- The following discussion is an archived debate of the proposed deletion of the article below. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.
The result was delete. Stifle (talk) 15:35, 12 September 2008 (UTC)[reply]
- Wizard sort (edit | talk | history | protect | delete | links | watch | logs | views) (delete) – (View log)
Unable to find any reliable sources to establish notability of this algorithm. Maralia (talk) 02:38, 8 September 2008 (UTC)[reply]
- Delete. Essentially, it looks to me like the author is proposing a (very naive and really bad) version of a Hash Sort--something that might warrant its own article. Anyway, there's a reason you've not found any evidence of notability for this algorithm: it's really bad. The algorithm does not even run in O(n^2) time, where n is the number of elements in the list, but in O(k), where k is the largest element in the list. That means if you have a list with a two elements, say, 1 and MAX_INT, that means that it would take on the order of MAX_INT operations to sort the list. Furthermore, as described this algorithm has no way of dealing with multiple entries in the list. This is very, very bad.
- Furthermore, lots of the stuff he has here is just wrong. To attempt to represent even a small double as a long could conceivably create a very, very large number--meaning that the Wizard sort would take a very long time to finish.
- This article is pretty much entirely OR, and not even good OR. It badly needs to go. However, I am going to temporarily copy this to my namespace--I know some people who might be interested in seeing this. TallNapoleon (talk) 06:34, 8 September 2008 (UTC)[reply]
- agree. I started smiling at the unique integer part. NVO (talk) 08:01, 8 September 2008 (UTC)[reply]
- Delete. I haven't been able to find any references either - Google gives nothing but false positives, and there's nothing on the arXiv - and so I think we have to call this OR. Scog (talk) 06:52, 8 September 2008 (UTC)[reply]
As mentioned in the first paragraph: this method is a synthesis of different concepts (as most of the sorting algorithms, by the way). It is using Hashing technique with F(x)=x, however this method involves more steps then just hashing and it is not listed on Sorting algorithm page, which states that best known practical case is a Radix Sort, which is slower then Wizard sort.
I think author of the remark also missed Wizard Sort (non-unique integers) part, which explains how Wizard sort deals with multiple entries.
Also, please describe how in a world this algorithm will be O(n^2) as stated and not O(n+j), where j is the largest element in a set, as stated in the article?
It is correct, for small array with large values, Wizard Sort will waste a lot of memory and time, but issue of speed in sorting comes up, usually, only in large arrays.
As for the last remark: it is a blatant original research, but that does not reduce its validity.
And my last point: You can say anything you want about it, but Benchmarks do not lie!
djwizard Yevgeniy Dukhovny, September 08, 2008. —Preceding undated comment was added at 19:23, 8 September 2008 (UTC) — djwizard (talk • contribs) has made few or no other edits outside this topic. [reply]
- Response WP:NOR says "Wikipedia does not publish original thought: all material in Wikipedia must be attributable to a reliable, published source". That's policy. This algorithm is original thought, as you've stated yourself. —/Mendaliv/2¢/Δ's/ 19:35, 8 September 2008 (UTC)[reply]
- Response Source code is a source. By not publishing original thought Wikipedia means a subjective opinion. In this case it is a fact, which you can verify by running code references in the article. 2+2=4 is a fact (in most cases), just because someone stated it, does not mean it is an original thought. Yevgeniy Dukhovny --Djwizard (talk) 20:49, 8 September 2008 (UTC)[reply]
- Do Not Delete May be I am missing something, but arguments stated in TallNapoleon's message are simply incorrect, and I thought I showed how are they are incorrect. It is not that I agree or disagree with them: but they are just stating something about the article (like Wizard Sort not dealing with mulitple values) which are simply not true. Please someone explain this to me.--Djwizard (talk) 22:36, 8 September 2008 (UTC)[reply]
- Even if TallNapoleon is completely wrong and this is the best sorting algorithm ever, independent reliable sources need to be provided to establish notability. You needn't defend the algorithm's mechanics here; a deletion decision would not be based on a judgement of its value, but rather its notability. Maralia (talk) 22:55, 8 September 2008 (UTC)[reply]
- To all, except for Maralia (who has a valid point, with which I disagree though), who are saying delete:
You might throw theory at it as much as you can, but I provided a Java code which you can run on you computers and use in your code (and some of you probably will use it in your code). If you run source code, it, almost certainly, will be faster then ANY sort you have available.
Show me a code which runs faster then this!
--Djwizard (talk) 23:32, 8 September 2008 (UTC)[reply]
- Response For any one of us to run the algorithm and benchmark it ourselves and then put information in the article based on those benchmarks would constitute original research; "Wikipedia is not the place to publish your own opinions, experiences, or arguments". The algorithm doesn't matter for the sake of this debate, rather it's the failure of WP:RS; "Articles should be based on reliable, third-party, published sources with a reputation for fact-checking and accuracy". —/Mendaliv/2¢/Δ's/ 23:44, 8 September 2008 (UTC)[reply]
- Response: as far as I can tell this is a rather clunky implementation of Counting sort. I'm afraid benchmarks are not considered "proof" when it comes to developing algorithms, because they have a tendency to miss pathological cases. In this case, one such case would be a small list containing MAX_LONG, which would cause your sort to literally run most modern computers out of memory. It is true that for large numbers of values with a small range, counting sort will sort more quickly than any other sorting algorithm. But a single large element breaks it horribly. DJ, it's nice to see that you've taken an interest in computer science, but I highly recommend that you take a university level algorithms course. As for the policy on original research, for an algorithm to be notable it can't just be something you came up with. Rather, it has to be something that has been recognized by the computer science community as a whole, by being frequently used in applications, published in a scholarly journal, or something of the sort. Oh, and if Wizard Sort does deal with multiple objects, you have my apologies for missing that. It appears to do so in a very strange manner, when frankly the best way to do it is to have your Wizard array store the number of times each element appears--as, indeed, counting sort does. Best of luck, but still delete. TallNapoleon (talk) 23:50, 8 September 2008 (UTC)[reply]
- Delete. Does not appear to be notable at this time. If this algorithm is the best thing since sliced bread, get it published in the Communications of the ACM or a similar journal, and then it may become notable enough to have a Wikipedia article. --Itub (talk) 09:10, 9 September 2008 (UTC)[reply]
- Note: This debate has been included in the list of Computing-related deletion discussions. -- Fabrictramp | talk to me 17:52, 9 September 2008 (UTC)[reply]
- Delete self admitted original research. It may one day, after peer review and independent coverage, become notable but it is nowhere near now.
- Delete - as original research. Whether it's good, bad, indifferent, or the best thing since sliced bread, it's still clearly original research which is contrary to policy. -- Whpq (talk) 16:58, 10 September 2008 (UTC)[reply]
- The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.