There are several questions that request analytical solutions of statistical problems, and there are answers that give nicely derived equations using mathematical calculus. These are however not currently distinguished, so I propose the analytical-derivation tag.

Why is it useful?

  • This new tag could complement the existing numerics tag and distinguish between questions that are opinions, collections, applications of well known methods, or numerical approximations – and symbolic, theoretical answers.
  • There are questions that explicitly ask for deriving formulas.
  • Some questions received both analytical and numerical solutions or numerical approximations as answers. Reading through multiple numerical solutions may give the key idea to solving a new problem. The same applies to this tag, reviewing calculus used in statistical problems may give examples, hints or solutions.
  • Some users are outstandingly active in deriving statistical formulas in their answers. Now they only receive reputation scores, but the new tag would honour their efforts with tag-badges, too.

Alternatives: the above or similar goals can be achieved using other similar tag names, possible words include analytical, algebraic, symbolic, formal, mathematical, calculus, derivation, formula, equation.

A few answers to potential criticisms: It is OK to use a relevant tag to couple the question and answer together. This is not intended to be used as a meta tag (tags that do not describe the content of the question, rather some other aspect of the question, like the author’s skill level, or the author’s motivation for asking it, or generally what “kind” of question it is (poll, how-to, etc.)), but where the application of the specific method of algebra/calculus to statistical formulas is a key to get to the answer.

I ask for opinions

  • how the description of this tag should be compactly worded
  • what should it cover and what shouldn't it cover
  • if there is a better expression than analytical-derivation.

StackExchange is not static, so tags evolve even after a good amount of usage, but a good working definition will help launching the new tag.

  • 5
    $\begingroup$ (+1) I appreciate that you finally decided to open a new and fresh thread on that particular issue. $\endgroup$ – chl Apr 29 '11 at 18:56
  • 4
    $\begingroup$ it really would be best if you consult with the community here on meta before retagging 50 questions with a new tag of your own creation. $\endgroup$ – Jeff Atwood Apr 30 '11 at 9:25
  • 5
    $\begingroup$ Why do't you just add the tag to your own (relevant) post and see how it shakes out? $\endgroup$ – JMS Apr 30 '11 at 16:04

I am sympathetic with the idea but believe it could not be used much.

Most people asking questions on this site have statistical concerns. If they want derivations they usually begin (and end) on the math.SE site (where "analytical derivation" would certainly be meta). Moreover, your long list of putative synonyms belies the inherent ambiguity of the concept. "Analytical," "algebraic," "symbolic," "formula," "derivation," and "calculus" mean distinctly different things and have different applications and uses. If somebody really did label a question here with any of those things, I suspect the high-rep community members would quickly migrate it to the math site.

It is especially telling that to date (3000+ questions) nobody has ever attempted to tag their question here with something like "analytical-derivation" or "symbolic", etc. Although on occasion such a tag could usefully discriminate more theoretical questions from practical ones, it tells me your proposal is acting like a solution in search of a problem. Why not wait until people with actual questions create tags like this?

I like your use of the word "evolve": that suggests that a tag should emerge from a few actual successful uses by questioners, become recognized through repeated use as legitimate, acquire a wiki entry, and accumulate a few followers or appear in searches. At that point it would make sense to retroactively apply the evolved, mature tag to older questions.

Note that tags are applied to questions, not replies. It is tempting to retag questions in light of the answers received, but one should be cautious about such retroactive labeling. It is appropriate to do so when the question itself is edited to reflect a new understanding of possible solutions.

