Case study: How can I typeset the date/time at compile time?

A perfectly good answer was left, but without a MWE (minimum working example). It looked easy enough to do, so I figured I'd try to augment the answer with such. It was a bit long for a comment, so I left it as an additional answer (making it clear, I hope, that it was based on the original answer). In working out the MWE, I did add one minor bit of information (a recommendation to use the latest version of one of the packages). As I see it, I had the following options here:

  1. Leave a comment asking the original answerer to add a MWE
  2. Edit that answer including the MWE
  3. Do as I did (leaving a separate answer), but make my answer CW
  4. Do as I did (leaving a separate answer)

Obviously, the best situation is where the original answerer includes a MWE in the first place, so this is in the situation where that hasn't happened. Here's my thoughts on my options:

  1. This puts the onus on the original answerer to make it clearer. I'd rather have the convention that any information is potentially useful and that people should not feel that they have to put in a MWE - I think that making this a de facto requirement will put people off answering. Of course, it's desirable, but we want the "barrier to entry" to be as low as possible. In this case, working out a MWE was fairly simple.

  2. I think that this is actually the ideal situation and is what I would like the convention to be. However, given that people tend to be a bit possessive of their answers, I'm hesitant to promote this as a first line. Perhaps a combination of this and the above: leave a comment saying, "Do you mind if I add a MWE to your answer?". The problem with this (and answer 1) is that it puts a delay in the system.

3 and 4 are pretty similar. The issue here is in deciding when a fresh answer adds sufficient information to warrant it not being CW. Perhaps doing a MWE is enough to warrant the extra credit since it can take a few minutes to iron out the wrinkles, or figure out the dependencies. I'm not sure.

Certainly, I'd like the environment here to be such that we do take a look at people's answers and look for ways to tweak them to be better; but I would like it to be the case that that is interpreted as meaning:

Good answer! Let me tweak it a little to make it a great answer.

Rather than,

Hmm, along the right lines. If you do X, Y, and Z it might just be okay.

Just remembered another example where I wanted to edit a question: passing options to the newenvironment, in that case, I left a comment.

  • OT: I wonder why the answer in you second example got accepted. It shouldn't even compile.
    – Caramdir
    Aug 4, 2010 at 11:36
  • 1
    To your question: It might depend on how acquainted with the SE way the poster is. While somebody new to the site might find it offending if you change their answer, most long-term users will probably just see it as a way to get a better answer (and more rep) for free.
    – Caramdir
    Aug 4, 2010 at 11:40
  • 1
    @Caramdir: it does compile! \begin{hello} ... \end{hello} is a synonym for \hello ... \endhello. I wondered if it would complain about there being no \endhello but it didn't. Aug 4, 2010 at 12:12
  • Thanks for the clarification. LaTeX is more forgiving than I thought (though I whether this is a good thing is another question).
    – Caramdir
    Aug 4, 2010 at 12:30
  • Does anyone know what the SO policy on this is? And MO?
    – Caramdir
    Aug 4, 2010 at 12:31
  • I don't think the policy was ever really spelled out on SO, other than "respect the original author". I added an answer describing how I'd do it on SO though.
    – jalf
    Aug 6, 2010 at 12:23
  • @Andrew Stacey: \begin{foo} ... \end{foo} is essentially \begingroup\foo ... \csname endfoo\endcsname. (There's some sanity checking, but functionally, that's all there is to environments.)
    – TH.
    Aug 26, 2010 at 5:24

4 Answers 4


I've always just followed my gut feeling on this, so I'm not really sure how to spell it out.

On SO at least (and I assume on all SE sites, but I haven't checked), there is a bit of advice in the right sidebar when you edit someone else's question. Among other things, it says to respect the original author.

So I tend to edit when I can (if I have the rep for it), and if the answer is small and noncontroversial and is what the original author wanted to say.

For something like adding a minimal working example, I'd make a separate answer (I don't think I'd CW it either), and for completely pointless typos (teh->the, or other stuff that doesn't really affect the meaning or readability) it just feels too "grammar nazi" to edit it, so I just leave it. I don't want anyone to feel they have the grammar police hot on their heels, eager to call you out the moment you make the slightest misstep.

For minor typos or errors that actually disturb the meaning of the answer, and where I know the author would agree with the correction, I just edit it and leave a comment explaining what I did.

On SO, where code snippets are common, people often make minor errors in their sample code, and so editing it seems the simplest course of action.

On this site, I'd edit something like this:

\end{tabular} // oops, missing *

because it's clear that the author just made a mistake, he intended to write it correctly, but made a typo, which might confuse the reader.

If I'm not 100% sure what they meant to write, I add a comment pointing out the error, suggesting a fix, and asking if that is what they meant.

But I wouldn't edit a MWE into their post, because that's kind of intrusive, putting words in their mouth. I'd make my own answer instead

  • 7
    If I might attempt a summing up: "polish, but no substance". I think I'd go along with that. Aug 6, 2010 at 12:28
  • yeah, but with the extra qualifier "don't polish unless it matters". Like I said, fixing individual typos that don't detract from the meaning just seems overly aggressive.
    – jalf
    Aug 6, 2010 at 17:45
  • I think that SU has a somewhat more interventionist culture, following quack's preferences: questions aren't really edited more often, but questions that don't make clear what the problem is and what information is being sought are at risk of having their innards pulled out, and wrought into a proper introductory paragraph. People don't often do that on SO, though I have done that a few times. Aug 13, 2010 at 9:43
  • 1
    @jalf Actually, I would disagree on that. If you comprehend this network as collective accumulation of knowledge and advice, enhancing will always make something better. Say polishing non-substancial "teh" to "the" improves readability and reduces disambiguation. For that the minimum edit limit of some (10?) characters is there. I would always do so while of course respecting the original author and his ideas/style.
    – Cadoiz
    Aug 19, 2020 at 10:13

Yes, I'm also often left wondering about this. In fact I just edited the accepted answer here, and it wasn't even me the one who came up with the additional information! I stole it from a comment Dima left on the question because I thought it was an important enough clarification to add.

As the site develops we will probably figure out where is the "line" for acceptable edits, but these are my thoughts as of now:

  • If we can get away with a minor edit, or even just a comment, then that should be the preferred solution.

  • If the required edit is slightly mayor, then editing and then adding a comment along the lines of "Hey, good answer! Hope you don't mind I added XYZ to make it clearer/more complete/whatever", so that no feelings are hurt and also we don't have to wait for the changes to happen. Moreover, as far as I understand, the person posting the original answer keeps the rep. points if the answer gets accepted/upvoted. (Which is actually the reason why I was hesitant to accept your answer on the datetime question, he was first and he had a good answer!)

  • Of course, if the answer is completely different and would require a complete rewrite then a new answer is the way to go.

  • Indeed, I was actually surprised to see that you had accepted it! Though now I've added the work-around for TL2009 then perhaps it's more acceptable. Aug 4, 2010 at 12:13
  • On a second thought, I've just accepted Grigory's answer. First we don't rob his points, and then your answer starts "Expanding on Grigory's answer" so I guess this ordering of the answers is more useful for people stumbling on the page. Hope you don't mind and find it fair! Aug 4, 2010 at 12:19
  • I certainly didn't feel robed (but thank you)
    – Grigory M
    Aug 4, 2010 at 12:32
  • 1
    Certainly! I think that this is what ought to happen. Indeed, any post that starts (either explicitly or implicitly) "Expanding on X's answer ..." should be viewed as linked to the parent post and anyone who votes for the child should vote for the parent. Since you only get one accept vote, it ought to go to the parent by default (ie unless the original is so terse as to be unusable). Aug 4, 2010 at 12:38

The crucial convention on SO is, as jalf says, respect the original author, and I think what Andrew thought of & Juan did is a kind of interventionism that is generally avoided on SO. We can consider edits on a scale from conservative: fixing typos, putting in pre and code tags, indenting code examples; to interventionist, where one substantially rewrites part of a question, say, to turn a subjective and argumentative question into a productive question.

While SO's culture avoids interventionist edits,

  1. The harm done by even very interventionist edits need not be so very high: users always have the power of rollback with their own questions and answers. If you are doing something interventionist with a user's content, you can explain why you thought the edit was helpful, and point out the rollback option in a comment.
  2. There are bounds to respect: you don't need to bend over backwards to avoid annoying people, if you think your edit makes the question/answer clearly more readable. From the SO FAQ:

    Other people can edit my stuff?!

    Like Wikipedia, this site is collaboratively edited, and all edits are tracked. If you are not comfortable with the idea of your questions and answers being edited by other trusted users, this may not be the site for you.

  3. We can easily move the boundaries here on tex.sx by making it policy. Adding hyperlinks is normally considered to be conservative editing on SO; we can say that adding examples is conservative editing here.

There was a question on MSO that I found helpful, What to do when people with very high rep are completely changing the meaning of a question?

  • 1
    Can you clarify: in both of the cases I cited, I did not edit the answer that I thought about answering. In one case, I left another answer, in the other case, I left a comment. So I'm unclear about your comment about "overstepping the bounds". Aug 12, 2010 at 19:34
  • @Andrew: I'm sorry, there were so many things wrong with my answer, which I have substantially revised. I did understand that you hadn't made that edit when I read your qn, but somehow when I was typing... "overstepping the bounds" is much too strong, even on SO, and I was not saying what you were considering, and what Juan had done, are that here on tex.sx. I've changed to the language of conservative vs. interventionist editing. Aug 13, 2010 at 9:20

Probably, (2) is the best (generally, I'm not too comfortable with someone editing my answers, but just adding MWE or something is certainly OK). But this is an option only for high-rep users.

And for everyone else I'd suggest (4'): posting a separate answer with MWE if/after OP (or anybody) requests it.

  • Your comment about high-rep is true, but also it's worth thinking about getting conventions in place before loads of people have enough rep to edit answers. I like (4'), but I'd add the extra that it's okay to add a MWE as a separate answer without it being asked for once the initial activity has died down - namely, to leave the Q&A in the best possible state "for posterity". Aug 4, 2010 at 12:16
  • @Andrew guess, I'm just too lazy to do such thing w/o request :)
    – Grigory M
    Aug 4, 2010 at 12:33
  • 3
    That's partly my point: I don't want to expect this behaviour of people, but some (like myself) might feel like going round tidying up a little here and there to make things more usable by future questioners. Having to get permission every time is going to be a bit of a chore. Aug 4, 2010 at 12:39

You must log in to answer this question.

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