Category talk:Discrete distributions

Linking Java applets as external links to various probability distributions

Would it be encyclopedically valuable and informative to include external links at the bottom of probability distribution articles that point to specific Java applets, which provide interactive graphical and computational interfaces for exploring the concrete distribution of interest? Please enter your comments below. Some obvious pros/cons include:

Cons
- WP:ELNO: (under "Links normally to be avoided") reads:

Direct links to documents that require external applications (such as Flash or Java) to view the relevant content, unless the article is about such rich media. If you do link to such material make a note of what application is required.

- This Java applet is signed by an authority not recognized by a vanilla Firefox install requiring several security questions to be answered before the page can be viewed.
  - It appears this note is regarding this applet, which seams to be Thawte-signed.
- There may be many, diverse or/and inconsistent implementations of applets for one or many distribution families.
- The editor adding the links may have a WP:COI because he appears to be associated with the same institution as the links. He has not stated what the COI situation is when asked.
- Errors are present in the implementations (i.e. the Negative Binomial with r=30 and p=0 does not have a positive variance). Notably, this is the default value for this distribution.
  - Is this regarding this applet? This is a general discussion about the encyclopedic and informative value of including external links to distribution Java applets. It may be more constructive to address specific reviews about concrete applets to the corresponding developers Wikipedia:Etiquette.
Pros
- Significantly enhanced ability to understand a distribution, and its properties, via graphical manipulations and interactions with the distribution using a platform agnostic Java applet (e.g., setting of parameters, control of shape, etc.)
- Applets integrate back-end statistical computing libraries, which may use complex Special functions, with dynamic and intuitive graphical user interfaces.
- Enable computation (e.g., critical values, probabilities of interest, etc.), discovery of applications and identification of possible limitations for each distribution.
- Wikipedia uses nofollow tags, and thus additions of external links will not alter search engine rankings.
- TBD
Example: This is just one instance of a static image and a corresponding interactive graphical Java applet for Poisson distribution.

The main question is: How valuable to the community may such external links be? Iwaterpolo (talk) 20:14, 18 May 2009 (UTC)[reply]

My initial reactions is: who doesn't have Java? But maybe that's naive. Michael Hardy (talk) 20:39, 18 May 2009 (UTC)[reply]

Michael Hardy, I think you have picked up an issue of essentially measure 0 here. A better question (but still less important to this discussion) would be this: do these applications render correctly on all browsers? The answer is no. On my browser the axis labels (a) are cut off, and (b) when they change do not first clear themselves but just write over so that there is a 4, superimposed on a 5, superimposed on a 6, ... PDBailey (talk) 23:37, 18 May 2009 (UTC)[reply]

previous discussion

This discussion started as a result of discussions on other pages. Here are the links.

question at hand

The question at hand is this: Does this tool / set of tools merit linking to from every distribution that it links to. General policy for external links is at WP:LINKS and the nutshell description reads, "External links to an article can be helpful to the reader, but they should be kept minimal, meritable, and directly relevant to the article." Later it reads

Some external links are welcome (see "What should be linked", below), but Wikipedia's purpose is not to include a comprehensive list of external links related to each topic. No page should be linked from a Wikipedia article unless its inclusion is justifiable.

In addition, there is a general reticence to link to non-accessible or annoying content such as those that require Java, see the text of the Wikipedia policy above in cons. PDBailey (talk) 01:53, 19 May 2009 (UTC)[reply]

There have been other discussions of this in the past, besides those mentioned above. A major alternative approach not mentioned above is to link all/most distribution articles to a Wikipedia article listing this type of online software and not have them listed elsewhere unless they deserve their own articles. Then, when/if they become dead they can be removed from a single place. There would also be scope for some description of capabilities and notes of any foibles such as pointed out above. Melcombe (talk) 09:04, 19 May 2009 (UTC)[reply]

Defining Java applets as "non-accessible" or "annoying" content probably does not represent NPOV. It's absolutely true that Java applets may have slight platform, network, version and preference specific appearance. In general, however, most people consider them helpful and such applets are widely accepted as useful. Graphical and dynamic distribution applets that enable interactive exploration of probability distributions included as external links may be useful for many users (perhaps not all users). The alternative proposal below (a new Wikipedia article listing this type of online software, that all distribution articles link to) may also be a valuable solution. Iwaterpolo (talk) 05:39, 21 May 2009 (UTC)[reply]

