User talk:DavidMCEddy

From Wikiversity
Jump to navigation Jump to search
Welcome!

Hello DavidMCEddy, and welcome to Wikiversity! If you need help, feel free to visit my talk page, or contact us and ask questions. After you leave a comment on a talk page, remember to sign and date; it helps everyone follow the threads of the discussion. The signature icon Signature icon.png in the edit window makes it simple. All users are expected to abide by our Privacy policy, Civility policy, and the Terms of Use while at Wikiversity.

To get started, you may


You don't need to be an educator to edit. You only need to be bold to contribute and to experiment with the sandbox or your userpage. See you around Wikiversity! --Draubb (talk) 4:32pm EDT, 22 May 2013 (EDT)
Crystal Clear app gnome.png
Smiley.jpg

Hi DavidMCEddy, I am Draubb. Nice to meet you. I love the beach. And I love to play cricket and soccer. I am currently creating historical and learning pages! Love to see you around Wikiversity. I am also learning to play piano. I know how to play Ode to Joy! What do you like? Can I be your happy mascot?

Need help?[edit source]

Hello DavidMCEddy, If you need any help - just go to my my discussion page! Thank you. --Draubb (discusscontribs) 00:19, 26 May 2013 (UTC)[reply]

Effective defense and ISIL[edit source]

I just stumbled across this page and was quite impressed. Great work! I've added it to Category:Featured resources. (The preceding unsigned comment was added by Mu301 (talkcontribs) 3 January 2016)

Naming Conventions[edit source]

In a recent edit you mentioned Wikimedia naming conventions. Note that each project has its own naming conventions. At Wikiversity, either Sentence case or Title Case may be used. In my experience, academics seem to prefer Title Case. See Wikiversity:Naming conventions#Casing for more information. -- Dave Braunschweig (discusscontribs) 20:28, 16 December 2016 (UTC)[reply]

Thanks. I'm in the process of creating 11 stubs to be completed hopefully in part by others under a soon-to-be created "Category:People's agenda". I'm already roughly a third through this process, and I think I'll continue with "Sentence case", since you say that's still permissible on Wikiversity. I expect I'll be introducing more people to Wikimedia editing, and they'll need to use Sentence case if they do anything with Wikipedia. (And thanks for the "quality resource" recommendation of my earlier piece.) DavidMCEddy (discusscontribs)
Thanks again for the suggestion. I updated the titles for other reasons and converted them all the Title Case, as you suggested. The "Move" facility made it easier than I anticipated, because it automatically created the needed redirect. DavidMCEddy (discusscontribs) 12:30, 18 December 2016 (UTC)[reply]

Effective defense[edit source]

Hi DavidMCEddy!

The resource Effective defense appears to be well developed and ready for learners! Would you like to have it announced in our Main Page News? --Marshallsumter (discusscontribs) 01:00, 13 March 2017 (UTC)[reply]

Would I like to have it announced on the Wikiversity Main Page? Yes, but I think it should be better connected than it is right now.
What have you seen that encouraged you to suggest this?
I ask, because I'd first like to hear from you what you've seen. Then I'd like to invite your comments on other things I've done that may have escaped your notice. Then I'd like to make sure everything is linked from a common category, so the announcement could refer to that.
In particular, have you seen my latest "Winning the War on Terror"?
That just went live yesterday, and I'm still polishing it.
Thanks very much for your interest. DavidMCEddy (discusscontribs) 02:03, 13 March 2017 (UTC)[reply]
I've read Effective defense! It is effective when initially read; hence, my request! Good authors always want to make their works better! Ultimately, it's your call. You will likely continue to make it better independent of its being announced! Everyone announced so far has seen further improvement and the bump up in readership helps! --Marshallsumter (discusscontribs) 02:57, 13 March 2017 (UTC)[reply]
I've read Winning the War on Terror! Good so far! It's on my list to watch. When you appear to have reached an effective plateau, I'll ask the same of it. --Marshallsumter (discusscontribs) 02:57, 13 March 2017 (UTC)[reply]
I've also read many of your earlier resources, but did not ask about them. Perhaps when you return to them, I'll ask then. The announcements usually are for resources, not categories. Sometimes portals are included. Cheers! --Marshallsumter (discusscontribs) 02:57, 13 March 2017 (UTC)[reply]
I very much appreciate the recommendation. Have you also seen Effective defense and ISIL? Someone gave me a star, "Quality resource: this resource is a featured learning resource," for that one. It received 115 views in the past 90 days vs. 61 for "Effective defense." "Winning the War on Terror" has already received 138 pageviews in the first 2 days! But I'm still polishing it. It should be ready in a week at the most, I think. DavidMCEddy (discusscontribs) 23:45, 13 March 2017 (UTC)[reply]

People's Agenda[edit source]

Hello. I'm awfully sorry, I should have left redirects for them but I was too focussed on linking the various pages together. Green Giant (discusscontribs) 07:12, 11 May 2017 (UTC)[reply]