  • $\begingroup$ Tags apply to questions and answers together: meta.stackexchange.com/questions/26913 $\endgroup$ – GaBorgulya Apr 29 '11 at 16:32
  • $\begingroup$ @Ga Good point; thanks for the link and the clarification. That makes your proposal more viable. I still think it needs refinement, though. For instance, how are we supposed to recognize answers that benefit from the proposed tag? Would it be all answers that use TeX? What about answers where the math is contained only in code? What about answers that use statistical formulas but don't derive them, or explain the nature of the derivation in English only, or merely link to a formula or derivation (like today's question at stats.stackexchange.com/q/10142/919 )? $\endgroup$ – whuber Apr 29 '11 at 16:35
  • $\begingroup$ The tag numerics has been used 4 times although it could have been used many more times. Once agreed on it I will immediately put the new tag on one of my questions (stats.stackexchange.com/questions/8812/…, you can think about it as already created), so at least one, but in my opinion many more questions need such a tag. $\endgroup$ – GaBorgulya Apr 29 '11 at 16:55
  • 3
    $\begingroup$ If "numerics" refers to numerical algorithms, then the tag clearly does not belong on one of those questions and is useless for two others (which ask only for references). Only the last thread actually discusses a numerical algorithm. This brings us back to the basic questions: what, precisely, would "analytical-derivation" mean and how would one know when it should be applied to a thread? I'm not asking these questions rhetorically: I'm asking them in the spirit of having recognized some merit in your idea and now am wondering how to make it operational. $\endgroup$ – whuber Apr 29 '11 at 17:10
  • 4
    $\begingroup$ One of those was mine. You're right, the numerics tag was useless. IIRC the first tag I tried was numerical-optimization, but it didn't exist so I wound up with numerics and optimization. In retrospect I also don't think the numerics tag meets the standard set in the blog post linked above. I've removed it from that post. $\endgroup$ – JMS Apr 30 '11 at 0:08

Although it is OK to use a relevant tag to couple question and answer it is not obvious that this is a expected behaviour of the site. The tags are shown only under question, so the natural implication is that they apply only to the question, not to answers.

Although the idea is nice, searching only for answers which are analytical for example might be useful, it requires changes to SE engine.

Coupled with the very good points raised by @whuber, I think this tag does not merit the major intervention needed to justify its creation. Starting to use it on question by question basis seems entirely appropriate. I think any new tag must first of all be used only for new questions and then if it shows its usefulness it can be applied for previous questions.

  • $\begingroup$ +1 These sound like good principles for originating any tag, not just the one in question. $\endgroup$ – whuber May 3 '11 at 14:01

Ok, so now we clearly see that you really want to supplement the tag system with a bunch of meta-tags describing what kind of answer does the OP want.

This is:

