Jump to content

Wikipedia:Articles for deletion/Fastflow (computing)

Add links
From Wikipedia, the free encyclopedia
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. I have ignored votes from brandnew users. The issue is the sourcing. The only halfway useful sources are the academic ones written by the authors of the software. This does not cross the notability bar and N/GNG require independant sources. Sorry but this isn't suitable for inclusion just yet Spartaz Humbug! 03:54, 10 September 2010 (UTC)[reply]

Fastflow (computing)

[edit]
Fastflow (computing) (edit | talk | history | protect | delete | links | watch | logs | views) – (View log • AfD statistics)
(Find sources: Google (books · news · scholar · free images · WP refs) · FENS · JSTOR · TWL)

Apparently non-notable software project. The references listed and that I have found that I have checked do not mention the software, mention it in passing or do not meet the bar for academic sources. --Nuujinn (talk) 23:22, 2 September 2010 (UTC)[reply]

  • Keep I simply am mystified by Nuujinn's comment above. First of all, this is a field I understand reasonably well. Does Nuujinn claim to understand this field? Unless he/she does, a statement that software is "apparently non-notable" is vacuous. (This is a serious problem, IMO, in many of the software AfD nominations I have seen.) Also, on the Fastflow website, there are the following references (which should be incorporated into the WP article) that more than meet the notability requirements AFAIK:
Papers
[ART10] Marco Aldinucci, Salvatore Ruggieri, and Massimo Torquati. Porting Decision Tree Algorithms to Multicore using FastFlow, in: Proc. of European Conference in Machine Learning and Knowledge Discovery in Databases (ECML PKDD), volume 6321 of LNCS, pages 7–23, Barcelona, Spain, Sept. 2010. Springer. (16% acceptance rate)bib
  • Comment, document exists as a tech report from Aldinucci's uni (local tech reports are not generally reviewed--I do not know how other unis handle these, but where I work anyone can write up a tech report and we'll put it up. It's a quick way of getting the data out into the world, establishing what you have done, without having to wait for acceptance and publication via peer review.). The conference will occur starting september 20th, see the conference web page, but this has apparently not yet been published. If anyone can find a copy of the proceeding, please advise. --Nuujinn (talk) 14:47, 6 September 2010 (UTC)[reply]
[ABL10] Marco Aldinucci, Andrea Bracciali, Pietro Lio'. Formal Synthetic Immunology, Ercim News 82:40–41, July 2010. bib
[ABL10] Marco Aldinucci, Andrea Bracciali, Pietro Lio', Anil Sorathiya, and Massimo Torquati. StochKit-FF: Efficient Systems Biology on Multicore Architectures, in: Proc. of the 1st Workshop on High Performance Bioinformatics and Biomedicine (HiBB, in conjunction with Euro-Par 2010), LNCS, Ischia, Italy, Sept. 2010. Springer. To appear. bib
[ADK10] Marco Aldinucci, Marco Danelutto, Peter Kilpatrick, Massimiliano Meneghin, and Massimo Torquati. Accelerating sequential programs using FastFlow and self-offloading, Università di Pisa, Dipartimento di Informatica, Italy, number TR-10-03, February 2010. bib
[AMT09] Marco Aldinucci, Massimiliano Meneghin, and Massimo Torquati. Efficient Smith-Waterman on multi-core with FastFlow, in: Proc. of Intl. Euromicro PDP 2010: Parallel Distributed and network-based Processing. IEEE. Feb. 2010.bib
[ATM09] Marco Aldinucci, Massimo Torquati, and Massimiliano Meneghin. FastFlow: Efficient Parallel Streaming Applications on Multi-core, Università di Pisa, Dipartimento di Informatica, Italy, number TR-09-12, September 2009. bib
[ADM09] Marco Aldinucci, Marco Danelutto, Massimiliano Meneghin, Peter Kilpatrick, and Massimo Torquati. Efficient streaming applications on multi-core with FastFlow: the biosequence alignment test-bed, in: Proc. of Intl. Parallel Computing (PARCO), September 2009. bib
  • Comment, Science direct is an Elsevier press service, whereas the paper is published by IOS press (those two companies are both from the Netherlands and they are in competition). This paper is in a volume that collects the best papers from the conference Parco 2009 (whereas the conference itself has not electronic edition but only a printed book). This paper has been reviewed before the conference and accepted, then selected at the conference, then reviewed again and accepted by other reviewers. IOS press sells it for 20$. I have a copy. I'm involved in fastflow. Regards. Aldinuc (talk) 17:40, 6 September 2010 (UTC)[reply]
Talks
FastFlow: a pattern-based programming framework for multicores. Dagstuhl seminar 10191, Schloss Dagstuhl, Germany. May 2010. Invited. Slides available on-demand.
  • Comment, at my uni we get guest lectures 1-2 a week, sometimes from folks who asked to come to take a look at what we're during, sometimes from friends of the person giving the talk. Such do not confer notability. --Nuujinn (talk) 14:47, 6 September 2010 (UTC)[reply]
FastFlow: why we need yet another programming framework. Guest lecture, Computer science Dept. Queen’s University Belfast, UK. March 2010. Invited. Slides available on-demand.
  • Comment, at my uni we get guest lectures 1-2 a week, sometimes from folks who asked to come to take a look at what we're during, sometimes from friends of the person giving the talk. Such do not confer notability. --Nuujinn (talk) 14:47, 6 September 2010 (UTC)[reply]
Efficient Smith-Waterman on multi-core with FastFlow. IEEE PDP 2010: Parallel Distributed and network-based Processing, Pisa, Italy. February 2010.