@Green Giant: Thank you for your work. What are the benefits of the changes you made? Those changes look like they were made to conform to a standard, but I'm not familiar with that standard. How can I find documentation on that standard? DavidMCEddy (discusscontribs) 11:38, 11 May 2017 (UTC)[reply]
I moved the central page from a category to a resource page. The individual pages were renamed to become subpages of the central page. This enables them to be linked back to People's Agenda automatically. This creates a hierarchy and makes it a little easier to navigate and find topics (in addition to having a category). The benefits of this are described at WV:Subpages and mw:Help:Subpages Then I've added {{CourseCat}} to each subpage, which automatically categorises them in the category that matches the main page (the bit before the slash in the page name). Then I've reformatted your main page to include both a search box (specifically for these subpages) and a dynamic page list at the lower half of the page - it lists all the pages in the related category without having to manually list them. One of the settings in that list is "offset=1" which stops it listing the main page (a circular situation). If I've missed anything please ask. Also if you want ping someone, you can do so by typing the following literally: {{ping|Green Giant}} but obviously changing the name to the relevant user. Green Giant (discusscontribs) 14:26, 11 May 2017 (UTC)[reply]
Great. Thanks. DavidMCEddy (discusscontribs) 14:56, 11 May 2017 (UTC)[reply]

Winning the War on Terror: Essay or research?[edit source]

I just noticed the discussion at User talk:Ijon. Ijon is active on other projects, but hasn't been active here. From my perspective, the move was inappropriate. I've moved the content back to Winning the War on Terror and relabeled it as an essay rather than research. Perhaps that will help others accept it as is. It's good work. Please continue your efforts. Let me know if you have any questions. -- Dave Braunschweig (discusscontribs) 13:40, 4 July 2017 (UTC)[reply]

Thanks.
Where should I look to understand the categorization system and what ljon was trying to do with that move?
I put it in with other things.
Someone else moved a collection of essays I started on into a subdirectory called "People's Agenda". How can I understand the rules you are trying to create and enforce with this kind of effort?
What's the difference between an "essay" and "research"?
Wikiversity exists to support research. It's not just a free blog site, right? I think my "essays" or "research" are appropriate for Wikiversity, because they are intended to be research, inviting others to collaborate as they see fit. So far, that hasn't happened, but that doesn't mean it won't.
DavidMCEddy (discusscontribs) 14:01, 4 July 2017 (UTC)[reply]
Wikiversity is the only Wikimedia project that allows original research / essays. When users from Wikipedia see original research, their immediate reaction is that it doesn't belong in main space, and they want to move it to user space instead. Whether or not we do that here depends on the quality of the article. Sometimes we move it, sometimes we make it a subpage of an introductory article, sometimes it stands alone. If anyone else objects, we could create a Terrorism or other introductory learning project to add this under. Category didn't make any difference in this case. It was the the content, not the labeling.
People's Agenda appears to have been an appropriate move. The easiest way to recognize this one is your own description of "collection of essays". As a collection of essays, they belong together rather than as separate pages. Wikipedia is 5 million individual articles, without duplication. Wikiversity is a collection of learning projects with subpages for each, allowing different courses and authors to present their own perspectives. Many users may want to create content about LGBT rights. Having these essays in a People's Agenda learning project allows you to present your ideas without disruption from or disruption to others. The other approach would be to create a learning project on LGBT rights, but then the essay would be one of a number of subpages of different essays on the subject.
Research should ultimately be supported by peer review. Essays provide a personal perspective. Whether either label or both is more appropriate depends on who's complaining about the content. In this case, ljon complained about it being an essay, so I thought labeling it as an essay rather than as research might help future readers. It also states in the introduction that it is an essay, so it is probably better labeled as an essay rather than as research, but both tags could be applied if you prefer.
Wikiversity exists to support free learning materials and activities. That includes research. What you have created is appropriate for Wikiversity, but it isn't currently designed for collaboration. Winning the War on Terror is informative, but does not appear to be participatory in nature. It's a wall of text too long to provide an easy entry point for engagement. When you are ready for others to participate, you might consider breaking it down into subpages and providing ways for others to engage. For example, Terrorism is minuscule as a cause of death could be a subpage with background information and then instructions on how users might do their own research into this subject to extend the discussion. But that would be a different approach than the content you have. You would need to decide whether your goal is to inform or to provide an environment where others learn with you.
Dave Braunschweig (discusscontribs) 16:08, 4 July 2017 (UTC)[reply]

Winning the War on Terror: Announce on main page news?[edit source]

Hi DavidMCEddy!

Your essay Winning the War on Terror appears well-developed and ready for learners! Would you like to have it announced on our Main Page News? --Marshallsumter (discusscontribs) 18:31, 9 July 2017 (UTC)[reply]

Thanks. Of course, I'd love to have the additional exposure.
I have one question in that regard: In the immediately preceding section, "Winning the War on Terror: Essay or research?", user:Dave Braunschweig suggests ways this could be reorganized to make it more inviting for others to contribute.
What might be appropriate to do in that regard? In the announcement invite others to review it from that perspective? Or just note that it's available and leave it at that?
Unfortunately, I have very little time to even think about that until after wm2017:Wikimania2017, August 9-13. I've proposed a "Birds of a feather" session for that conference on ""Building Wikinews into the premier news site worldwide".
In that regard, I've been developing a proposal for "Everyone's Favorite News Site" under n:Wikinews:Water cooler/policy. Do you think it might make sense to move (or copy) that discussion into Wikiversity, where the ideas could be developed more in the format of a proposal with notes at the end? The software settings under n:Wikinews:Water cooler/policy make it seem difficult or inappropriate to include references in a sensible way: It's more like a "user talk" page, where notes appear at the end, separated from the text to which they belong, possibly by completely unrelated material.
I'm also considering a similar use of Wikiversity to develop a proposal for something completely different: The recent useR!2017 conference in Brussels included a session soliciting input on how to make the R (programming language) easier to use.
What do you think about posting an article to Wikiversity that would, if we do it right, would evolve into a proposal for work in that area that might be funded by some granting agency like the R Foundation? DavidMCEddy (discusscontribs) 21:52, 9 July 2017 (UTC)[reply]






Sure. That would be great

Everyone's favorite news site[edit source]

Hi DavidMCEddy!

Your journalism resource Everyone's favorite news site appears to be well-developed and ready for learners and participants! Would you like to have it announced on our Main Page News? --Marshallsumter (discusscontribs) 02:08, 29 August 2017 (UTC)[reply]

Net neutrality and 'Restoring Internet freedom'[edit source]

Hi DavidMCEddy!

Your resource Net neutrality and 'Restoring Internet freedom' appears to be ready for learners! Would you like to have it announced on our Main Page News? --Marshallsumter (discusscontribs) 20:56, 30 October 2017 (UTC)[reply]

That would be great. Thanks. DavidMCEddy (discusscontribs) 21:33, 30 October 2017 (UTC)[reply]

Searching R Packages[edit source]

Hi DavidMCEddy!

Your participatory research project Searching R Packages appears well-developed and ready! Would you like to have it announced on our Main Page News? --Marshallsumter (discusscontribs) 16:44, 6 February 2018 (UTC)[reply]

Please. Thanks. And please excuse my tardiness in responding.
What do you think about also featuring Draft Proposal for improving the ability of R users to search R packages?
These two are companion pieces: I wrote the first to support writing the second, and I wrote the second to try to solicit collaborators to actually do the work outlined in the second.
DavidMCEddy (discusscontribs) 01:22, 27 February 2018 (UTC)[reply]

Media and politics[edit source]

Hi DavidMCEddy!

Your research essay Media and politics appears to be well-developed and ready for learners! Would you like to have it announced on our Main Page News? --Marshallsumter (discusscontribs) 22:39, 26 February 2018 (UTC)[reply]

I actually have three essays with "Media and *" titles under Category:Freedom and abundance: Media and corruption, Media and politics, and Media and taxes. Over the past 90 days, the first has averaged 20 views per day; the latter two have averaged less then on a day each -- roughly one per day together.
I need to review and probably merge them -- update and expand the one that gets the most traffic and redirect the others to that one. However, I also need to review and prioritize that.
Thanks very much for your interest and support. DavidMCEddy (discusscontribs) 03:56, 27 February 2018 (UTC)[reply]

hello[edit source]

Dear Mr David. I will like to have a short impered conversation with you about me and you doing some projects together, If you don't mind kindly get back to on my Facebook acct so we can talk much better you can fine on Adam Omar (Nickname) Time. I will be waiting to get a response from you. Thanks Realtimesongs (discusscontribs) 18:26, 3 April 2018 (UTC)[reply]

@Realtimesongs: What do you have in mind?
I assume you are familiar with my work on Everyone's favorite news site and more generally? DavidMCEddy (discusscontribs) 20:31, 3 April 2018 (UTC)[reply]

Media and corruption[edit source]

Hi DavidMCEddy!

Your original research resource Media and corruption appears well-developed and ready for learners! Would you like to have it announced on our Main Page News? --Marshallsumter (discusscontribs) 04:00, 24 July 2018 (UTC)[reply]

I'd be pleased and honored.
However, you should know that I have other plans for that article:
  1. Most importantly, I'm thinking of creating a Wikipedia article by the same title, which would consist of the summary of relevant research in this article -- with a link to the Wikiversity article.
  2. If the extract on Wikipedia is not speedily deleted, I plan to condense the background discussion to a very brief section with a link to the Wikipedia article. The remaining material on Wikiversity would primarily be a discussion of policy implications that might be rejected on Wikipedia as not being "encyclopedic".
  3. I also have Wikiversity articles on "Media and politics" and "Media and taxes", which I planned to migrate into the current article on "Media and corruption" and replace with a redirect.
My rationale for this is that the article on "Media and corruption" has been averaging around 14 page views per day, while the other two have been attracting maybe 1. I don't know, but I think an article on this subject would likely attract more views on Wikipedia than Wikiversity, though I won't know until I try. Second, I think the other two "Media and *" articles have largely been overlooked, because people search for "media and corruption" and find it, but they don't think of searching for "media and taxes" nor "media and politics". However, that remains to be seen. DavidMCEddy (discusscontribs) 18:56, 24 July 2018 (UTC)[reply]

The Great American Paradox[edit source]

Hi DavidMCEddy!

