leave a note that it has been moved, and copy it to the al-K. talk page. Bill Wvbailey (talk) 20:20, 22 April 2008 (UTC)Reply[reply]
If everyone is happy with moving this discussion to Khwarizmi page, so lets do it. Maybe Carl who protected the topic and opened the discussion in this page would be the best person to move this discussion. ( (talk) 20:52, 22 April 2008 (UTC))Reply[reply]

Proposal to remove "Persian"

There seems to be some support in the previous section to just remove the "persian" part altogether, as I did in this edit. Then the article Muhammad ibn Mūsā al-Khwārizmī will have to be discussed separately. What do people think about that? — Carl (CBM · talk) 10:30, 28 April 2008 (UTC)Reply[reply]

Nobody doubted about the fact that he was "Persian" which is also mentioned in Muhammad ibn Mūsā al-Khwārizmī. All the discussion was about adding Iranian after Persian. Lets leave it as it is, and change it only in case of any changes in Muhammad ibn Mūsā al-Khwārizmī.( (talk) 15:01, 28 April 2008 (UTC))Reply[reply]


I see that someone quicky reverted Carl's edit so that the article says "Persian", again. I'd suggest that if we agree to omit the word Persian that the dates of al-K.'s life be added in parentheses (slightly unnecessary, as the approx date of his book is given, but the dates put a frame around the history), together with a footnote noting this dispute and to refer the reader to al-K.'s biography. In regards to Iranians, I still can't find any deep historical reference to "Iranians". I've been reading Herodotus (c. 484 - c.425 B.C.). He begins his history with a detailed discussion of the Persian empire (the name given to the map on p. 317), in particular the war between Cyrus and Croesus. The map of the Persian empire at that time (its capital was Babylon -- located on the Euphrates river, about 50 miles due south of Baghdad). The Persian empire ranged in the west from ancient Trace (modern eastern Greece, including Athens) into ancient Libya (modern north Africa in particular modern Egypt), north to the Caucucus Mountains, north well beyond the Oxus and just north of the Jaxartes to include Taskent, east to the Indus River, and south into the Syrian Desert (modern Saudia Arabia). On the map Khiva is located toward the northern boundary (as noted in the history above). Teheran is shown in the region called Media, which also includes Parthia. So, the extent of Persia at the time of the fall of the Roman and Persian empires (at the hand of the Moslems) was farther to the east (i.e. modern Egypt and Greece -- the Romans had pushed them out of modern Turkey, Egypt, and ancient Syria/Palestine), but about the same in the east and north, i.e. clear over to India. I'd recommend the following to anyone curious about the ancient Persians at the time of the ancient Greeks:

Robert Maynard Hutchings, Editor in Chief, 1952, Great Books of the Western World: Volume 6: The History of Herodotus, and Thucydides: The History of the Peloponnesian War, Encyclopedia Britannica, Inc. Chicago. The translation of Herodotus is by George Rawlinson. Maps begin on p. 317, and an index begins on p. 325. The reading is surprisingly easy excepting the use of formal English in the conversations.

Bill Wvbailey (talk) 17:48, 28 April 2008 (UTC)Reply[reply]

I am ok with the suggestion to leave the article as it is (same as it was before) unless a change is made to the main biography article first. — Carl (CBM · talk) 18:33, 28 April 2008 (UTC)Reply[reply]

I agree. Bill Wvbailey (talk) 16:10, 29 April 2008 (UTC)Reply[reply]

Get rid of the flaky nationalism for heaven's sake. Its misguided, misleading and irrelevant to the subject at hand. As for that asinine parenthesizing: Since when does uninformed usage prescribe that we keep readers ignorant? By the same measure we would need to add "(ugga ugga era)" after every instance of "Paleolithic," and delete every article for every topic that didn't exist before 1935. -- Fullstop (talk) 11:03, 4 May 2008 (UTC)Reply[reply]

Discovery v. Invention?

Is algorithm a discovery or invention? Anwar (talk) 11:47, 10 June 2008 (UTC)Reply[reply]

In general the same question can be asked of "Mathematics": Is "doing mathematics" an act of discovery or invention? If you are a Platonist, doing mathematics is discovery: Just as a new land was discovered by Columbus, and dinosaur bones were discovered in the Gobi desert by Roy Chapman Andrews, the mathematics (new land, dinosaur bones) exists (always before and after and) independent of the mind and is merely discovered (e.g. by cleverness, persistence and luck). On the other hand (I believe that) a Formalist would say that doing mathematics is act of invention -- a particular mathematics is a game played with rules set up in advance (like Monopoly, Risk, chess, or checkers) so there can be many mathematics (as there are many games). An Intuitionist (I believe) would agree with the Formalist in this but go maybe a step further -- an act of "doing mathematics" is not only a human invention but also reflects something in particular about the construction (i.e. wiring) of the human mind. Maybe someone else has a take on this. This matter has come up in other venues. Bill Wvbailey (talk) 14:17, 13 June 2008 (UTC)Reply[reply]
Yes, this is a long-debated philosophical question. In short, it's not appropriate to discuss in this article. Dcoetzee 19:51, 13 June 2008 (UTC)Reply[reply]

Data structure query

do u have data structure test papers. if yes, pls sent the link. —Preceding unsigned comment added by (talk) 06:36, 4 August 2008 (UTC)Reply[reply]