Efficient streaming applications on multi-core with FastFlow: the biosequence alignment test-bed. ParCo 2009, Lyon, France. September 2009.
  • Comment, could you prove, or at least, argument about this? Publishing a non-original material is a serious infringement of copyright; claiming an author has published duplicate material is a serious attack to the good name of the authors. I kindly ask you to either prove it or delete the previous comment. (I am involved in the project). Regards. Aldinuc (talk) 16:30, 6 September 2010 (UTC)[reply]
  • Comment, sure. It appears to me that the papers published or to be published in the conferences listed above have the same titles as these talks. I'm assuming that the papers were written, submitted to the conference, accepted, and both presented and published as part of the proceedings of the conference. If this is correct, I would suggest that the publication of the proceedings and the presentation of the paper at the conference are together a single reference, not two. --Nuujinn (talk) 17:39, 6 September 2010 (UTC)[reply]
  • Comment, probably a misunderstanding. Of course the talks have same title of the papers. In the FastFlow page they are linked to PDF of the paper and to PDF of the slides. This is the only reason because they appear duplicated. If the suggestion is: put the PDF of slides in the same raw of the paper, well, this is good idea for improving the fastflow website. Thanks. (otherwise I did not really understand ...). I'm involved in the FastFlow project. Regards. Aldinuc (talk) 17:50, 6 September 2010 (UTC)[reply]
PDF versions of all the papers and slides are on the Fastflow website. (I haven't bothered to link to them, but I hope that someone will.) —Preceding unsigned comment added by HowardBGolden (talk • contribs) 02:17, 3 September 2010 (UTC)[reply]
  • Comment. The above refs can be viewed in-context at the Fastflow web site, here. Besides the above user, is there any group or project of established users we can appeal to here on Wikipedia that has the in-depth knowledge of C++ that would be helpful in determining the notability of what its author is calling "Fastflow"? A few folks who could look more closely at those papers, for example, to see which are directly applicable to the technology, and render an opinion as to whether they constitute a sufficient basis for notability?  – OhioStandard (talk) 12:10, 3 September 2010 (UTC)[reply]
I'm striking the above !vote as an obvious sock or collaborator of one of the co-creators of this programming framework and the creator of this article about it under his account name Aldinuc (talk · contribs) There's no outing issue here, btw. User Aldinuc has self-identified as one of the creators of the framework. For the quacking I base this action on, take a look at the revision history for the user page belonging to Pomello (talk · contribs), which account was just created today, and whose only edit has been to !vote here. Also, Aldinuc, since your editing experience certainly didn't begin with your creation of your Aldinuc account, will you please disclose the other account name(s) or wiki location(s) where you've edited previously?  – OhioStandard (talk) 13:27, 3 September 2010 (UTC)[reply]
Dear Ohiostandard, I copy-pasted Aldinuc (talk · contribs) user description because I did not know how to make a brand-new one, and I would like to disclose that I'm C++ expert. Template usage in wikipedia is pretty obscure to me, that's it, you spotted me  :-) Then I discovered (because of inexperience) that aldinuc page includes personal information (such as skype address and the fact that he/she like guitar and I deleted these information from my user page). I would like to add that I was pretty undecided if copying his/her one or your one, then I discarded your one because it is much more complex and refer to obscure-to-me things such as sandbox, etc. Moreover I'm also Italian and there was a Italy tag there that I can copy. I did not created the account today, I've created my user page today, I've created my account before august for sure, don't remember exactly when, there should be a way to check if it somehow matters. By the way, I've also copied the source from this page, since I did not knew how to indent things, sign a paragraph, make a link and so on. I confess I'm a computer scientist, and this is technically a wiki, and I edit dozen of other wiki pages for other reasons. I hope this not a problem. -- Pomello (talk) 14:25, 3 September 2010 (UTC)[reply]
and, yes, he shown me how to edit the user page (because I've asked), and yes we cooperating since I'm using his software, and I think it is good piece of software. And yes, I found this software on the web and asked him to cooperate (starting from July). I did not expected that, due to this, I cannot write my own opinion. And I decided to do it today - as my own initiative - because I've seen the page is under deletion and you asked have opinions, well, IMO the page is worth, and I tried to argument on why I think it is worth. Anyway, I absolutely would not like to violate any rule. I'm really just a newbie in editing wikipedia that would like to give a technical opinion on a software that I know quite well because I'm using it. Best. Pomello (talk) 15:47, 3 September 2010 (UTC)[reply]
  • Comment. HowardBGolden, I've worked in IT at a major uni since the early 1980s, although that's not really relevant here. If you look at the article, please note that most of the references cited came out before the software. Also, I have read some of the papers posted on the project's home page (I assume the homepage is where you got them, since it appears that you cut and pasted the list from there, have you read them?), some cover it in details, some in passing, but all are written by the authors of the software and if you RTFM, none of those can be considered reliable since they are PDFs hosted on a self published website. The software may be notable, but we need good references to show that it is. Pomello, I see that you are apparently new around here, glad to have you on board. Download stats thankfully do not make a piece of software notable. See for example this iphone app with more that a 1000 downloads a day (which, ironically is probably notable given the daily mail article). Basically, what we're looking for here is significant coverage in reliable sources. Blogs generally speaking do not count, but Advanced Acceleration Technologies for Biological Sequence Analyses might, do you have the volume/issue/date for the article in question? --Nuujinn (talk) 11:33, 3 September 2010 (UTC)[reply]
Nuujinn, in the bibliography I pasted above, I didn't read the papers. However, I see that some of them are in reputable journals and computer conferences. Generally, these are subject to peer review. Therefore, I believe that they are considered reliable due to the independent peer review process. (If I'm mistaken about this, please advise.) Please note that LNCS (Lecture Notes in Computer Science), published by Springer, is a prestigious source as well. Each volume is independently edited. It is fairly common for academics to put "pre-prints" of their published articles on their websites so the information can be accessed with paying the high fees charged by some publishers and professional associations.
