Q. Should the FAQ advise members to use code delimiters for output?

[Bruce Weaver]

Section 12.3 of the FAQ explains how to use code delimiters when posting to the forum. Currently, it says nothing about enclosing output in code delimiters. I think it should explicitly advise members to enclose output in code delimiters. YMMV, but I certainly prefer this...

. clear

. sysuse auto
(1978 automobile data)

. regress price weight

      Source |       SS           df       MS      Number of obs   =        74
-------------+----------------------------------   F(1, 72)        =     29.42
       Model |   184233937         1   184233937   Prob > F        =    0.0000
    Residual |   450831459        72  6261548.04   R-squared       =    0.2901
-------------+----------------------------------   Adj R-squared   =    0.2802
       Total |   635065396        73  8699525.97   Root MSE        =    2502.3

------------------------------------------------------------------------------
       price | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
      weight |   2.044063   .3768341     5.42   0.000     1.292857    2.795268
       _cons |  -6.707353    1174.43    -0.01   0.995     -2347.89    2334.475
------------------------------------------------------------------------------

...over this:

. clear

. sysuse auto
(1978 automobile data)

. regress price weight

Source | SS df MS Number of obs = 74
-------------+---------------------------------- F(1, 72) = 29.42
Model | 184233937 1 184233937 Prob > F = 0.0000
Residual | 450831459 72 6261548.04 R-squared = 0.2901
-------------+---------------------------------- Adj R-squared = 0.2802
Total | 635065396 73 8699525.97 Root MSE = 2502.3

------------------------------------------------------------------------------
price | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
weight | 2.044063 .3768341 5.42 0.000 1.292857 2.795268
_cons | -6.707353 1174.43 -0.01 0.995 -2347.89 2334.475
------------------------------------------------------------------------------

.

[Nick Cox]

Sounds a good idea to me. If there's dissent, please signal one way or another. Thanks for the suggestion.

[Richard Williams]

Some people are nicer than me, but I often don't even try to answer a question when code delimiters have not been used. I agree the FAQ should be explicit about this if it isn't already.

[Jared Greathouse]

Yeah I'd support this. I already don't answer many questions that don't use them, so I think being more explicit about this is fine.

I've said before that people (new users who make an account) should arguably be forced to undergo such a process before they can ask anything. In other words, nobody should have the ability to say "I've never used dataex" or "I don't know how to put the code in delimiters".

Of course, this is much more work on the backend, so I don't expect it to be implemented, but it would at least do some kind of quality control in terms of the way the questions are asked.

[Mike Lacy]

Describing them as "[plain] text delimiters," rather than "code delimiters," might be worth considering.

[Carlo Lazzaro]

I do agree with Bruce's proposal.

[Girish Venkataraman]

A sticky link to a Chuck Huber style 2-minute video tutorial on 'Best practices with posting' would not be a bad idea. The # icon is just not intuitive (even though I had worked with Stata for over 2 years before joining the the forum). I am at fault for not reading the rather lengthy FAQ section myself when I registered and probably did so for the first time only when I did not get a response to one of my posts.

[Nick Cox]

First up, just a reminder that Statalist is run by StataCorp. It didn't start that way. Back in 1994, and for 20 years, Statalist was an email-based server run out of the Harvard School of Public Health and maintained largely through the goodwill and hard work of a few people there. It's remarkable that was sustainable for so long.

It's diverting that whenever one reads a post and thinks "That's just too messy or confused or lazy to allow or even deserve a reply" and so moves on, no one knows but you, while if you ask people in public to follow the FAQ Advice -- please show your code formatted readably, give a data example, explain the exact error -- some fraction of the world thinks that you are being hostile, rude or obnoxious.

I've heard of forums in which nothing is posted publicly without getting a moderator's approval first, but that's not going to happen here.

Jared Greathouse wants people to be obliged, even forced, to take a test before they can use Statalist. That's not going to happen either.

Many people manage to read the FAQ Advice before posting and follow its suggestions, so it can't be that hard.

[Jared Greathouse]

I've heard of forums in which nothing is posted publicly without getting a moderator's approval first, but that's not going to happen here.

Yeah to me this is kinda nutty, to be very polite about it😂 without even getting into the other problems with this, my first question would be "What kinds of people are attracted to these jobs or these positions? What incentive structure would this create?"