RFC

Community feedback is requested regarding the pros (informative and encyclopedic value) and cons (potential for platform/browser specific behavior) of external links listed at the bottom of Wikipedia probability distribution articles. Such external links typically point to Java applets, which provide interactive graphical and computational interfaces for exploring the concrete probability distribution of interest, but may introduce browser specific appearances. Iwaterpolo (talk) 19:01, 29 June 2009 (UTC)[reply]

Can you provide some links to a few examples that are/were in articles? My initial thought is that it's fine to link to Java applets but note that it is an applet. If the applet becomes unavailable it will eventually get removed much like a dead link. I really don't see why anything should be done differently. If an applet is particularly demanding of computer resources, then it may be wise to note that in the external links. Sifaka ^talk 07:46, 3 July 2009 (UTC)[reply]

Here are some examples of links that were suggested for specific Probability Distribution Wikipedia articles:

Wikipedia article: Poisson_distribution --> suggested link: Interactive Poisson Distribution Web Applet (Java)

Wikipedia article: Uniform_distribution_(discrete) --> suggested link: Interactive (discrete) Uniform Distribution Web Applet (Java)

Wikipedia article: Negative_binomial_distribution --> suggested link: Interactive Negative-Binomial Distribution Web Applet (Java)

Wikipedia article: Rice_distribution --> suggested link: Interactive Rice Distribution Web Applet (Java), etc.

An example of how these links were suggested for inclusion in articles is available here: [1]. Iwaterpolo (talk) 22:50, 7 July 2009 (UTC)[reply]

Sifaka, I think you are suggesting that the blanket policy is just wrong. I think this kind of comment would be best left as a proposed change to WP:ELNO (see above or the link for the relevant text of the policy). I could go into why I don't even have Java on in my browser, but I think that would be better said over there too. PDBailey (talk) 04:29, 8 July 2009 (UTC)[reply]

Iwaterpolo must clearly state COI. Before this RFC moves forward, Iwaterpolo has some clarifying to do here. There are really two issues here, is Iwaterpolo also 128.97.126.109 and is Iwaterpolo affiliated with UCLA.

A brief history of this proposal.

Iwaterpolo and 128.97.126.109 added these links to SOCR pages and they were removed by myself and User A1. Iwaterpolo and 128.97.126.109 objected and User A1 pointed to COI and WP:SPAM (see User_talk:User_A1#link_spam_by_Iwaterpolo). I also asked him to state if he was affiliated with 128.97.126.109. He made a request for a third opinion and the 3O responder did not get involved because there were already three people involved.
Iwaterpolo then started the above discussion. After a lull in comments he suggested that "There seem to some community evidence to suggest these may be useful." (see User_talk:Pdbailey#continuation) and suggested readding them or (if I really insisted) creating a page just for links to SOCR. User A1 and I prevailed upon him not to do this. I also asked him if he is affiliated with 128.97.126.109 and to clearly state any COI (see User_talk:Iwaterpolo#other_account.3F).
Now, he started this RFC.

Iwaterpolo has been focused on getting these links posted and has used a substantial amount of what appears to me to be wikilawyering to get them posted (look at his edit history, one edit since this dispute started on May 17 has been on any other topic). To date Iwaterpolo has not responded or commented in any way on if he is 128.97.126.109 or his affiliation with UCLA, or SOCR. I submit this must happen before any links to SOCR should even be considered or this RFC should move forward. PDBailey (talk) 10:58, 8 July 2009 (UTC)[reply]

(moved down from after the second bullet of PDBailey's July 8 comment by PDBailey) PDBailey, as you are providing background on this discussion, please avoid misrepresenting personal interpretations as facts! Iwaterpolo did not ask for "creating a page just for links to SOCR, as you indicated above! Here is verbatim what Iwaterpolo stated : "If you still strongly object to this, will it be better to create a brand new page, say Interactive Probability Distributions Java Applets, that can be listed as a category at the bottom of all distributions." This is not a discussion about SOCR or Java, per se. It is a discussion about the informative value of adding external applet links that interactively demonstrate properties of distributions. There are many other sites that provide varieties of applet types (e.g., Flash) that can be included as external links. Thanks. Iwaterpolo (talk) 18:12, 10 July 2009 (UTC)[reply]

Iwaterpolo, I'm sorry, you are right you proposed adding a page that links to java applets, not just SOCR applets. PDBailey (talk) 12:11, 12 July 2009 (UTC)[reply]