Of course the software came out after the papers. The software was open-sourced on Sourceforge once it was sufficiently developed. It's clear that the authors of the papers were using earlier versions of the software to advance their research.
The issue of conflict of interest is significant. If one of the authors of the papers wrote part or all of the WP article, this should be disclosed. This may lead to a decision to delete the article (I'm not familiar with the COI rules). However, the notability of the topic has been established based on all the publications in reliable, refereed journals and books. As far as the sock-puppet charge, I have no information about any posters other than myself. I post under my real name, and I don't use any pseudonyms or noms de plume. You can e-mail me using the WP feature. — HowardBGolden (talk) 01:53, 4 September 2010 (UTC)[reply]
I'm confused, you say "It's clear that the authors of the papers were using earlier versions of the software to advance their research" but also "In the bibliography I pasted above, I didn't read the papers"--if you didn't read the articles, how could that be clear? Also, please note, I said "If you look at the article, please note that most of the references cited came out before the software." (emphasis added). Some of the sources referenced in the article are from 2002, 2003, 2004. Regarding the question of COI, the author of the article is one of the designers of the software, see this. And whatever practice is for academics to prepublish, we cannot use pdfs from a self-published web site as sources. Of course, we can use the versions from the journals themselves. ----Nuujinn (talk) 12:55, 4 September 2010 (UTC)[reply]
Let me clarify: I didn't read the papers. I searched for the papers on GScholar and I skimmed some of them. From this I realized that the papers were substantial scholarly papers that incorporated significant material about Fastflow, or were about Fastflow itself. (Have you read/skimmed any of the pasted references? If so, what is your opinion?) I don't have no-charge access at home to the academic journals/books, but I do at work. I will confirm that the PDFs on the website are actually published in the academic journals/books. I can't do this until 7 September 2010 (UTC) when I return to my office. — HowardBGolden (talk) 17:10, 4 September 2010 (UTC)[reply]
Yes, I read some of the references above from the author's web site. I'm unsure about them--on the one hand, assuming they are valid publications, and I'm sure they are, the sources are reliable. On the other hand, it troubles me that they are all by the author of the article and of the software--there lies a clear COI. The relevant portion of policy is, I believe, WP:SPIP, and my reading of that suggests that although the sources are reliable as peer reviewed journals, articles written by an author of the software alone are insufficient to establish notability regardless of where they are published, and what we require are some reliable sources written by "people independent of the topic itself (or of its manufacturer, creator, author, inventor, or vendor) [who] have actually considered the topic notable enough that they have written and published non-trivial works of their own that focus upon it – without incentive, promotion, or other influence by people connected to the topic matter." So far, I haven't been able to find such references via google scholar. On a related note, I'll point out that the term "FastFlow" has been used before, such as FastFlow: A Framework for Accurate Characterization of Network Traffic bySR Kundu, B Chakravarty, K Basu, SK Das in 2006, and in reference to lasers and chemical compounds. Makes the searching a bit tougher. --Nuujinn (talk) 17:33, 4 September 2010 (UTC)[reply]
  • Comment. I'm partly at blame for this being brought here at this time: I should have addressed this sooner. Some history will be useful for other editors to help put this article in its correct context. User Aldinuc (talk · contribs) is one of the co-creators of this programming framework; he created this article and, about nine months ago, co-released the Fastflow programming framework and template library that it documents. He disclosed his identity as user Aldinuc here, btw, something I was informed of after I'd made this 30 May 2010 post to ANI about the Fastflow article and its author. I was too quick by half in marking that ANI post as "resolved"; I did not, for example, realize at the time that the only user to respond to it while the thread was open was not an admin ( although he became one the following month ). When I closed the ANI thread I said I'd take the matter to COIN for further action. I didn't do that, though.
Although I strongly disliked that the framework's author had used Wikipedia to promote his own new creation, and although I believed at the time that what he is calling "Fastflow" is not notable, I nevertheless wanted to get some opinions from C++ programmers on whether this article is useful on Wikipedia. ( Here's one recent attempt, for example. ) Probably the strictly-correct thing to do would have been to AfD right away since my search found, at that time, only a single paper had been presented about it, at an IEEE mini-conference, as I recall. When informed of my ANI post by another user – perhaps I should have informed him myself, even though I referred to him and his software only by pseudonyms at ANI – he responded with this defense of his actions and his article.
I'm still not quite sure what to do with this article. Here's a run-down of my conflicting motives that concern it: (1) I don't like to see useful articles about topics that get little popular-press coverage deleted from Wikipedia. (2) I don't know whether this is a useful article, since this area of programming isn't one I know much about, and since it's so new - version 1.0 was just released this month. (3) I really don't want every computer scientist who creates a new algorithm or implementation to be able to immediately create an article about his work on Wikipedia. (4) I'm personally aware of a researcher, in pharmacology, btw, who has over 15,000 citations to his papers, but doesn't have an article about him or about any of his discoveries here. I raise this fourth item to point up my impression that publishing a couple of papers and giving some talks about one's intellectual creations doesn't necessarily mean they merit an article on Wikipedia. (5) If you look at the contributions history for user Aldinuc (talk · contribs) you'll see that this is a single-purpose-account and may also come to the same conclusion that I alluded to in my ANI post, that this user almost certainly has a significant edit history under some other account. Perhaps that would be on the Italian Wikipedia, I don't know, but no new user creates a first article like this one, via such rapid-fire edits, without prior experience. Anyway, I'm going to think about this topic some more before I !vote, but I did think other editors should have the benefit of this context in coming to their own decisions. Best,  – OhioStandard (talk) 12:10, 3 September 2010 (UTC)[reply]
  • Useful article I had the chance to use the article and related links. I used the library for teaching activity at my university and the students found it useful and somehow precise. I have no concerns relative to the way the matter has been presented. And I think the framework discussed is worth being included. As far as the programming perspective is concerned, the C++ implementation of the framework is a quite good piece of work: simple, effective coding from the user perspective, quite efficient in targeting common multicore systems. It compares well with existing, much more famous libraries. —Preceding unsigned comment added by Mjaroslav (talk • contribs) 13:04, 3 September 2010 (UTC)[reply]
Mjaroslav, I see you just created this account and that the foregoing is your only edit. Would you mind telling us how this matter came to your attention?  – OhioStandard (talk) 13:53, 3 September 2010 (UTC)[reply]
OhioStandard, as I said, I used information from this article for teaching purposes. Coming back to the page, I read comments and I thought I could express my opinion on the subject even I never contributed to wikipedia before. I registered and wrote my comment. —Preceding unsigned comment added by 93.47.42.201 (talk) 22:18, 4 September 2010 (UTC)[reply]
  • *Comment. Nuujinn, yes I'm a newbie as wikipedia editor (I hope that I did not violated any rule), I'm not new as computer scientistic. I'm using this software to build a Monte Carlo simulator, it will appear soon on sourceforge. I knew fastflow via wikipedia, and this is the main reason of my support. Reliable sources for PDF can be easily found on the web. Many of them are referred to reviewed conferences; people usually refer to self-published sources (as Technical Reports) to avoid to violate the copyright of the publishers, look at:
LNCS vol 6321 Springer (ECML/PKDD 2010 that is an A-class scientific venue with over 650 submitted papers and 18% acceptance rate this year, look at the conference webpage).
Ercim news (Journal, article are usually invited).
IEEE PDP 2010.
IOS Parallel Computing 2009.
As I said, I'm a newbie here. I might be wrong, but I honestly don't support the idea that blogs and download count don't contribute to notability in general sense (as far they are third-party sources). They might not contribute to scientific notability, but in order to to discuss it I think we should begin a scientific discussion, thus you should raise scientific problems in the approach, e.g. if the approach is not new, not sound, not motivated or whatever else. Scientifically, I think it is a promising approach, and as I said I believe third-party citation will appear, IMO it is matter of time. The same kind of approach has been recently pushed by Intel with TBB even if with a different back-end that use interlocked operations instead of lock-free approach (see HotPar that by the way cites - at citation 1 - a paper from FastFlow authors proposing FastFlow approach, don't know why they still don't call it fastflow). About the handbook I've mentioned, unfortunately I don't have yet the electronic access to it (I see the paper, I've asked my company to get it).
As a final comment I should say that computer science is not pharmacology, each discipline has its own numbers; the most cited paper ever in computer science get about 4800 citations, the second one 2700 (take a look to cireseerx). In computer science a paper reaching 100 as citation count is rare.
I think the article itself may evolve in the sense of describing the approach: high-level programming coupled with lock-free approach, maybe moving the accent from FastFlow itself to the FastFlow approach to parallel computing. I was planning to modify the article (I did not since I see it is under discussion). I hope it may help the discussion. -- Pomello (talk) 13:37, 3 September 2010 (UTC)[reply]
Excellent, I'll check out the refs later. As for the blogs and downloads, it's policy--personally, I'd like to extend the policy for blogs relative to opensource projects since they do not generally participate in the more traditional media. Downloads wouldn't work, as it would be very hard to verify, and doesn't really mean much. Also, the cite count's not critical--my concern about the papers was that all the ones covering the software appear to be from the author of the software--if we have a couple more from other sources, and we have some from the author that are peer-reviewed, we're in good shape. Thanks for the help--for a noob you're off to a fine start. --Nuujinn (talk) 13:49, 3 September 2010 (UTC)[reply]
  • Comment. I would ask those involved here to consider a couple of issues. One is the wikipedia policy on conflict of interest. If you are involved in the software project, you should probably not be editing the article, and if you do, you should be very careful to maintain a neutral point of view. The other issue is wikipedia policy on sock puppetry. There is an appearance of socking here, and folks will be looking into this. If you are in violation of these policies, it would be a good idea to admit it now. Aside from the impact engaging in such behavior might have on wikipedia, it could also have unintended effects in the real world as these conversations are available to anyone with an interest. --Nuujinn (talk) 17:21, 3 September 2010 (UTC)[reply]
I support Nuujinn's suggestion, and would further suggest that such a disclosure could reasonably be made via e-mail, if that is preferred, to either him or myself. I notice he has the e-mail feature activated for his account, as do I. Nuujinn and I have never interacted before, btw, but from what I've seen here I trust his judgment, and would intend to forward a copy of any e-mail I might receive concerning this matter to him and to EdJohnston (see below), an admin who's familiar with this as well. My purpose in doing so would be to confer as to how we can proceed with appropriate care to minimize any possibility of the kind of unintended consequences Nuujinn refers to above.  – OhioStandard (talk) 18:01, 3 September 2010 (UTC)[reply]
Yes, email is an excellent idea, and I'm happy to receive same, and will also forward as needed. It would be shame if a momentary lapse in judgment affected a promising career, it is now normal practice to google for information on job applicants. Wikipedia is not interested in punitive actions, but we do need to prevent socking so we can all deal with one another in good faith, and I am sure that any misjudgments admitted to would be quickly forgiven. --Nuujinn (talk) 18:15, 3 September 2010 (UTC)[reply]
  • Delete - For previous discussions on user talk, see [2] and [3]. This article doesn't pass our notability standards at the present time. I am particularly concerned that User:Aldinuc has been adding links to Fastflow in other articles. The Fastflow system may eventually win the attention of other computer scientists who are not part of the Fastflow research group. Whenever that happens, and the publications appear, this article can be recreated. To predict now that the Fastflow system will gain wide usage and respect within the field is an exercise in speculation, since the third-party reviews and articles do not exist. EdJohnston (talk) 17:43, 3 September 2010 (UTC)[reply]

Procedural note - Ohiostandard has suggested that there may be sockpuppets or meatpuppets participating in this AfD. In particular, he observes that the User:Pomello account was created by User:Aldinuc, as is shown in the logs. I invite any editors who are affiliated with the Fastflow research group to declare that fact in their comments, to avoid problems later on. Editors who pretend to be independent, but are not, may be viewed dimly. The admin who closes the AfD should be prepared to make any necessary allowances. EdJohnston (talk) 19:10, 3 September 2010 (UTC)[reply]

  • Comment. I'm one of the co-authors of the FastFlow framework. I would ask to consider just a couple of issues. First, from the scientific viewpoint FastFlow is notable since many peer-reviewed articles have been accepted and published into international journals and conference proceedings (some other are under review and hopefully will appear). For example take a look at the accepted paper list of the PKDD2010 forthcoming conference Accepted papers. This is a fact and not an opinion. Furthermore, on wikipedia there are lots of articles which discuss specific software (for example TBB, aMule, ntop, troff, memcached, postfix, just to mention some I know and use) and many of them don't have any scientific articles published on peer-reviewed journals or conferences. Second, probably due to inexperience in using wikipedia, some errors during the writing of the FastFlow wikipedia article have been committed, but I think such errors can be fixed with your comments and suggestions. All in all, I think it is worth to have the FastFlow article on wikipedia. massimot (talk) 20:50, 4 September 2010 (UTC)[reply]
Of course, you're entitled to your opinion. But we have policies that guide our decisions. I see that this is your first edit, so I'm assuming you may not be familiar with them. The primary problem thus far as I see it with the question of notability, as WP defines it, is that all of the articles I've been able to find are authored by an author of the software. That's a conflict of interest, and my opinion is that we must have reliable sources from disinterested parties to establish notability. If you can provide references to some, that would be most excellent. Also, please note that we are concerned here only with FastFlow, not with other articles which may also fail inclusion criteria--other stuff exists, but we're not dealing with that at the moment. Finally, if someone suggested you join the discussion here, that might be consider a violation of the no canvassing policy. People with a potential conflict of interest should be cautious in their comments and edits, as it can be difficult for such to maintain a neutral point of view. I hope all of that is clear, please feel free to post questions here or on my talk page if something is not. --Nuujinn (talk) 21:01, 4 September 2010 (UTC)[reply]
Nuujinn, you're entitled to your opinion as well. As I read the conflict of interest and notability guidelines, they don't interact in the way you describe. A developer of an idea can write an article that is published in a peer-reviewed (reliable) publication. That doesn't conflict with the conflict of interest guidelines (as I read them). It isn't necessary to have the article written by an independent party if it is peer-reviewed. (If you wish, I'll walk you through the various WP policies and guidelines that I believe support my position.) As I said, when a WP article is written or edited by someone with a conflict of interest, a higher standard is necessary to assure WP:NPOV, but WP:COI doesn't prohibit such articles absolutely. — HowardBGolden (talk) 21:12, 4 September 2010 (UTC)[reply]
By all means. As I said, I am uncomfortable with using only sources written by the author of the software, since I do not believe they establish general notability, but by no mean certain. What is quite surprising to me is that we don't yet have any sources that are not written by people involved in the project--if the software is truly notable, I would think it easy to find such sources. --Nuujinn (talk) 21:21, 4 September 2010 (UTC)[reply]
  • Comment. It's not just this article that's under consideration, btw. Aldinuc also inserted information about his software framework into at least a dozen related articles, after he created this one. If this article is deleted, those edits should be reverted as well.  – OhioStandard (talk) 10:36, 6 September 2010 (UTC)[reply]
  • Keep per HowardBGolden's findings at the top of this discussion. Journal articles and other scientific publications are reliable sources, even if they're all by the same people; this isn't an issue of self-published sources or improperly citing oneself. Nyttend (talk) 13:09, 6 September 2010 (UTC)[reply]