Your original research project The Great American Paradox appears well-developed and ready for learners and participants! Would you like to have it announced on our Main Page News? --Marshallsumter (discusscontribs) 02:53, 3 December 2018 (UTC)[reply]

Can you wait a few days?
I'm revising it. It should be ready in couple of days.
I'm actually making a movie from it. I'm thinking of posting the movie to Wikimedia Commons and including that in this article, similar to Communication’s Challenge to Democracy.
I should have that done this week.
Thanks, DavidMCEddy (discusscontribs) 03:09, 3 December 2018 (UTC)[reply]
@Marshallsumter: I just posted the revisions I was working on. I hope to have a movie of this to add by next Monday. Maybe wait until then to announce it? Thanks, DavidMCEddy (discusscontribs) 19:50, 4 December 2018 (UTC)[reply]
@Marshallsumter: It's now ready to be "announced on our Main Page News", if that's convenient for you.
Yesterday, I posted the video to Wikimedia Commons and linked to it from Wikiversity. That video refers to "en4j.org/2", which I just got working within the past hour.
If you have any concerns or suggestions for improvement with any of this, please let me know. I believe, based so far on no evidence, that I might be able to reach a substantially larger audience with a modest additional effort by making a video based on a research article, have that video end by asking peopled to go to something like "en4j.org/2" and comment on Wikiversity, etc.
Thanks for your support and interest in this work. DavidMCEddy (discusscontribs) 21:06, 10 December 2018 (UTC)[reply]

Time to extinction of civilization[edit source]

Hi DavidMCEddy!

Your original research essay Time to extinction of civilization appears well-developed and ready for learners and participants! Would you like to have it announced on our Main Page News? --Marshallsumter (discusscontribs) 19:14, 30 December 2018 (UTC)[reply]

Thanks for the vote of confidence. Permit me to share with you some of my plans for this:
  1. I hope to recruit collaborators to refine this article and submit some version of this piece to one or more refereed academic journals. I've already attempted to contact some, and I have others in mind.
  2. I also want to add estimates of the uncertainty in the estimate of time to the next crisis and Monte Carlo simulations to improve the estimates of uncertainty summarized in the paper.
Adding estimates of uncertainty in the "time to the next crisis" shouldn't take too long. Adding a Monte Carlo might take a bit longer, though not much.
Recruiting collaborators and getting them involved could take much longer.
However, it's fine with me if you'd like to announce it now. Or we could wait another week or two to see how far I get with #2.
DavidMCEddy (discusscontribs) 19:28, 30 December 2018 (UTC)[reply]

Empty Citation[edit source]

See Category:Pages with empty citations. Citation 12 references an invalid Q number. -- Dave Braunschweig (discusscontribs) 02:52, 1 December 2020 (UTC)[reply]

Thanks. Fixed.
Might it be feasible for you to find out how long that was listed as an invalid Q number?
It's probably not important, but I'm pretty sure the reference was real when added it to that article on 2018-12-04.
I cannot imagine that I would have cited an invalid Q number, planning to fix it later. If the "invalid Q number" complaint dates from that time, then maybe I did.
However, it if it's more recent, then there would seem to be something that destroyed that Wikidata item.
??? Thanks again, DavidMCEddy (discusscontribs) 03:47, 1 December 2020 (UTC)[reply]
Sorry, I'm not aware of any way to see category membership dates. You'd need to check with a Wikidata admin regarding history of that Q number. -- Dave Braunschweig (discusscontribs) 14:50, 1 December 2020 (UTC)[reply]
Users Randykitty and Jianhui67 decided it was "Completely unreferenced and deleted it at least 15 months before it appeared on Category:Pages with empty citations. See the thread on this I initiated at your suggestion: Wikidata:Wikidata:Project chat:Q59318277 seems to have disappeared.
Thanks for suggesting Wikidata admins. I've not seen replies to all of my posts there, but this actually generated something useful. DavidMCEddy (discusscontribs) 21:23, 1 December 2020 (UTC)[reply]

wikiquote[edit source]

Hi DavidMCEddy, I happened to see your posting regarding google summer of code. Do you happen to know if an intern may be useful for other wmf-wikis such as wikiquote for example? Thanks in advance, Ottawahitech (discusscontribs) 20:14, 29 January 2021 (UTC)[reply]

I don't know more about Google's summer of code than that it exists and that link provided one way to find out more. If you have an idea for a project, I suggest you do more research to determine what their criteria are, etc. DavidMCEddy (discusscontribs) 20:38, 29 January 2021 (UTC)[reply]

Wikiversity Grant Proposal: Call for Support or Feedback[edit source]

Hi DavidMCEddy!

A collaborator (bwsulliv) and I have submitted a grant proposal for a new Wikiversity learning project, which we are calling Eventmath. We aim to help math and statistics teachers to teach their students the mathematical thinking skills needed to make sense of current events. Specifically, the project will pair math lesson plans with news articles (or social media posts, if the goal is to debunk misinformation, for example).

I'm writing to you because we would like to gather more feedback from the Wikiversity community. I noticed that you've made contributions related to statistical forecasting and journalism, and one of the use cases for our project is a quantitative literacy course for journalists, so I thought you might be interested in the proposal. If you could either endorse the proposal or provide constructive feedback, we would really appreciate it!

Thank you for your consideration.