This discussion, see above, is about the encyclopedic and informative value of including external links, at the bottom of probability distribution articles, that point to specific Java applets. It is not about SOCR, user Iwaterpolo, or user PDBailey. If PDBailey has personal problems with Iwaterpolo, then this should be addressed using other means. Please keep this discussion clear and focused to the point - pros and cons of including such links. These links need not be links to the SOCR distribution Java applets. As pointed out above, there are a number of sites with excellent Java applets demonstrating properties, calculations and interpretations of various probability functions. If PDBailey does not have Java-enabled browser, Michael Hardy pointed above that "most people" do. Iwaterpolo is associated with the UCLA SOCR project and does not use any other aliases. PDBailey described Java applets as "non-accessible" or "annoying" content, which can explain his stand on this issue, ref. NPOV. It would be more constructive to focus this discussion on the informative value of such external applet links - e.g., see the example above comparing the static-image vs. one example of a dynamic interactive and applet for Poisson distribution. Iwaterpolo (talk) 17:52, 10 July 2009 (UTC)[reply]

Iwaterpolo, do you then agree that it would be inappropriate for you to add links to SOCR on a page and that instead you should, i.e. point out the existence of the link on the talk page while stating your conflict of interest? PDBailey (talk) 12:11, 12 July 2009 (UTC)[reply]

In retrospect, I (Iwaterpolo) should have initiated this discussion thread (under continuous or discrete distributions), before I submitted my initial edits on this topic (e.g., [2]). Now I can see how one can perceive these additions as COI. I do believe, that such interactive resources for dynamically exploring properties of any probability function (e.g., distribution, density, generating, etc.) would be extremely valuable to learners, educators as well as developers. There are a variety of groups that provide such resources using different technologies (e.g., Java, Flash, VM, services, etc.) I apologize for my initial single-handed action - it was intended to contribute to our collective efforts, not to spam, distract or provoke a controversy. I support any of the 2 solutions proposed so far:

Within each probability distribution article, include low key links (on the bottom of the article, under "External links") to appropriate and functional external interactive content (e.g., applets), clearly specifying the requirements for viewing this content, or

Create a brand new page, say Interactive Probability Distributions Applets, that can be listed as a category at the bottom of all distributions. One slight problem with this approach is that it directs to one fixed page from all probability distribution articles. So, the reader will not be able to infer apriori if a collection of tools contains an interactive tool for a specific distribution. Many sites provide tools for a small number of distributions.

In either case, the community will continue to validate, update and remove such external links to ensure their appropriateness, correctness and functionality. Iwaterpolo (talk) 21:53, 12 July 2009 (UTC)[reply]

oppose. So I think that the question is this: why does Wikipedia have this rule and is this a case where that reasoning doesn't apply. The policy is this

Direct links to documents that require external applications (such as Flash or Java) to view the relevant content, unless the article is about such rich media. If you do link to such material make a note of what application is required.

On the talk page there are two discussions of this, Wikipedia_talk:External_links/Archive_2#Explanation_of_edit_by_User:Tony_Sidaway_on_Flash.2C_Java.2C_etc and Wikipedia_talk:External_links/Archive_1#External_linking_to_Flash.2C_Java.2C_etc

The objections raised in these discussion seem (to me) to be easily remedied by simply noting that the link is to a Java applications. In addition, their content is inherently interactive so something like java, javascript, or flash is a must. While we could argue about javascript versus java, I think there is only so much training wheels Wikipedia must have and this could be over the line. Again, this is more of a discussion for external links than here.

However, in any case, I would submit this: any link should not request complete control of your computer regardless of who signs it. This is apropos because the above proposed links do make this request when I last clicked on them. PDBailey (talk) 22:41, 16 July 2009 (UTC)[reply]

Notwithstanding the outcome of that discussion, I still oppose these links or the creation of another page just for links to interactive applications. the reason is this: WP:NOTLINK. WP:LINKS reads (in the nutshell header), "External links to an article can be helpful to the reader, but they should be kept minimal, meritable, and directly relevant to the article." I don't these these sorts of links meet that criterion. I think they could be useful, but I think lots of things could be useful. If these types of links are truly useful, is it really encyclopedic to make links to them just because they are there and useful? we already have our answer in the header of What Wikipedia is not (Wikipedia:What Wikipedia is not#Content reads, "merely being true or useful does not automatically make something suitable for inclusion in an encyclopedia."). That would be more appropriate for a page that advises people on what websites might be useful to them based on keywords, i.e. a search engine. PDBailey (talk) 22:41, 16 July 2009 (UTC)[reply]