It's my understanding that there are no references provided *that actually mention Fastflow* that are not written by members of the Fastflow research group. (The reference list included a couple of articles on the general issues of parallel computing). If projects are allowed to establish notability by simply getting their own publications accepted in journals, then the gate is wide open. EdJohnston (talk) 13:47, 6 September 2010 (UTC)[reply]
I've just made a run through the reference listed above, most, IMO, fail the bar of reliable sources as they have not yet appeared in print, or are essentially self-published. My !vote will be delete unless we can find some additional sources written by parties not closely connected to the software project, as I believe WP:SPIP applies. I also encourage other editors to examine the sources carefully, there may be some degree of academic puffery going on here. --Nuujinn (talk) 14:47, 6 September 2010 (UTC)[reply]
EdJohnston, the crux of the issue (IMO) is what it means to have an article published in a refereed academic journal. I believe this establishes a presumption that the article is notable within that field, because that's what the referees do. This is especially true when the journal is well-known and selective. An article published in such a journal has risen above WP:SPIP IMO. You worry that "the gate is wide open" if this is allowed. I view that as a hypothetical that needs evidence. I think it makes sense to allow refereed academic journal articles to establish notability and see what happens. If this leads to bad results, then the criteria would need to be tightened. Let's wait and see. — HowardBGolden (talk) 16:36, 6 September 2010 (UTC)[reply]
For the sake of completeness take a look to: Current Bioinformatics, 2010, 5, 176-194 Advanced Acceleration Technologies for Biological Sequence Analyses by Xiandong Meng, Yanqing Ji and Hai Jiang that can be found here. FastFlow is reference [61]. I'm involved in fastflow. The electronic version is under copyright not sure it can be freely downloaded, I'll be happy to send a copy to anybody willing to have a look by private mail. Regards. Aldinuc (talk) 16:55, 6 September 2010 (UTC)[reply]
I requested a copy of the article from Aldinuc. It contains the following brief statement about Fastflow: "OpenMP has been used to parallelize the Smith-Waterman algorithm. Aldinucci et al. [61] has demonstrated that the OpenMP approach outperforms several parallelization schemes such as Intel Threading Building Blocks (TBB) and Cilk." Reference [61] is "Aldinucci M, Torquati M, Meneghin M. FastFlow: efficient parallel streaming applications on multi-core, 2009; arXiv:0909.1187v1." In my opinion, this demonstrates the notability of Fastflow. It is written by independent researchers in bioinformatics, not by members of the project. — HowardBGolden (talk) 17:57, 6 September 2010 (UTC)[reply]
  • Question Does the text of the article mention fastflow directly? I believe that would count as passing mention of OpenMP, but if the text of the article does not even mention fastflow directly, I fail to see how that could be considered even a passing mention, much less significant coverage, of fastflow. --Nuujinn (talk) 18:21, 6 September 2010 (UTC)[reply]
  • Response The text of the article refers to the reference (which is about Fastflow) demonstrating using OpenMP (and, by implication, Fastflow) to outperform some other parallelization techniques. I think this is a direct reference. It also makes clear that the independent authors recognize the notability of Fastflow. (I think we are quibbling at this point. The article is about advanced techniques to do sequence analysis. The discussion of Fastflow is not a passing mention, since it is one of the advanced techniques. This isn't like a mention of a rock band in a biography of a famous politician (IMO), which would be a passing mention). — HowardBGolden (talk) 19:06, 6 September 2010 (UTC)[reply]
  • Comment, respectfully, I disagree. Passing mention means that a reference merely mentions a topic, but does not provide significant coverage, and I don't think we even have passing mention here. The Advanced Acceleration Technologies for Biological Sequence Analyses article does not directly address fastflow at all, but rather uses an article by Aldinucci covering fastflow to support the statement that OpenMP can be used to parallelize a particular algorithm in a manner that is more efficient than other approaches. To connect the mention of OpenMP in source A (Advanced Acceleration Technologies for Biological Sequence Analyses) to coverage of fastflow in source B (FastFlow: efficient parallel streaming applications on multi-core) may well be crossing the line established by WP:SYNTH--as you note, you're relying on an implication, rather than an explicit connection drawn by the authors of source A. Rather than showing that "independent authors recognize the notability of Fastflow", the article shows that some researchers recognize the notability of OpenMP, which has been around for a while now. But even if drawing the connection between the mention of OpenMP in the one article to fastflow in the other does not violate WP:SYNTH, it is still no more than passing mention--if that's all that's in Advanced Acceleration Technologies for Biological Sequence Analyses, there's no significant coverage of either OpenMP or fastflow there. --Nuujinn (talk) 20:06, 6 September 2010 (UTC)[reply]
  • Comment, respectfully, the discussion here is no longer on notability, but on how a "related work" section in a scientific paper should be written according to some interpretation of wikipedia rules. If a scientist would like to tell "OpenMP is fast" just put there a reference to OpenMP not to fastflow. The cited paper present fastflow not other topics. I'm really curious to know why, according to this vision, these authors mentioned a fastflow paper, since, as you said there plenty of papers discussing openMP and Smith-Waterman algorithm. The fact that the "fastflow" is not explicitly written in the text is not relevant at all. Scientific publishing does not have the same rules as wikipedia. I'm still part of the fastflow team. Aldinuc (talk) 21:43, 6 September 2010 (UTC)[reply]
  • Comment, I'm sorry, but I believe you are incorrect, as the primary issue here is notability as defined by wikipedia policy. I'm curious, too, as to why these authors mentioned a fastflow paper--curiosity not withstanding, the problem is they do not say why, and any interpretation we make as to the reasons why they made the decision they did constitute WP:OR. We are supposed to follow reliable sources, not make assumptions or interpretations. You are, however, correct when you state that "Scientific publishing did [does?] not have the same rules of [as?] wikipedia"--wikipedia has its own set of policies and guidelines, and those are the ones we follow. --Nuujinn (talk) 22:01, 6 September 2010 (UTC)[reply]
