What is the optimal length & level of detail of tag wiki excerpts?

Question

I find tag wiki excerpts extremely useful, much more so than full wikis (I doubt that many people ever come across full wikis because SE design buries them quite deep).

When I write wiki excerpts I have been trying to put as much information in it as possible given the strict limit of 460 characters. If a wiki excerpt is much shorter than that, I usually try to expand it until it reaches the maximal capacity, because that's how I believe it will be most useful. When I mouse-over a tag I prefer to get more information and not less.

For example, I have very recently written tag wiki excerpts for elastic-net and total-least-squares; both approach this hard length limit and both were approved by the community.

Today I rewrote tag wiki excerpt for distance-covariance, but my edit was edited and strongly shortened by @whuber with the comment

Excerpt was too long: moved excess material to the wiki itself.

I am unhappy about it because the excess material moved to the main wiki will be seen by almost nobody, whereas in the excerpt I think it could be useful. I thought that there can't be such thing as "too long" an excerpt: the upper limit is given by the 460 allowed characters, and any length below that is not too long.

What is the optimal length of tag wiki excerpts? What is the appropriate level of detail? The more detailed the better (as I thought)? Or the more concise the better (as per @whuber's comment)?

Answer 1

Like @Scortchi and @whuber, I prefer shorter excerpts. (Besides aesthetic preference, I suspect they are more likely to be read.) I think excerpts should mostly proffer tag usage guidance, and that more substantive and detailed information is better placed in the full wiki.

That said, what is "shorter" exactly? To some degree "shortness" is in the eye of the beholder. Moreover, it may not be fully possible for a reader to infer how a tag should be used without a brief note clarifying the nature of the thing in question. As a result, I don't object to including some of that when: (a) appropriate, (b) made as briefly as possible, and (c) placed after the definition / basic usage. That is, it would be somewhat less prominent, in what newspaper reporters used to call the inverted pyramid style.

To put this more concretely, when the question is asked in the abstract, I would have responded as @Scortchi (+1). Indeed, I have edited excerpts to shorten them and move information into the full wiki. On the other hand, in the example given in @amoeba's answer ([elastic-net]), I think the excerpt was perfectly appropriate and not excessively long.

---

Edit:
As I say, I prefer shorter excerpts, but sometimes some brief information about the topic is necessary to help people could understand how the tag should be used. For example, I wouldn't find "a regularization method for regression models" helpful for the elastic net, because that is equally descriptive of the LASSO and ridge regression. Mentioning orthogonal and Demming regression in TLS could help people recognize that it is the tag to use. Likewise, "a measure of statistical dependence..." is too nonspecific for my taste.

My suggested text for the three excerpts in question is below. I do believe that every character you can trim increases the likelihood the excerpt will be read. That said, I have attempted to retain as much of the informational content as possible. I will acknowledge that these may need some work before they are posted; I'm not familiar with distance covariance and only somewhat with total least squares.

• elastic-net

  A regularization method for regression models that shrinks coefficients towards zero. The elastic net combines the penalty terms used for LASSO and ridge regression.

  The numbers of characters are 326 (amoeba) > 165 (gung) > 46 (whuber).

• total-least-squares

  A method to fit a linear model by minimizing the errors in both X & Y (OLS minimizes errors in Y only). Orthogonal and Deming regression are special cases of weighted TLS. It is often used when both Y and X have measurement error.

  The numbers of characters are 425 (amoeba) > 230 (gung) > 113 (whuber).

• distance-covariance

  Distance (aka Brownian) covariance is a measure of statistical dependence between two random vectors of any dimension. DC(X, Y) = 0 iif they are independent. Sample DC is computed via the pairwise distances between all points.

  The numbers of characters are 431 (amoeba) > 226 (gung) > 106 (whuber).

---

Update by @amoeba

Following some additional discussion in the comments in this thread, we converged on the following excerpts:

A regularization method for regression models that combines the penalties of lasso and of ridge regression.

A technique to estimate parameters $\beta$ of the linear model $Y=X\beta$ when both $Y$ and $X$ are subject to measurement error. Includes Orthogonal and Deming regression as special cases.

A measure of dependence between two random variables (or two random vectors of any dimension). Also called Brownian covariance.

109 / 174 / 127 chars. I am accepting this answer as this level of detail roughly seems to be the consensus.

Answer 2

I'd go for the more concise the better, to increase the chance people actually read the tag wiki excerpt before using the tag & perhaps therefore use it correctly.

The usage guidance, or tag wiki excerpt, is a short blurb that describes when and why a tag should be used on this site specifically.

Mousing over a tag & reading the excerpt involves deciphering tiny white text on a grey background while simultaneously maintaining the cursor positioned over a small grey box. This aren't circumstances conducive to learning what distance covariance is; but to quickly checking what the distance-covariance tag is for, & what it's not for: I, for one, am grateful for brevity. When adding a tag to a question, you're presented with a box containing six tag wiki excerpts, which is updated as you type: concise excerpts aid a speedy choice of the correct tag. If 460 characters is a wise maximum bound it doesn't follow that approaching 460 is an optimum.

I don't think the full wiki is buried very deep—it's one click away from the excerpt pop-up—, & it's the right place to read, in comfort, a fuller explanation of what the tag refers to.

Answer 3

The purpose of a Wiki excerpt is not to supply as full and complete a definition as you can. It is to help people apply tags appropriately. To be effective, the excerpt therefore should:

• Be as short as possible, for otherwise it will be ignored.

• Be clear about what the tag means--but only to the point of distinguishing it from other possible tags that could reasonably be applied.

• Distinguish it from related tags.

• Provide guidance concerning how the tag is intended to be used (or not used).

For some of the more effective examples, read the excerpts for the r, self-study, and clustering tags. (To fully appreciate how those excerpts will appear, you will need to find a page on the main site that displays a tag--which can be done by clicking through the links in the preceding sentence--and hovering your mouse over the tag wherever it appears.)

I warmly encourage qualified community members to review tags in subjects you are interested in, to create or modify their excerpts to meet these objectives, and--if you are so inspired--to supply additional information in the tag wiki itself. If you want to write a review of a subject, though, then please consider our blog page: it's a much better alternative.

Answer 4

I will post my own answer so that we could have some sort of voting here. I am obviously for detailed excerpts.

After my post, @whuber edited all excerpts that I mentioned to shorten them drastically. Here are the edits so that people can compare which version they prefer:

• elastic-net: https://stats.stackexchange.com/posts/184175/revisions
• total-least-squares: https://stats.stackexchange.com/posts/184871/revisions
• distance-covariance: https://stats.stackexchange.com/posts/177139/revisions

I find that shortening e.g. the first excerpt from

A regularization method for regression models that penalizes the size of regression coefficients $\beta_i$ and biases them towards zero. Elastic net includes two penalty terms, one proportional to $\sum |\beta_i|$ and another proportional to $\sum \beta_i^2$. When used alone, these penalty terms lead to Lasso regression and ridge regression respectively.

to

A regularization method for regression models.

makes it useless. Lasso, ridge, and elastic net could have exactly the same excerpt then. Honestly, I am baffled.

---

Update (May 19, 2017)

Other answers (+1 to each) convinced me that conciseness is important and made me change my opinion. After all the discussions, we arrived to something like a compromise (but more on the @whuber's side of the spectrum). For [elastic-net], the excerpt reads as follows:

A regularization method for regression models that combines the penalties of lasso and of ridge regression.

and by now I agree that this is better than my version above.