Greg at Higher Math Help (discusscontribs) 20:11, 19 February 2021 (UTC)[reply]

@Greg at Higher Math Help: Done. Thanks.
FYI, if you could use help with R (programming language), I might be able to help, including posting data analyses with "R Markdown" script files that make the analyses 100% reproducible, because the methodology is 100% documented and transparent.
Are you aware of Investigative Reporters and Editors and their National Institute for Computer-Assisted Reporting (NICAR) at the University of Missouri in Columbia? It is my understanding that they do a lot with R and Python. If you haven't already, you might consider contacting them. If you make the right connections with them, they might provide you with a steady stream of great examples. DavidMCEddy (discusscontribs) 21:12, 19 February 2021 (UTC)[reply]
@DavidMCEddy: Wow, thank you so much! We would love it if you could help. We already have mentioned the inclusion of code in supplementary materials, and our first example entry does include a simple spreadsheet analysis. There is no limit to the level of math or the types of technical skills we can support, so R analyses are entirely welcome! Our existing template allows for an Assumed Knowledge section, so if knowledge of R is required, this would simply be noted there. If you would like to add yourself as an interested contributor in the proposal itself, we would be happy to have you!
Also, that is fantastic information regarding NICAR. I had planned to contact the University of Missouri journalism department (I know of that department's reputation, and they came up again in some of my research). However, I was unaware of NICAR, which sounds very relevant. I just added a link to our proposal indicating we will contact the Investigative Reporters and Editors organization regarding a collaboration with NICAR. Thanks again!
@Greg at Higher Math Help: Have you seen "Time to extinction of civilization", "Forecasting nuclear proliferation", and "US Gross Domestic Product (GDP) per capita"? All contain one or two "Companion R Markdown vignettes" as appendices.
The statistics in "US Gross Domestic Product (GDP) per capita" are relatively introductory. The analysis considers alternative epocs in the macroeconomic history of the US since 1790. First, it concludes that the variance in the annual increments in log(real GDP per capita) is not constant; in fact, the variability since 1947 has been the lowest in US history. Secondly, it concludes that the 6-epoc models considered are not statistically significantly better than the 5-epoc model.
"Time to extinction of civilization" illustrates combining subjective Bayesian logic with parametric analysis, then uses Monte Carlo to estimate a probability distribution of the time to extinction of civilization, provided the assumptions are credible.
"Forecasting nuclear proliferation" does Bayesian Model Averaging of Poisson regression of the probability distribution of the number of new nuclear weapon states each year. It then uses that to do Monte Carlo simulations of the future.
The "US Gross Domestic Product (GDP) per capita" might be useful toward the end of a solid introductory statistics class. "Forecasting nuclear proliferation" could be used in a class that discusses generalized linear model. "Time to extinction of civilization" might be used with an advanced study of subjective probabilities.
I'm not going to take the lead in preparing simpler examples, but I'm happy to help others.
I'm currently working on developing discrete time state-space models of real GDP per capita in different countries using data in the Penn World Table and the Maddison Project to discuss the macroeconomic history of all the world's economies for which I can get the data. I have two primary goals:
  1. Make it easy for people to understand the different variables that have been developed to explain this history.
  2. Make it easy for people to get plots of any part of this macroeconomic history to illustrate whatever they want, as long as it's consistent with the available data and model fits. This includes making it easy to produce plots showing the macroeconomic history of any given country that could be used to illustrate Wikipedia articles about that country. This also includes making it all available under the Creative Commons Attribution-ShareAlike.
This work is going slowly. I think I know how to do it all in outline, but I'm currently struggling with some of the details. DavidMCEddy (discusscontribs) 23:17, 19 February 2021 (UTC)[reply]
@DavidMCEddy: I haven't read through the resources you've listed yet. Thanks for explaining your work! I totally understand that you have your own projects, so no worries about not taking the lead on simpler examples. We welcome any type of help.
Out of curiosity, are you comfortable in Python as well? If you have a preference for R, I'm curious to know if you have a rationale. I could see either being used for a course on computing and quantitative methods for investigative journalists, which could potentially be based on Eventmath materials. Right now, we are proposing to recruit math educators, which may include statisticians who are familiar with Python or R.
Actually, this discussion is leading me to wonder if we should widen our campaign and spend more time reaching out to data scientists; although, perhaps it's better to avoid scope creep, at least at this early stage. In any case, I'm open to discussing new ideas.
@Greg at Higher Math Help: Regarding R vs. Python, Matlab, etc.
  1. I started getting into S-Plus around 1990, around the time that Python was invented. I'm a statistician, and I didn't hear about Python until much later.
  2. In 2010 I made a presentation for a chapter of the Association for Computing Machinery in Santa Clara County, CA, about Package development processes, comparing R with other languages, and inviting collaborators to help produce a more complete comparison. I claimed that the Comprehensive R Archive Network (CRAN) gave R a very valuable edge over other languages, outlined in a table in a section on selected repositories in the Wikipedia article on software repository. Every package on CRAN has documentation, which is supposed to include unit tests that are run routinely to make sure that changes on one package don't break other things. Other languages may have this also, but I don't know them -- and haven't found it worth my time to investigative.
  3. I've used Python, but I have not found it to be as easy to use as R -- and I so far have failed to find a repository for contributed packages that compares with CRAN. This could be my deficiencies. I have not had much interaction with other Python users. I've written some Python code to do real time data processing that I could not figure out how to do in R. Around 2010 I started with Python code that someone else wrote and expanded it to do substantially more than the toy code I was provided. Maybe 2-3 years ago I wrote python code to try to detect outages in transmissions from a radio station. The intermittent outage problem was solved by other means before I got my Python to a point that it was actually useful.
My bottom line: I recommend R over Python for multiple reasons. First, I think R is structurally better for statistical analyses than Python; Python is more general purpose Second, I believe that CRAN provides software that's easier to access and use than anything I've seen available in Python. However, both of these things could be just my limited use of Python. Third, the package development process for R and CRAN is relatively easy to learn and use. There may be comparable processes for other languages. However, the Wikipedia article on "Software repository" has averaged 300 views per day since 2015-07-01. If Python or some other language had a package development and sharing process comparable to CRAN, I would think that someone would have added more info to the table I created there. I know that absence of evidence is not evidence of absence, but ... .
I'm not sure I know the difference between statisticians and data scientists. The former may know more math and less data analysis, but they are fairly close.
I said years ago that R is increasingly the language of choice for people engaged in new statistical algorithm development. I think that's still true, though I could be wrong. DavidMCEddy (discusscontribs) 03:08, 20 February 2021 (UTC)[reply]
@DavidMCEddy: Thank you for your detailed overview! Ultimately, I think we would welcome data analysis using any language, especially open-source languages. However, it's useful to hear your expertise. I have experience with Python but have not used it for any serious statistical purposes, and I haven't used R myself, so the overview you've provided is informative.
Thanks again!
--Greg at Higher Math Help (discusscontribs) 22:08, 22 February 2021 (UTC)[reply]
@Greg at Higher Math Help: Might you have time to review the table of w:Software repository#Selected repositories, especially to confirm (or correct) the listing of "PyPI" as the primary repository for sharing Python and adding "Autochecks" if appropriate.
I occasionally get "nastygrams" from CRAN complaining that a package for which I'm the maintainer generates errors (or warnings) on at least one of the platforms tested. I can think of three things that have caused this: (1) The people managing the R language develop and / or CRAN added new autochecks for, e.g., references in a function to undefined variables (most likely a latent bug from misspelling a variable name). (2) Someone changed a package I used in a way that broke my code. (3) The basic language itself was changed in a way that broke my code. I'm not happy to get those emails. However, long term, they have made R more user friendly and reliable. DavidMCEddy (discusscontribs) 22:37, 22 February 2021 (UTC)[reply]
@DavidMCEddy: Alas, I am not well versed in the Python ecosystem. I've developed a physics simulation with Python, I've done some basic Python coursework online, and I've used Python while tutoring a data science student on basic combinatorics (and maybe some probability), but my knowledge of libraries like NumPy, Matplotlib, pandas, and TensorFlow is mostly or all secondhand. Most recently, I've been working in JavaScript (primarily p5.js and associated libraries).
However, I did a little digging and found some resources that may help. According to the python.org glossary, "PyPI is the default Package Index for the Python community." Also according to Python.org, PyPI migrated to a new platform, and "As of April 13th, 2018, pypi.org is the URL for PyPI." More updated information is available in the PyPI Wikipedia article, but I see that's already linked to from the article on software repositories that you've asked about. Based on these sources, it does seem that PyPI is the primary repository. I haven't tried looking into "Autochecks."
For information on publishing to PyPI, managing Python packages, and Python packaging and installation tools, the following resources may be helpful:
I know this may not be new to you, but I hope some of it helps!
--Greg at Higher Math Help (discusscontribs) 05:06, 23 February 2021 (UTC)[reply]