  • Against official SE policy clearly formulated on blog. (I should stop now)
  • Redundant. It's like ordering books by cover colour -- it may look organised, but has no real value.
  • Complex. This is a whole, non-obvious system that must by manually managed. No new user would get an idea to follow it; also it has no homologue on other SEs, so it would be confusing for SE-aware new users too.
  • Revolutionary. Implementing it would require to manually (!) retag all the old questions, leading to a total chaos.
  • Short-circuiting. It gives user an option to tag the question with meta tag only, so effectively giving no topic tag. This leads either to the destruction of real tag system or necessity to make manual tagging.
  • It would make a properly used [numerics] a meta-tag.
  • 1
    $\begingroup$ I give you 72 hours to calm down, read my proposal thoroughly and rewrite/revoke your post. Please be objective, specific and constructive. Many thanks. $\endgroup$ – GaBorgulya Apr 29 '11 at 20:46
  • 5
    $\begingroup$ @GaBorgulya No personal attack, please. @mbq's points reflect his opinion and echo parts of SE policy, and they are not directed against you specifically. If you don't like them, or feel they are counter-productive wrt. question at hand, simply downvote them. It is the way to go on meta questions that aim to bring potentially conflicting responses. $\endgroup$ – chl Apr 29 '11 at 21:16
  • $\begingroup$ @chl mbq deleted my tagging work, gave only late explanation, labelled it "damage" to "minimise", expressed that s/he thought that organising tags was "just janitorial" work; here s/he states s/he "clearly sees" that I deliberately want to violate SE policy, compares my initiative to colour book covers, believes it is "leading to…total chaos", even ruining an other tag. S/he talks as "we". I FIND THESE UNACCEPTABLE FROM A MODERATOR. Decisions will be made: either (a) s/he revokes/rewites the post (b) his/her moderator status will be revoked or (c) I leave stats.SE. $\endgroup$ – GaBorgulya Apr 29 '11 at 22:00
  • 4
    $\begingroup$ @Ga An ultimatum like this is counterproductive. We will be glad for you to stay and sorry to see you go, but the choice is yours. $\endgroup$ – whuber Apr 30 '11 at 1:44
  • $\begingroup$ @whuber It's outrageous that my constructive intentions are believed to be deliberate destruction. All you originally should have had to do is to say: "Gabor, your new tag is not OK, pls stop, read these guidelines and discuss". What mbq does is not logical and not benevolent. What I had originally expected as moderator behaviour I found written here, search for "what to expect": blog.stackoverflow.com/2009/05/a-theory-of-moderation. And I add that a moderator should assume good intent, otherwise he may easily get into a position where he himself is destructive. $\endgroup$ – GaBorgulya Apr 30 '11 at 10:18
  • 5
    $\begingroup$ @GaBorgulya Well, IMO I'm very constructive; I just stated clearly that your idea is wrong, so you could waste no more time and energy on it. As for b, you may write to the SE team using contact us link in footer, however the logs will show them that all your accusations are false. As for c, I believe there is no point in that -- I really appreciate your efforts, and that's why I want to direct them on a right path. $\endgroup$ – user88 Apr 30 '11 at 10:29
  • 4
    $\begingroup$ @Ga It seems you are reacting to the perceived tone of @mbq's messages rather than their content. We have to make allowance for the different styles and idioms of English used around the world. What to you seem like unfortunate or denigrating choices of words, like "damage," or hyperbolic idioms like "total chaos," are not likely to be meant in such ways. The last sentence by mbq above ("appreciate" and "right path") is consistent with the attitude and intentions expressed by mbq elsewhere. Why not take that as a point of departure for further discussion? $\endgroup$ – whuber Apr 30 '11 at 14:13
  • 4
    $\begingroup$ @Ga Respectfully, you're not doing your argument any favors by taking every criticism as a personal attack. Even if you disagree with what @mbq did I think its been made plain that it was done with the interest of the site and the community at heart. If you're really interested in helping the community and not just making a new tag so you can snag a badge by retagging old posts then why quit the site? That doesn't help anyone. If you truly find the action of the mods so terrible then might I suggest d) hang around until the next moderator election. $\endgroup$ – JMS Apr 30 '11 at 16:03
  • $\begingroup$ @JMS @whuber mbq stated s/he "clearly see" that I "really want…a bunch of meta-tags". In other words s/he is sure that my intention is to violate SE policy – despite I stated the opposite ("This is not intended to be used as a meta tag"). That also explained why s/he so quickly deleted my work and was so late to give any feedback despite SE moderating guidelines. What s/he "clearly see[s]" is an assumption, which is wrong and malicious. No moderator should be thinking like that. But s/he still has time to rethink. $\endgroup$ – GaBorgulya Apr 30 '11 at 19:14
  • $\begingroup$ @GaBorgulya Whacking someone with hammer is equally harmful regardless of how the whacker stated his intentions ;-) $\endgroup$ – user88 Apr 30 '11 at 19:36
  • $\begingroup$ I can't recognise the signs of that you "appreciate [my] efforts". I also can't imagine how the idea of "whacking someone with hammer" came to your mind and how that can be funny. No aspect of my stats.SE activity is comparable to that. I never thought about that and few things would be more antithetical than that to my physician's oath. $\endgroup$ – GaBorgulya Apr 30 '11 at 21:11
  • 2
    $\begingroup$ This discussion seems to be diverging (exponentially fast) from the question topic. There is already another thread here addressing the merits of the recent extensive tagging and subsequent rollback. $\endgroup$ – cardinal Apr 30 '11 at 23:04
  • $\begingroup$ @cardinal I started this new thread to separate the objective question of the optimal tag definition from the previous debate on the rollback. From those previously involved Chl welcomed the clean and new thread, whuber answered with constructive criticism. Mbq, however, answered ("Ok, so now we clearly see that you really want…") with an assumption about my motivation, assuming it to be malicious, and later reconfirmed it. S/he is also committed to "minimise" the "damage" I make. Given that s/he is a moderator with superuser deleting power there is no sense to create until this is settled. $\endgroup$ – GaBorgulya May 1 '11 at 0:09
  • 4
    $\begingroup$ link The rest of us will continue to enjoy asking and answering... $\endgroup$ – JMS May 1 '11 at 15:02
  • 1
    $\begingroup$ @GaBorgulya, mbq is not a native English speaker. That is why his language might seem crude and not respective. $\endgroup$ – mpiktas May 3 '11 at 7:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .