@darijgrinberg If you do not find link-only answers to be all that bad, please post a meta question suggesting that the policy of deleting link-only answers be revisited. Frankly, I suspect that consensus will be against you, but if it turns out that I am wrong, I won't bother flagging such answers in the future.

Best solution for link-only answers is to convert them to comments, IMO

@AlexanderGruber Sounds good to me. Should I flag such answers for moderator intervention, then?

@XanderHenderson I would say yes, when the mood strikes you. I usually convert to comment when I see them anyway.

Groovy.

The exception would be when it's a reference request or something like that, where the link is the answer. What I would do then is just edit to make sure the source is properly referenced via text in the post in case the link breaks.

@darijgrinberg that's are very odd comment. First, it does not only makes sense for that. It also makes sense to guarantee that there actually is some readable content on the site. (For one thing, on some devices pdfs are a pain.) Second, why is it hardly appropriate to discourage self-PR?

because sometimes it isn't as much self-PR as "it makes more sense to collect a bunch of things in a PDF than to spread them across 5 different posts on m.se"

i wouldn't claim PDF is a pain, at least not on tablets

i have much worse experience with m.se on tablets

mupdf works way better than i'm used to see software working :D

@darijgrinberg MSE is meant to be self-contained. If the content doesn't work in the format of a self-contained question/answer pair, then that content is not welcome on MSE. That doesn't mean that it isn't good content---it may very well be excellent content. It just is appropriate for the format.

@darijgrinberg if somebody wants to post here, they should post here. If not then not.

The site is not a link farm.

And not to single out Pete specifically, but personally I do see problems with self PR.

It's less of an issue on this site by now though.

re the pdf remark, well and I don't read pdfs on my phone.

@darijgrinberg an often used image that explains what the NAA flag is for explicitly includes link only answers. It is a rule. (Of course there can be rare exceptions as already mentioned.)

That said, I agree that an inconvenience of the NAA is that it's "all or nothing" and the options of convert to comment might sometimes fall through the cracks. Thus, I agree with @Alex that flagging for mods can be better.

@darijgrinberg that's a bit funny since those that mostly argue against deletions often stress helping the poor (new) users.

There is much less useful content creation than there could be, because 'everybody' is busy writing the same thing yet again and dealing with all that stuff in one way or another.

@darijgrinberg maybe so, but it should still be reasonably scoped and clear what the intent is. Plus there is zero reason not to put a moderate amount of effort into fixing the writing prompt. It also improves the general appearance.

@darijgrinberg it might make it much harder to find though, with poor title etc.

Yeah, I was explicitly talking about cases where the question has been made readable and been given some context (not necessarily the one in which the OP encountered it).

While browsing those many link only answers I noticed another issue they pose, e.g. here one user comments that the linked answer is identical to an earlier posted answer (by me). Of course it is not so easy to notice things like this when the link is offsite. And you can't change your vote if the link target changes (e.g. to something incorrect), as you could with an edited answer. Just a couple of many serious issues posed by link-only answers.

@darijgrinberg the question that started the discussion include you given context in what I find a very odd form.

It's fundamentally unclear to me why you did what you did.

It's just seems weird, like a willful provocation.

And, of course, there is no onsite bump if the offsite ink target changes.

As for ethics, I don't see any problem if you post answers here which are supplemented by such links, e.g. "see Theorem xyz in my notes for generalizations".

@darijgrinberg well, you didn't. It's not clear if you actually think so. You should maybe try not to be dismissive of those standards and still somehow make it by them but actually try to appreciate what the intent is.

del-pls polynomials Missing context or other details (No Roomba: accepted answer) $p$ and $q$ be the roots of the polynomial $mx^2 + x(2-m) + 3$.‭ - starunique2016‭ 2016-05-29 15:35:54Z

@darijgrinberg really what it comes down to is that pdfs are frequently moved or deleted, and even if they are archived, many users may not know about that site or think to look there. In either case, it's certainly a suboptimal solution compared to having (at least an outline of) the answer written here. Adding the link is a good supplement, just not sufficient on its own.

@user21820 C2 needs some attention. (since it's a previous HNQ).

[ SmokeDetector | MS ] Bad keyword in answer (86): If $f(x) = x^3-3x+1.$, then no. of distinct real roots of $f(f(x)) = 0$ by Balakrishna Murthy on math.SE

[ SmokeDetector | MS ] Repeating characters in body, repeating characters in title (135): y = mx+b??????????????? ✏️ by blakekeke on math.SE

titles I have moved \DeclareMathOperator from the title into the body in this post - based on the advice on meta and some recent discussions: $\gcd(ord(a),ord(b)) = 1$ then $ord(ab)=ord(a)\cdot ord(b)$ where $a,b$ do not commute. The problem is with the length of the new title if I want to use \operatorname{ord}.

"$\gcd(\operatorname{ord}(a),\operatorname{ord}(b))=1$ then $\operatorname{ord}(ab)=\operatorname{ord}(a)\cdot\operatorname{ord}(b)$ where $a,b$ do not commute" has 157 characters.

Ok, so maybe this: "$\gcd(\operatorname{ord}(a),\operatorname{ord}(b))=1$ then $\operatorname{ord}(ab)=\operatorname{ord}(a)\operatorname{ord}(b)$ if $a,b$ do not commute"

@MartinSleziak \rm works as well

Doesn't it have different spacing?

I mean, one of the two commands just changes the font, the other makes it into an operator.

At least, I vaguely recall that this is how it works in LaTeX. (Of course, I might be wrong.)

$\gcd({\rm{ord}}(a),{\rm{ord}}(b))=1$ then ${\rm{ord}}(ab)={\rm{ord}}(a)\cdot{\rm{ord}}(b)$ where $a,b$ do not commute

@MartinSleziak yes

In any case, that's a good idea for shortening. (Although some LaTeX purists might object.) Thanks for that!

No worries, those objectors might have a hard time trying to fit \operatorname in though :)

Well, at least in the current situation, it was fixed in some way.

Of course, there are also other posts with definition of a macro in the title.

HNQ missing context

@YuiToCheng And it got 6 upvotes. :-(

delete Not an answer to the question being asked.

HNQ duplicate

Nonsense question

del-pls 20k+ geometry this doesn't answer the question Answer to: Confused with equation of plane‭ - Broly‭ 2019-04-06 21:49:42Z

[ SmokeDetector | MS ] Bad keyword in link text in body (67): You are flipping n fair coins, putting aside those that come up heads after each flip. What's the probability that you'll end up with one coin? by johndoe8971 on math.SE

D1, D2, D3.

D4, D5, D6.

D7, D8, D9.

D10, D11, D12.

D13, D14, D15.

C1, C2, C3.

C4, C5, C6.

C7, C8, C9.

[ SmokeDetector | MS ] Mostly dots in answer (36): Probability that 2 Boys or 2 Girls are selected by user656977 on math.SE