Jump to content

User talk:Tupui

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Welcome!

[edit]

Hi Tupui! I noticed your contributions and wanted to welcome you to the Wikipedia community. I hope you like it here and decide to stay.

As you get started, you may find this short tutorial helpful:

Learn more about editing

Alternatively, the contributing to Wikipedia page covers the same topics.

If you have any questions, we have a friendly space where experienced editors can help you here:

Get help at the Teahouse

If you are not sure where to help out, you can find a task here:

Volunteer at the Task Center

Happy editing! You should have gotten one of these already!  — SMcCandlish ¢ 😼  15:23, 17 February 2024 (UTC)[reply]

Sampling, etc.

[edit]

This is really getting too long and detailed to continue at WP:AN, so I'm moving to user talk.

What you said over there is a common complaint and misunderstanding from subject-matter specialists, especially academics: we experts are not welcomed to contribute our knowledge. ... it’s a chance for Wikipedia that we would be willing to spend time to share our knowledge. The problem here is basically in thinking of yourself as a reliable source. Academics and other experts are used to being treated with subject-specific deference and consulted and listened to, but on WP there is no way to be sure anyone is who they claim they are, and even if we could be sure, they still would not be a source, of any kind for anything. A source (in our sense, not the journalistic or intelligence-gathering sense) is a publication, not a person. And for our purposes, the only ones that count for most things are secondary sources, not primary research (which is often enough wrong, and much more often rapidly surpassed).

WP does in fact badly need subject-matter experts in many fields, but we need them to be Wikipedians (builders of the encyclopedia, following its rule set), not promoters of themselves "as sources" and not as promoters of novel ideas in a field to which they are deeply personally bound. We need experts to help find and evalute (give due weight to) such sources and weed out fringe material. Experts often have a very clear professional understanding, for example, of which ideas are taken seriously in a field and why, which are in dispute, and which are generally rejected, and how the concepts interrelated in various often complicated ways. These are matters that may be difficult for non-experts to discern (in large part because most published academic material is paywalled). But even that kind of internal analysis of the reputability of the source material is in turn dependent on sources about sources; even if you know that so-and-so's theory is considered ironclad, while some other person's is laughed at, you still have to be able to prove it, not just assert it, and you probably have better access to the materials with which to do so that most other people.

The second half of what you said in the same bit, random folks, which clearly have no clue about a subject, are more welcomed to share their non existing knowledge, is not actually true at all, for the very same reason: we don't want anyone to try to "share their knowledge"; Wikipedia is not a classroom or any other form of instruction. Rather, random folks and experts alike are expected to follow policy and rely on reliable (and mostly secondary) sources. WP gets the facts from the sources, not from the editor. We have an essay at WP:Randy in Boise, about the problem of no-nothing people who try to act as if they are experts, and this can sometimes be a problem, but that is not the problem here. MrOllie is expecting that our sourcing policies be followed, and they're not being followed. In theory, whether an editor is an expert or not should have no bearing on the content outcome if the same sources are used. But in reality, of course, the value that experts provide here, when they get used to the fact that this is an encyclopedia not an academy or standards body, is in the ability to assess and interpret the sources with professional nuance and clarity, and weigh them with (sometimes again) other such source material, producing encyclopedic summaries that better reflect the professional understanding of the subject. Experts on WP are invaluable, when they play by the WP rules, but become individually problematic and more of a drain than a gain when they act as if sources themselves instead of finders and weighers of sources. (Plus frequent conflict of interest problems.)

The central issue in your and Saltean's dispute with MrOllie really comes down to "the researchers and professionals making new methods and driving the field": This kind of stuff is what is "sexy" in journals, the bleeding edge, but it is not material for encyclopedic coverage, not until the real word has vetted it throughly and it is no longer the vergue but well-centered, and extensively covered in secondary source material. [N]ot that we are trying to add our latest stuff to Wikipedia: Understood, but if there's a dearth of secondary coverage, then even work you did a decade ago is still probably too new, or too obscure, or too something for encyclopedic inclusion. (A lot of academic and technical work is actually too obscure for this publication, even if influential within a particular professional circle. There's nothing wrong with the work, it's just not what an encyclopedia covers.)

[W]hat if there are no secondary independent sources? Does that mean we could not write anything on a topic? Pretty much yes. E.g., if secondary sources indicate there is a theoretical, factual, processual or other dispute within a field, we might use primary sources to indicate the specifics of what each side in the dispute is arguing (per WP:ABOUTSELF, a publication itself is an accurate source for what the publication actually says, though not per WP:AEIS for analysis, evaluation, interpretation, or synthesis of those claims). But if secondary sources don't cover it, then neither set of primary sources and the arguments they make about the matter will be any encyclopedic interest. That doesn't make Wikipedia broken, it just makes it a tertiary work, providing an overview of the central, best-accepted state of the subject (as demonstrable with independent secondary sources); WP is not a journal and not a monograph.

On a side matter: sensitivity analysis is not the same topic as sampling (in the computational modeling sense or otherwise, even if a connected topic), and we do have quite an article on the former. You're seem to be commingling different things here: the desire for a devoted Sampling (computational modeling) article has nothing to do with EU regulations about sensitivity analysis (and the regs don't appear to have anything to do, either, with the exact specifics the pair of you want to write into our material about sampling and its relation to that analysis – that is, "this topic is provably important" does not yield "therefore, I get to write about it to reflect my professional but personal view of it"). If there's some error in the SA article with regard to the CM meaning of sampling, then that is something to raise on the article talk page, and proving it will require reliable secondary sources, not primary research. You might not like it, but some of WP (and some of any detailed publication) is always going to be "wrong" according to someone, on a lot of things, and is always going to be wrong according to a large group of people on a few things, because (for one thing) there is a gap between when facts, theories, methods, etc., appear and when they are subject to enough secondary coverage to be something a publication like this will include. It's the nature of the beast.

By way of analogy, it personally irks me a bit that the potentially ascendant "Celtic from the west" hypothesis (that the Celtic language family developed in Iberia and spread from there, rather than arising in Central Europe, as is the long-term and best-accepted model), has virtually no coverage in Wikipedia, despite being around since at least the early 2000s, and the subject of three major edited volumes and a lot of supplementary material, and general discussion among the interested, and significant (even heated) debate within the relevant specialties. However, it lacks actually reliable secondary coverage like literature reviews (even weaker "pop science" coverage of the subject is basically non-analytical and simply regurgitating what the primary-source papers say, and some academic debates which are themselves primary sources). As far as I can tell, WP doesn't even briefly mention this hypothesis much of anywhere except at John T. Koch and Barry Cunliffe, the biographical articles of the two leading proponents, where they are attributed as ideas propounded by these figures, not as facts that Wikipedia is declaring about language origins, or even as facts of major debates about those origins. The evidence increasingly suggests Cunliffe, Koch, et al. are largely if not entirely correct on most aspects of the new model, but it faces generations of ingrained thinking and has not yet achieved much non-primary coverage, much less a scientific consensus. It just is as it is. I raise this as an example of where Wikipedia may necessarily be "wrong" to someone, and even to a significant and informed body of someones, but that it can't actually be any other way, because our role is not advocating for the truth of a particular doctrine, method, claim, or other thing that a professional/academic/expert sector might want to see promoted here.

I apologize for this being lengthy, but I'm trying to provide reasons and reasoning, explanation and examples, causes and effects, not just a flippant response, which you might get more of at AN.  — SMcCandlish ¢ 😼  15:23, 17 February 2024 (UTC)[reply]

Thank you for taking the time to write this response.
I am not sure how to respond though, here is my attempt.
Andrea is arguably the second most famous person in the sensitivity analysis community for good reasons. From his discoveries he made but also on his efforts to create conferences, summer schools and also his tremendous work on policy making. He lead a EU Join Research Center on the topic and made EU policies. But he is also one of the kindest person I know, always there to share his knowledge anywhere, including on WP since 10 years as he pointed out.
This was my main grip with this whole argumentation. Seeing an editor just dismiss his work without any consideration and then refusing to engage is outrageous to me. I am not even talking about the lack of consideration and mocking that took place while raising a formal complain. Specifically the sandwich part, which to me already is a Code of Conduct violation, at least the ones I practice daily in the Scientific Python community. Again, a joke is only one if everyone is laughing. This whole thing has kept us awake trying to find a solution for days.
To the citation part, my personal explanation of why I do think that using Andrea's work is a sensible thing to do is simply because I honestly did not find better articles on the subject than what he writes. We all agree in the community about that. He is precise and accurate. So sure, we could find some other article which talk about the subject, but they most often time are found to be inferior. As a reader, I personally prefer to see the orifinal thing if it does not contain mistakes as opposed to reading the same material and method section copied a thousand time. I personally don't buy into the secondary/tertiary/best argument here when the whole community would collectively agree that Andrea's book for instance is a paramount reference, all course don't just cite it but invite students to really read it.
Sampling and sensitivity analysis have a close relationship thanks to a certain Sobol' (I actually created my account here to fix a typo in his name, and got myself shoot by an editor reverting everything without blinking an eye...). The most performant method in Quasi-Monte Carlo sampling is the sequence of Sobol' and there is a large body of literature which applies this sampling scheme to be used for calculating the sensitivity indices of Sobol' (yes, the same person.)
All in all, I think this has already gone too far and I feel sorry for Andrea. He just wanted to share something, was shoot without care and now nobody seem to really care and take the time to go over the actually edit and do something about it. If that's just a citation thing, fine restore the page and ask for an edit there. I mean, just engage and respect the time people put into this.
As for me, personally, I am afraid I will just abstaint from contributing at all to Wikipedia moving forward. This is too much work for me especially if people (not you) are not respectfull and inclusive. This alone goes against my values of sharing knowledge, information and making it accessible to everyone.
Thank you for reading this, and I hope Andrea's situation get resolved in a way that will keep him coming and contributing to WP as he's been doing for years.
Pamphile Tupui Roy 213.142.96.4 (talk) 21:27, 19 February 2024 (UTC)[reply]