Wikipedia:Articles for deletion/Commitment ordering

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 keep. The consensus is that the article should not be deleted. The quality of the writing and the style are issues which can be addresses through normal editing. HJ Mitchell | Penny for your thoughts? 23:09, 23 November 2011 (UTC)[reply]

Commitment ordering

(Find sources: Google (books · news · scholar · free images · WP refs) · FENS · JSTOR · TWL)

Wikipedia is not a textbook, or scientific journal, this is one big scientific article, not an encyclopedia article. Wikipedia is not a means of promotion, author is using wikipedia to promote himself and his research, Wikipedia:Articles for deletion/Yoav Raz, Wikipedia:Articles for deletion/History of commitment ordering. Wikipedia is not a publisher of original thought, while sections do reproduce Raz's research published elsewhere other sections are original synthesis to further the authors position on the subject, creating a personal essay. Wikipedia is not a soapbox, article is based primarily on Raz's publication in a way that is clearly to advocate his position on the subject. duffbeerforme (talk) 07:41, 15 November 2011 (UTC)[reply]

Delete all the references that actually mention this topic seem to be by Yoav Raz. Stuartyeates (talk) 08:53, 15 November 2011 (UTC)[reply]

Incorrect. See the following examples:

1. the following quotations on CO appears in a 2009 book: Philip A. Bernstein, Eric Newcomer (2009): Principles of Transaction Processing, 2nd Edition, Morgan Kaufmann (Elsevier), June 2009, ISBN 978-1-55860-623-4

Quotations:

"Not all concurrency control algorithms use locks... Three other techniques are timestamp ordering, serialization graph testing, and commit ordering. Timestamp ordering assigns each transaction a timestamp and ensures that conflicting operations execute in timestamp order. Serialization graph testing tracks conflicts and ensures that the serialization graph is acyclic. Commit ordering ensures that conflicting operations are consistent with the relative order in which their transactions commit, which can enable interoperability of systems using different concurrency control mechanisms." (quotation from page 145)
"Commit ordering is presented in Raz (1992)." (page 360)

Bold fonts in source

Phil Bernstein is a known researcher and authority in database concurrency control. --Comps (talk) 23:22, 19 November 2011 (UTC)[reply]

2. The textbook on transaction concurrency control: Gerhard Weikum, Gottfried Vossen (2001): Transactional Information Systems, Elsevier, ISBN 1-55860-508-8 , has two section dedicated to and named commitment ordering: One starts in page 102 and the other in page 700. The book references Yoav Raz's CO articles for the sections. Both authors are known researchers in this area.--Comps (talk) 23:47, 19 November 2011 (UTC)[reply]

Commitment ordering is a central element today in Concurrency control theory. It has enjoys increasing utilization and interest (see History of commitment ordering). Will be a mistake to delete, to my opinion, independently of all the reasons that you guess caused writing it. --Comps (talk) 11:29, 16 November 2011 (UTC)[reply]

Note: This debate has been included in the list of Computing-related deletion discussions. • Gene93k (talk) 15:20, 15 November 2011 (UTC)[reply]

Comment presumably the paper and the conference proceedings were peer-reviewed, so that there are actually experts in the field who find this notable enough to publish? Cusop Dingle (talk) 18:32, 15 November 2011 (UTC)[reply]
- Was that a direct request for input from experts, or a rhetorical argument? So far as I know, Wikipedia does not automatically assume that subjects covered in peer-reviewed papers are notable under the GNG. Chris Cunningham (user:thumperward) (talk) 14:56, 16 November 2011 (UTC)[reply]

As noted in the deletion discussion of History of commitment ordering all referenced CO articles have been peer-reviewed by prestigious program committees and reviewers. Every expert looking into it will endorse this claim. I agree that experts in database concurrency control need to decide about CO's importance, as well as the validity of this article's accuracy and proper references. Not by those not experts. The notability of CO is also argued in the History of CO deletion discussion, and solid reasons are given there for this. The Wikipedia article History of CO itself clearly gives these arguments for notability. Quality of article in terms of appearance and presentation is a separate issue. I'm sure it is not perfect, even after years of iterations and rewriting, but I'm also sure it can continue improving. --Comps (talk) 00:12, 17 November 2011 (UTC) I also believe it is very reasonable in term of quality. --Comps (talk) 00:17, 17 November 2011 (UTC)[reply]

The WP:GNG requires "significant coverage in reliable sources that are independent of the subject". Peer-reviewed articles and academic conference presentations on a subject would usually be considered sigificant coverage, and peer-reviewed journals are nornally reliable sources. The independence wpuld follow from the fact that the journals are peer-reviewed and edited independently. So it would appear that this topic passes the GNG threshold. Cusop Dingle (talk) 07:34, 18 November 2011 (UTC)[reply]

Abstain - Yoav Raz appears to have enlisted a bunch of grad students as Wikipedia editors. This creates WP:COI and WP:UNDUE issues which make it difficult to determine best how to proceed. This will not be resolved quickly though AfD. --Kvng (talk) 16:48, 16 November 2011 (UTC)[reply]

I see no problem if judged logically and properly. "Yoav Raz appears to have enlisted a bunch of grad students as Wikipedia editors." is unsubstantiated an irrelevant to the discussion, as long as facts are correctly presented, , with no phylibustering (which I have not seen here), and logical deductions are made. --Comps (talk) 00:12, 17 November 2011 (UTC)[reply]

Weak keep and stubify. There seem to be enough sources on the subject by people other than Raz to make this worthy as an encyclopedia topic. However, there are severe WP:COI and WP:NPOV problems; for instance, several of the non-Raz references I looked through said that CO is the same thing as strong recoverability, while our article (without adequate sources) argues that it isn't. I don't think most of the current articles text is usable, and if the problem editors can't be persuaded to let the article be edited down to a more encyclopedic state then it may have to be deleted outright — it isn't so important a topic that its absence would be a big problem. —David Eppstein (talk) 20:58, 16 November 2011 (UTC)[reply]

Strong recoverability has a definition identical to that of CO (in History of CO article). Please go to the strong recoverability articles (in History's refs) for definition. However the definition has not been utilized as in the CO invention (i.e., strong recoverability has no methods and algorithms which CO has) and thus has nothing to do with the CO patents. Otherwise the patents would have been disputed long ago. The strong recoverability articles are explicitly referenced in the CO patents, exactly to have this scrutinized by the patents' examiners. Independently of this CO preceded strong recoverability. --Comps (talk) 21:34, 16 November 2011 (UTC)[reply]

As said above all this needs to be determined by experts, not by not-experts. Specifically also the importance of CO that you (not an expert; please tell if otherwise) doubt. I bet that every expert will endorse CO's importance. Phil Bernstein, who is a known expert in the field, already did in the book Bernstein and Newcomer 2009 (referenced), listing CO as one of the four major database concurrency control methods, and describing its unique ability to allow different CC mechanisms' interoperability. Also any possible bias assumed here in the article needs to be detected by experts, if exists. Re stubifying, I see no reason for this, if CO is indeed important as claimed here, and if the article is accurate. --Comps (talk) 00:56, 17 November 2011 (UTC)[reply]

all this needs to be determined by experts - I suggest that you may be interested in brushing up on the concept of Wikipedia:Notability, on which this discussion is based. Stuartyeates (talk) 01:13, 17 November 2011 (UTC)[reply]

Ok. I see. Thanks. I'll concentrate on this. --Comps (talk) 11:42, 17 November 2011 (UTC)[reply]

See links below to Google scholar (100+) and to Google patents (250+) for citations. This establishes notability sufficiently, I believe. --Comps (talk) 08:19, 19 November 2011 (UTC)[reply]

Weak keep and rewrite As per David Eppstein. There are some independent references to this work, but all I've found are fairly trivial (i.e. a single sentence mentioning this exists.) —Ruud 13:27, 17 November 2011 (UTC)[reply]

I do not know why you call such citation "trivial". A citation is an acknowledgement in the validity and non-triviality of a publication (otherwise ignored). The CO articles have been cited in tens of academic articles, without any dispute; the CO patents have been cited in hundreds patents (see Google Scolar and Google Patents by Pat No.)

Google scolar for "Commitment ordering"

Google scolar for "Commit ordering" (many different articles for the alias)

Google patents for Patent 5701480 (One of three CO patents) --Comps (talk) 23:22, 19 November 2011 (UTC).[reply]

Keep and rewrite if necessary. Also someone to think about copyvio/OTRS. We should not reject material simply because it originates from a (or even the primary) researcher in the area, but we do need to be careful about areas where the information is not part of the generally accepted wisdom of the field, and of course be aware of WP:UNDUE. Rich Farmbrough, 17:04, 17 November 2011 (UTC).