Statalist isn't a paid service (and if it becomes one, I'll be the very first in line), and neither are many other forums (I think??), so the fact that some people become moderators on these forums just strikes me as odd, especially given the volume is traffic they tend to get. I think the worst case outcomes is getting posts like this, which were presumably authored by full grown adults, professional people who ostensibly do good work in their day jobs. It just seems like a MASSIVE opportunity to abuse a very small position of power, and I'm glad that Statalist (very very generally, I can think of like two exceptions) does not have a group of moderators that are super active.

Nick Cox once described Statalist to me to be like a bar where people in the community may come and speak freely, paraphrasing slightly. It's precisely that commitment to making people feel welcome (while also having reasonable standards for queries, like at the top of the thread), that I've used Stata for say, 7 years now, and proudly.

[Nick Cox]

I guess I started a digression by mentioning practices elsewhere (I call it context too), but Jared Greathouse pushed it further. His questions and comments about Stack Overflow (SO) can in part be answered factually (and in part just raise differences between people and their different choices). I am moderately active on the Stata tag of Stack Overflow and also on its statistical sister site Cross Validated. I am not a moderator at either place. Both sites are part of a wider network called Stack Exchange (SE).

The link in #9 to a post on Stack Overflow Meta is puzzling here. The discussion there was about editing (not at all about moderation). To explain if you want an explanation (see also https://stackoverflow.com/help/editing): Anyone on Stack Overflow can propose an edit to an existing post; if a user has more than 2000 reputation their edits are automatically accepted (although in principle and occasionally in practice they can be reversed or modified by anyone else with such a reputation). Thousands of people on SO have that level of reputation. Why would anyone want to edit someone else's posts? The over-arching ideal for a site like Stack Overflow is its goal of an archive of answered questions in good condition following a general house style: this is like, yet also unlike, Statalist, and in many ways Stack Overflow is closer in spirit to say Wikipedia than we are here.

The debate Jared links to is one about editing and I would summarize it as being about whether certain kinds of minor edits are just too pedantic or pointless to be worth it. or on the contrary such edits are worthwhile if minor contributions. I think Jared is implying that such editing, or arguing about such editing, is pretty silly, and if I am reading that correctly, that's his view, and I don't want to argue with it. It's easy enough to discover that I have edited some posts in the places I've mentioned, so that is enough of a signal on that.

Disagreements on SE about what kinds of editing are worthwhile are really minor and far from the worst things that ever happen there, which are mostly people getting very annoyed at their question being closed or deleted.

Now moderation on Stack Overflow (generally Stack Exchange) is a different deal. Moderators are elected by their communities; there are minimum qualifications before anyone can stand for election, or can vote in an election, but moderation is not a job and is not paid. Moderators have more privileges and indeed more responsibilities than other users. Why anyone would want to be a moderator is not too difficult to think about, and it just boils down to whether you think it would be. net, rewarding public service. I can't think that anyone interested in power should want to be a moderator anywhere on Stack Exchange: the most likely parts of the experience are tedium in dealing with multiple substandard posts and flak from dealing with irate members, either on SE or off it. That is, a thread can be closed (and ultimately deleted) if not a clear question, or being a duplicate question and not everyone appreciates that.

As a matter of history, Marcello Pagano is called the Moderator here, because he ran Statalist before 2014, but he would be the first, I imagine, to underline that it's a honorific title and that he has and wants no power in the present set-up. FWIW, I am the FAQ maintainer but I can't make edits unilaterally even to the FAQ, and that's as it should be. .

Statalist is quite different in respect of editing and moderation from SE. You can edit your own posts within a hour of first posting; otherwise only the administrators, certain employees of StataCorp, have any power to edit posts in detail or to delete them. Either action is pretty rare: as I understand it the leading reasons for StataCorp intervention are spamming and belated realisations that somebody has posted confidential material, meaning usually data examples that should not be public.

I don't recall comparing Statalist to a bar. It's a complete failure on that in some respects. I need to get off it to get any kind of drink or snack. But I have made some good friends here, I think.

[Nick Cox]

Back to #1: thanks to our never-slacking sladmin the FAQ has now been updated at 12.3 within https://www.statalist.org/forums/help#stata

Thanks to Bruce Weaver and all who contributed to discussion.