Wikipedia:Articles for deletion/Fastflow (computing)

Fastflow (computing)

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)

• Note: This debate has been included in the list of Software-related deletion discussions. -- • Gene93k (talk) 00:30, 3 September 2010 (UTC)
• Note AfD notice is not transcluded on the article page. Jclemens (talk) 00:51, 3 September 2010 (UTC)
  • 'tis now, I think. --Nuujinn (talk) 01:11, 3 September 2010 (UTC)

• 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

[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

Talks

FastFlow: a pattern-based programming framework for multicores. Dagstuhl seminar 10191, Schloss Dagstuhl, Germany. May 2010. Invited. Slides available on-demand.

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.

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.

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)

• Comment. The above refs can be viewed in-context at the author's 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)

• Keep A programming framework (not a videogame or a windows utility) less than one year old that stably receives 150+ downloads per month (60 from 1 to 3 sept 2010 since a new revision has been posted) and thousands of contacts can be hardly considered not significant IMO (see sourceforge stats). It has been mentioned in several independent blogs (e.g. Hack the market - lock-free programming), and cited in third-party scientific handbooks (e.g. Advanced Acceleration Technologies for Biological Sequence Analyses, also considering that a scientific works are hardly cited before one year due to revision and publishing time). By the way, the article appears, IMO, not written like an advertisement (as claimed in the box); it requires some updates, as noticed by HowardBGolden. -- Pomello (talk) 10:20, 3 September 2010 (UTC)

I'm striking the above !vote as an obvious sock or collaborator of Marco Aldinucci, the co-creator 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 the designer 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, Marco, 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)

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)

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)

• 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)

• 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 Marco Aldinucci; he created this article and, about nine months ago, co-released the programming framework and template library that it documents. Marco 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 – Marco/Aldinuc 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 Marco 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)

• 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)

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)

• *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)

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)

• 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)

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)

• Delete - For previous discussions on user talk, see [1] and [2]. 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)