It is new to me, but it also indirectly supports my claim of the superiority of CRAN over PyPI: I could be mistaken, but I believe that most R users learn about CRAN fairly early in learning R. However, this discussion inspires me to ask a Professor of Computer Information Systems about this:

Software repositories[edit source]

@Dave Braunschweig: Might you have time and interest to review the Wikipedia articles on Package development process and software repository, focusing especially on the table in the section of the latter on selected repositories?

This may not be worth your time. I created that table and the "Package development process" article with the help of a friend in 2010 as part of preparing for a talk I gave on this subject at that time for a Silicon Valley chapter of the Association for Computing Machinery. I believe that the package development process for R combined with how CRAN is managed is a vital part of what gives R its power and general utility. This actually wasn't my idea: I had heard in 2007 that Martin Maechler, the third person to join the R project, convinced the two originators to release the code under the GPL, turn over further development to an international team, and support contributed packages with what became CRAN.

I've since looked occasionally for someone who might know enough to actually give me a professional opinion on this conjecture. I'd be pleased to have your thoughts, even if it's just, "This doesn't seem worth my time." Thanks, DavidMCEddy (discusscontribs)

I briefly reviewed this. My initial reaction was that it needs something to differentiate it from Wikipedia:Repository (version control), even if it's just a brief mention. I suspect many / most / all of the packages in a given software repository are maintained in version control repositories. I'm not familiar with the history of R. CRAN is mentioned at Wikipedia:CRAN (R programming language). That mentions different names than the ones in the R history paragraph. I would think someone is still around who would know the answer. I've contacted people directly for things like this from time to time. I still have an email somewhere from Wikipedia:Ward Christensen where I asked him a question about how a protocol worked for one of my network communications courses. Really impressed the professor that I got the answer from the guy who wrote the protocol. -- Dave Braunschweig (discusscontribs) 02:57, 24 February 2021 (UTC)[reply]

I'd like your opinion on a study[edit source]

Hello there. I notice you're an active editor on Wikiversity, and if you have time I'd like to get your opinion on something related to statistics. Some time ago I had a dispute on Wikipedia at the talk page for an article on concealed carry. In particular, I did not want the Wikipedia page to cite a Washington Post article that misrepresented the findings of a study on concealed carry laws. The research paper that WaPo cited was "Right-to-Carry Laws and Violent Crime: A Comprehensive Assessment Using Panel Data and a State-Level Synthetic Control Analysis". The WaPo article stated incorrectly that many states saw an increase in violent crime following the introduction of right-to-carry laws. Actually, crime continued to decrease (in most cases) but for many states it was higher than for their corresponding synthetic controls. Anyway, I want to know what you think of the study itself. From what I understand, they fit a control model using data obtained from other states, prior to the introduction of RTC laws in the state of interest. They get a set of weights with which to combine data from those other states, in order to model how crime rates in the treated unit might look had it not adopted RTC. I noticed that their method was unable to find good-fitting synthetic controls for the states Maine, Montana, Minnesota and South Dakota, where their model suggested that shall-issue was a cause of decreasing violence. While they give an overview of how the synthetic control model works, they do not seem to provide many details or the data/settings they used to obtain their synthetic controls (or if they do, I must be missing it. There is a 126 page copy that includes an appendix with a few more details, e.g. data sources but the exact optimization process for fitting the synthetic controls is still not clear to me.). I'm not an expert in statistics, so it's difficult for me to evaluate the quality of this study one way or another, but I can't help being skeptical. What is your opinion on it? AP295 (discusscontribs) 21:59, 19 March 2021 (UTC)[reply]

Also, in some states where their synthetic control had a very good fit prior to the introduction of RTC in the treated unit (e.g. PN in figure 6 of the shorter version without the appendix), the treated unit seems to roughly continue the trend it was already on before RTC was introduced in that state, while the synthetic control diverges at that point. To me, this could just as easily suggest that the introduction of RTC laws had negligible impact, and that the synthetic control model was over-fit, but again statistics isn't my specialty and I don't have their data/software/settings. I realize the whole point of using synthetic controls is to have "counterfactual" units to compare with rather than simply looking before and after, but if RTC was a significant cause of increased violence in the states that adopted it, I'd still expect the treated unit to buck past trends after its RTC adoption more strongly/often than the control unit, which is comprised of data from states where there was no such policy change. After the spiel they give in section II, it's clear the sort of results they were after and it's hard for me to feel convinced by their results. (and much less by WaPo's gross misunderstanding and misrepresentation of the them) The general idea of using "synthetic" controls seems pretty clever but it's not hard to imagine how such a model could be tuned to favor specific results. AP295 (discusscontribs) 15:05, 20 March 2021 (UTC)[reply]

One last thought I had is that RTC itself does not necessarily imply (though it may be associated with in practice) lax or absent standards/vetting of applicants. As far as I know, it just means that if you meet the state's requirements, then issuing authorities can't simply deny you a permit at their own discretion. Conflating RTC with licensing/vetting standards themselves seems a bit underhanded. In many areas without RTC (but certainly not all), issuing authorities may abuse this prerogative and do not issue permits in most cases, even for applicants of good character and with a clean record. On the other hand, many counties in a given state could be shall-issue in practice, the exceptions often being urban areas with high crime where the authorities might be less inclined to issue someone a permit. In that case, passing a statewide shall-issue policy would only result in more permits being issued for high-crime areas, and the additional permits would be issued to people who would not have gotten them otherwise. But to say that issuing authorities should be given carte blanche to deny any applicant because introducing statewide RTC subsequently resulted in higher crime compared to some dubious synthetic control model is quite a stretch. I wouldn't even be skeptical if they achieved a reasonably consistent fit for the synthetic controls but it seems to vary quite a bit and I suspect there exits a better set of explanatory variables that one could choose to examine. Sorry for the digression, it just worries me that many people are losing sight of why we have such liberties, and there seems to be quite a strong bias on Wikimedia and in education. AP295 (discusscontribs) 19:09, 20 March 2021 (UTC)[reply]