Though I'm a big believer in CO, I also believe the article is neutral and fact-only based. However, any scrutiny is more than welcome. --Comps (talk) 19:45, 23 November 2011 (UTC)[reply]

Delete - incomprehensible to the average reader. . . Mean as custard (talk) 22:00, 18 November 2011 (UTC)[reply]

The average reader cannot understand also many other Wikipedia articles like Calculus of variations and Standard model (examples I already have given) beyond the trivial intro. He may be able to understand after studying the needed background. In this regard Commitment ordering is simpler than the examples since relying on smaller bodies of Mathematics and context. --Comps (talk) 07:30, 19 November 2011 (UTC)[reply]

Keep. This is an important article, on a central subject in concurrency control of databases and transactions. It is a fact-based article, neither an opinion or position in any form, nor a soapbox, contrary to the lead paragraph above. (to be elaborated) --Comps (talk) 17:12, 20 November 2011 (UTC)[reply]

Comment There are several problems with this article. The precise title "Commitment ordering" seems only to be used by Raz. I'm not sure whether he intends to use it as a general term or as a particular algorithm (mechanism?) for enforcing the "CO property". "Commit ordering" is used more often and by a much larger group of people. The name "dynamic atomicity" is also used more often and User:Comps even admits that it is equivalent and precedes "Commitment ordering". I think the title should be Dynamic atomicity.

Regarding the name Commitment ordering:

"Commitment ordering" is the full original name used by Dr. Raz. The name seems to be analogous to the name of another concurrency control method: "timestamp ordering" (In both CO and TO "objects" are ordered in transaction precedence order). As noted in the article this is the name of both the history (schedule) property and the methods, algorithms to produce such histories, which are patented. Like the name interchange "atomic commitment" <-> "atomic commit", also here we see the synonym "commit order", where people often prefer the shorter form.

Regarding "Dynamic atomicity": Dynamic atomicity was defined as a history property only (no generic algorithms). It is used primarily in the context of programming languages. It originated at the MIT and was compared to Commitment ordering thoroughly in the deleted article History of commitment ordering, now in user space for rewrite ([1]). Database text books and articles have used the name "Commitment ordering" which better describe the property in the database jargon than the name Dynamic atomicity, a jargon unique to the MIT group, typically used for Abstract data types, not databases (atomicity has a different meaning in databases and the MIT jargon). --Comps (talk) 19:17, 23 November 2011 (UTC)[reply]

It is not easy writing on a specialized topic for a general encyclopedia. WP:JARGON suggests that we should write at one level below that at which the subject is studied. I think this would mean an undergraduate computer science student. Even though User:Comps has put a lot of effort into this article I think it clearly misses that standard.

There is also the problem that the article over-emphasizes the work of Raz. I'm not saying that he should be ignored, but there are other people publishing in this field too.

I'm really not sure what to suggest should be done. User:Comps is probably the only person in this discussion who is an expert in this area, but he seems to have produced an article strongly biased towards Raz's work. On the other hand, he is of course a volunteer like all of us.

It is not easy writing on a specialized topic for a general encyclopedia. Dingo1729 (talk) 05:56, 23 November 2011 (UTC)[reply]

With respect to the comment: User:Comps is probably the only person in this discussion who is an expert in this area. It might be more true to say that User:Comps is the only person to have put themselves forward as an expert in this area. I have a PhD in Computer Science [2] and have taught concurrency at a university level; whether that counts as being an expert probably depends on where you draw your lines. I hold that expert status is irrelevant here, since this discussion is about balance and WP:RS. Stuartyeates (talk) 06:25, 23 November 2011 (UTC)[reply]

WP:RS is put to rest by the quotations above, in the the conversation of the first reply (Delete): The latest most notable text-books on transactions and transaction concurrency control respectively describe Commitment ordering! --Comps (talk) 19:31, 23 November 2011 (UTC)[reply]

My apologies, I accept the correction. Dingo1729 (talk) 15:33, 23 November 2011 (UTC)[reply]

The specific subject is Database concurrency control described in Concurrency control and covered in the text-books referenced there and in Serializability.

"User:Comps is the only person to have put themselves forward as an expert in this area." I do not remember if I explicitly said I was an expert, but I am willing to do it now. I wrote most text in the related articles here (over a period > 5 years), and will be happy to put them to any professional scrutiny. I'm glad Wikipedia has a good coverage of the area, and Commitment ordering is included as an important piece, now and even more in the future. It is pity History of commitment ordering was deleted, since it covers interesting issues like those raised here above with Dynamic atomicity. --Comps (talk) 19:17, 23 November 2011 (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.