  • Comment, we? Are you meaning that wikipedia pages in general accomplish your interpretation of notability rules? Check the pages of TBB, aMule, ntop, troff, memcached, postfix (to mention only the one mentioned the post above) and prepare to produce a lot AfDs ... Thanks for notifying the typos (fixed). Aldinuc (talk) 22:18, 6 September 2010 (UTC)[reply]
  • Comment, do I mean are we always successful in conforming to our policies? No, because wikipedia is not finished and never will be. Is it relevant that other articles on topics that are not notable exist? No, because other stuff exists. Are we going to have lots of AFDs every day, day in and day out? Yes, I think so. If you think those other articles cover non-notable topics, by all means, bring them to AFD. If you think they are notable and can be fixed, by all means, have a go. If you need help, feel free to ask me on my talk page. --Nuujinn (talk) 22:55, 6 September 2010 (UTC)[reply]
  • Question, Aldinuc, are you arguing that either of those meets the guidelines set by WP:RS? I do not believe that sourceforge or company blogs are considered a reliable sources. Do you know who the authors are? --Nuujinn (talk) 20:52, 6 September 2010 (UTC)[reply]
  • Answer, I'm just putting here information I'm finding on the web. All of them have been compiled much before this discussion. I (and the fastflow team in general) have no relations with the authors. I do believe that a person who download a software (an advanced software), implement with it a non trivial application (spending days or weeks of their time), then write a report on the web, or create a brand new project using this software is really worth of attention. Much more than any direct reference written in a paper. Why they should not be reliable sources? Is physical print that give you reliability? Thus should we conclude that wikipedia itself is, on the whole, an unreliable source? I'm still part of the fastflow project. Regards. Aldinuc (talk) 21:11, 6 September 2010 (UTC)[reply]
  • Self-published sources are not considered reliable, see WP:SPS. And yes, Wikipedia is not a reliable source, since it's a wiki, see WP:RSEX. Being in print is not a requirement, but blogs and software project pages are not considered reliable, esp. if the author is unknown. If an author is a known expert in a given field, and has appeared in reliable sources, their blog might be considered a reliable source, but generally speaking, some formal editorial oversight is required. There are some things that I am truly an expert, but my expertise does not make my blog a reliable source. Anyone can get a sourceforge account and say what they like. Yes, doing the kind of work you describe is, I think, valuable and worthy of attention, but the question is worthy of what kind of attention and where. Wikipedia is an encyclopedia, and it has rules governing what belongs here. Please read the relevant policies. --Nuujinn (talk) 22:55, 6 September 2010 (UTC)[reply]
  • Comment, well, it seems that I found another paper mentioning fastflow: Concurrent Evaluation of a Directed Acylic Graph by Rakesh Joshi (paper here take a look at beginning of page 3, there is a direct reference). It appeared at the ParaPlop 2010 conference held in Carefree, Arizona last march 2010. There cannot be any doubt that the mentioned fastflow is this fastflow, as the comment is about a SPSC queue bases lock-free/CAS-free programming environment. I'm really thinking that here the only issue is the time one would dedicate to this kind of searches (search on google: fastflow queue, then go to page 5 of the results, by the way in previous pages you can find several blogs also discussing fastflow). I should add that unfortunately I don't know the author, as he is advocating a nice work, and that I'm still involved in the fastflow project. Best regards. Aldinuc (talk) 22:25, 7 September 2010 (UTC)[reply]
I guess this is a mention of Fastflow:

The graph pattern resembles a network of pipeline filters, notably the recently available FastFlow framework, which relies on assemblies of SPSC queues, copy semantics and allocates dedicated copy threads (Emitters & Collectors) to realize lock- free/CAS-free operation.

Not clear if this is more than a passing mention, but Rakesh Joshi is actually seeing a resemblance between his technique and Fastflow. EdJohnston (talk) 22:37, 7 September 2010 (UTC)[reply]
Agreed, but I do think it is more passing mention than significant coverage. --Nuujinn (talk) 23:39, 7 September 2010 (UTC)[reply]
Fastflow is the only third-party programming framework mentioned in the paper. Aldinuc (talk) 00:25, 8 September 2010 (UTC)[reply]
With all due respect, can you point to a WP policy that suggests that being the only third-party programming framework mentioned in a paper confers notability? --Nuujinn (talk) 00:30, 8 September 2010 (UTC)[reply]
With all due respect, the Rakesh Joshi's paper presents a programming technique in detail. Toward the end of the paper, author recognizes that a similar technique has been already used in the fastflow framework. Of course the argument applies backward, as the description of the technique provided by the paper indirectly explains some of the techniques adopted in fastflow. By the way the structure of the paper the author chose is smart as enabled him to make the analogies in extreme synthesis and elegance (as I said, interesting paper). FastFlow is indeed significantly covered by this paper (I were editing FastFlow page, and I no longer do it, I'll certainly use this material). I hope nobody here is going to measure significance of a sentence by its length. With respect the last sentence (and only with respect to this), it might be useful to remember that Salvatore_Quasimodo, Nobel price 1959, got the the price thanks of a composition of 15 words (70 characters). I'm also genuinely curious to know EdJohnston 's opinion after this discussion as he/she expressed their vote before all this long discussion (also via e-mail, if he/she would not like to make it public). Best regards. Aldinuc (talk) 10:25, 8 September 2010 (UTC)[reply]
  • Delete, I've taken my fourth jaunt through google scholar and news, my local uni library, and I'm just not seeing anything that I believe meets the bar for general notability--all of the sources that meet the criterion of significant coverage are authored by a major contributor to the both the article here and the software itself, and my reading of WP:SPIP leads me to the conclusion that such sources should not be considered reliable. Many arguments have been presented for a keep, but thus far, few have a grounding in WP policy. --Nuujinn (talk) 00:43, 8 September 2010 (UTC)[reply]
  • Comment I reiterate my position that publication in a scholarly journal (which subjects the article to peer review) is sufficient to create notability. I will quote from Uncle G: "The places for publishing new theories and new discoveries are the appropriate scholarly journals. For new discoveries and theories, it is those journals that perform the process of fact checking, peer review, and publication, and from those journals that new things are accepted into the general corpus of human knowledge." [4] The point is that Fastflow has been published (multiple times). The scholarly journals and conferences have done their job. It is time to accept Fastflow into the general corpus of human knowledge. Every new idea has an originator. Wikipedia doesn't accept the originator's word for the notability, but Wikipedia should accept the judgment of scholars in the field that the subject is worthy of publication. WP:SPIP doesn't apply in this situation (IMO). — HowardBGolden (talk) 01:17, 8 September 2010 (UTC)[reply]
  • Comment. Well, first of all, fastflow is a software library, and really neither a theory nor a new discovery, it is rather more an invention. Uncle G is a very good editor, but I think you're quoting an essay, yes, rather than a policy or guideline. Also, in that same essay, Uncle G states "Evidence that something has been acknowledged by people other than the subject's own proponents/creators/authors/inventors, and thus become a part of the corpus of human knowledge, is that sources from those other people also exist. Evidence that a source is correct is that other people have performed research in the same area, and published material that concurs." Now, please forgive me for being a bit long winded here, but while the sources we're looking at are academic and reviewed, the article is about a software product, not a theory or discovery--a framework used to process data. If you look at the sources that are not authored by the creators of the software packages, you'll note that these are about other subjects, not the fastflow software. So we do not have a situation here where independent parties have also found the subject of the article notable enough, at least not yet, to provide significant coverage. And I think that is a key point--in that last quote from Uncle G's essay suggests that independent sources are required, and that's parallel to the intent of WP:SPIP. --Nuujinn (talk) 01:42, 8 September 2010 (UTC)[reply]
  • Comment, I'm really mystified by this last post; maybe I'm not the only one participant too much involved in this discussion (as I declared many times I'm part of the FastFlow team). FastFlow is not only a software library, at all. The FastFlow library is a software artifact developed to prove that the FastFlow methodology is sound and working. Cannot believe this is not yet clear. The FastFlow methodology brings in the scientific arena *several* new discoveries: one of those is how a multiple-producer-multiple-consumer queues can be implemented without any lock and without any interlocked operation. This has been considered *impossible* for long time now (to mention just one discovery). Maybe in this discussion is really lacking the opinion of an expert of the topic we are discussing. Moreover, even assuming - in very good faith - that people understanding is that FastFlow is just a library, then why looking academic articles as reliable sources? Is notability of software libraries established by academic, third-party articles and not download count, availability, maintenance, vitality, bug fixing readiness, are we reinventing software engineering discipline? I'm really afraid that discussion on notability issue is getting tangled up with the question of socking or other kind violation (that by the way has been separately discussed in proper places). I think any independent experienced editor coming here by change, is likely to get the same doubt. I say this because, IMO, a reasonable outcome of this discussion can be "1) the fastflow page needs several improvements, also finalized to explain the novel methodologies that have been published already in academic peer-reviewed articles, 2) these updates cannot be done by an editor that is in COI, as aldinuc for example, who should abstain to edit in this field" and not certainly that the fastflow topic is not notable. That's my very opinion. Thanks you all, anyway. Regards. Aldinuc (talk) 11:04, 8 September 2010 (UTC)[reply]
  • What you have just said above suggests not that the fastflow article exists, but rather that some of the material in this article belongs in a more general one on parallel processing or a related algorithm. Yes, "notability of software libraries [is] established by academic, third-party articles" which provide significant coverage, but in this case, there are not any to be found--only 3rd party passing mention and 1st party sources. If, as you say, there's a major new discovery based in fastflow, one would think someone else out there not associated with the software project would have written about it, but apparently that is not the case, at least not so far, and to be blunt, the lack of significant coverage by independent third parties in reliable sources is pretty much the only issue here. Perhaps that will happen soon, but in the meantime, I think we do not have a notable subject. Also, I don't know why you say "discussion on notability issue is getting tangled up with the question of socking or other kind violation" since we haven't talked about anything but notability for a while now. --Nuujinn (talk) 12:04, 8 September 2010 (UTC)[reply]
  • With all the due respect, you are not introducing any new argument, just continuing to repeat the same argument increasingly loudly (as clearly shown by your usage of bold typeface, such an inelegance) in front of any new evidence that has been produced. As an example, you marked as "in passing mention" the one in link to article, an article that has a primary topic StochKit-FF (i.e. StochKit-FastFlow shortened, as it is perfectly clear reading the paper), and you never stepped back from this position. This suggests that your evaluation of "in passing mention" is, at least, questionable. Again with all the due respect, about the sock issue I mentioned, I just cited Nuujinn's thoughts (you) who is, up to my knowledge, an expert editor (here), and notice that the discussion at this point in time was already well developed (see here). Regards. Aldinuc (talk) 14:35, 8 September 2010 (UTC)[reply]
  • Comment, sorry about the bolding, I started my standard Comment and got distracted. In regard to this article, I see that fastflow was used to build a new faster version of stochkit, but little about fastflow itself, hence my belief that it's passing mention. My main objection to using that as a reliable source, however, is that it is not from a 3rd party, and thus falls afoul of WP:SPIP. Yes, this is very repetitive--I'm trying to discuss these issues relative to WP policies, but not having much success, so I'll go do something else for a while. --Nuujinn (talk) 15:31, 8 September 2010 (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.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia:Articles_for_deletion/Fastflow_(computing)&oldid=1137844624"