Could you please provide complete citations to the complete article in WaPo and the research study you cite? (I would prefer that you cite Wikidata entries for both, but I will accept a complete citation in any form, preferably something that does not require me to pay for something that is behind a paywall to see it.) Thanks, DavidMCEddy (discusscontribs) 01:56, 21 March 2021 (UTC)[reply]
Here's a direct link to the study: https://www.nber.org/system/files/working_papers/w23510/w23510.pdf
The WaPo article: http://web.archive.org/web/20200815052813if_/https://www.washingtonpost.com/news/wonk/wp/2017/10/19/3-million-americans-carry-loaded-handguns-with-them-every-single-day-study-finds/
Sorry if I went off on a tangent, I just wanted your opinion on the quality of the study itself. AP295 (discusscontribs) 03:32, 21 March 2021 (UTC)[reply]
Thanks. From a cursory review, it looks to me like the work seem reasonably solid methodologically. It starts by citing important historical work and seems to have been written and reviewed by reputable scholars with reputable institutions. However, empirical evidence is never complete, and one study cannot be considered completely authoritative by itself.
Two other observations:
  1. They claim a 13-15 percent increase in "violent crime rates ten years after adoption". That effect is so small, it suggests to me that people concerned about these issues should look elsewhere for ways to change public policy to reduce violent crime rates. For example, Youth Ambassadors, Wikidata Q56879668 is a program in Kansas City that pays teens with terrible home lives minimum wage to stay after school and come in part time on Saturdays to do useful things. This reduces their opportunities to get into trouble and increases their motivation to stay out of trouble. I heard an unconfirmed report that they had nearly eliminated high school dropouts and crime among the youth in their program. I was told last year that a serious evaluation of that program was in progress. If the claims I heard are supported by that serious evaluation, we could cut teenage crime by 10 - 25 percent per year, with a cumulative effect of cutting crime at least by 40 percent in five years, by funding such programs at 100 percent of the need. AND the rate of economic growth would also increase, according to Stanford Economist w:Eric Hanushek#Economic impact of education. Hanuschek said that any such innovation would likely generate so much additional economic growth -- money from people able to produce more with less that they could not do without that education -- that they'd effectively get the program for free in perpetuity after the initial up-front investment. And that ignores the benefits of reduced crime, etc.
  2. Their crime rate data come from Uniform Crime Reports (UCR), which is notoriously poor, because it reflects what people are willing to report AND what the police are willing to actually record. The National Crime Victimization Survey (NCVS) is much better but harder to get: It asks random samples of thousands of households in the US about whether any members of the household had been a victim of one of six major crimes. Critics argue that (a) no verification is done of reported crime, and (b) domestic violence is almost certainly under reported. Nevertheless, a 13-15 percent change in UCR reports over a decade could easily be due to changes in reporting rather than changes in crime. After w:George Floyd was killed by Minneapolis police last year, the City Council voted to put a referendum on abolishing the police on the November ballot. Before that election the rate of reported crime increased:
Minneapolis City Council Surprised by Rise in Violence After Moving To Abolish Police Department.
I don't know if the increase in reported crime was accompanied by an increase in actual crime.
Thanks again. This is useful to know, especially that the effect is that small. DavidMCEddy (discusscontribs) 05:22, 21 March 2021 (UTC)[reply]
Thank you for having a look. My concerns were more about the synthetic control model itself and the fact that statewide RTC adoption was the treatment they chose to study. The parameters for the synthetic control were fit using data from states that did not adopt RTC, prior to RTC in the treated state. Consider figure 114 on page 96. The treated unit is on a clear upward trend immediately prior to RTC, and the synthetic control fits reasonably well right up until that point. However, it's the synthetic control that seems to respond strongly to to the treatment, while the treated unit continues its upward trend. This makes me doubt the predictive performance of their synthetic control, at least in this case. Additionally, I wasn't able to find the data/details they used to fit the synthetic control parameters aside from the footnote on page 28. Secondly, RTC/shall-issue is not an easing of gun control per se, and I don't think many people understand the distinction. It just removes the discretionary power of judges and other issuing authorities to deny applicants who meet whatever standards the state requires. There are several counties in my state where the issuing authorities are pro-gun-control and generally refuse to issue permits to anyone, and many other counties that generally will issue permits to applicants with a clean record. The process of applying is complicated, and many would say it's intentionally difficult. Conversely, I can see how "legitimate" profiling of applicants might suppress crime in other states/areas. The point is, RTC is not what the study ought to be looking at or advocating against. Generally, permit holders are not the source of most "gun violence", and the collection of anecdotes in section II does not change that fact. AP295 (discusscontribs) 16:05, 21 March 2021 (UTC)[reply]
Your analysis sounds sensible. See also w:Gun control#Studies. DavidMCEddy (discusscontribs) 16:20, 21 March 2021 (UTC)[reply]

Replace with new section title[edit source]

@Antandrus: Am I correct that your notice of speedily deletion pertains only to Talk:Effective defense and ISIL, and not to the article on Effective defense and ISIL itself?

It looks to me like the material you deleted was vandalism and deserved to be deleted. I didn't do it, because I've been overwhelmed with other work, and this looked like it may not be worth my time to worry about. However, if you can fix it without deleting the main article, that seems useful to me. Thanks, DavidMCEddy (discusscontribs) 02:15, 2 May 2021 (UTC)[reply]

Greetings -- yes, it was vandalism by a Foundation-banned editor, just on the talk page. Since he created the page, instead of just reversing his edit I put a speedy delete tag on it. That won't affect your main article, and will more permanently remove his spam link/vandalism. Antandrus (discusscontribs) 02:22, 2 May 2021 (UTC)[reply]