tl;dr: in the tooltip for upvoting/downvoting of questions is the semicolon(;) considered an and/or (meaning do we upvote if it is clear and useful but no research, or upvote for research but not clear or useful)...or is it "upvote because the question has research effort, and therefore is clear and useful"...or other?

Looking at a few of the questions today (not to pick on anyone, just as examples):

This question: Managing Windows 8 Start Screen Tiles (at the time of this post) has 4 upvotes and yet per Jeff Atwood here: In what cases am I meant to downvote a question? the tooltip should be the basis of voting up/down. In this case the question clearly has no research effort but has been upvoted as if it did.

A more recent meta question regarding voting on questions is here: What are the reasons a question can/should get an upvote or downvote on Serverfault? where Iain reiterates Jeff's point on why/when to up/downvote a question.

MDMarra also points out in that discussion:

...we fully expect you to have put some effort into solving your problem before you come here. We also expect you to have done some research before you come here....

So, not to pick on Mark, but his question recently: Is there reserved OID space for internal enterprise CAs? also shows a lack of research. Now, I know he wouldn't post a question without researching it, and I know his background and expertise, and the high rep count always lends to favorable treatment, but should the question be upvoted based on the tooltip? I like the question and upvoted it myself, so I'm just as guilty as the next guy (unless it is perfectly acceptable to upvote based on "usefulness to the community" and disregard research included).

The same could be said for this HIGHLY upvoted question: What are the advantages of tape drives?

In contrast this question appears to have research: What is /etc/apache2/sites-available used for and is it necessary? as does this one: Hyper-V File Server Clustering - at my wit’s end and both had multiple downvotes.

It would appear that while sometimes we vote on questions based on research, a lot of times we vote based on whether we like the merit/premise of the question regardless of the research done. So, is that perfectly valid?

  • I'm with @ward on this - the tooltps are as much advice as has been handed out. As with all voting - vote as you see fit, there is no formula and you absolutely will not get me to change the way that I vote as I'm comfortable with it. Given the general paucity of voting (@ward excepted) I can't see that it really matters.
    – user9517
    Commented Nov 9, 2013 at 13:35
  • 1
    Personally, I would have major trouble determining the amount of research effort put into a question if it is not within my area of expertise. It is way easier to just determine if it is "useful and clear" - I suppose folks just stick with that.
    – the-wabbit
    Commented Nov 11, 2013 at 11:50
  • This site should do what Reddit has done with delayed up/down vote results to avoid the group-think thundering herd. Commented May 21, 2014 at 2:49

5 Answers 5


Man, no wonder you guys don't vote enough, you're seriously overthinking things...

  1. vote up questions you think are good
  2. vote down questions you don't think are good

The tooltips are just that: tips. They're not definitive, legalistic definitions of when you must up or downvote, they're just short summaries of the most likely reasons you'd consider a question to be good or bad.

Looking at the other answers, my thought process for voting up is similar to voretaq7's: if it's clear what they're asking and I think it's useful to myself or another professional, that's good enough.

For downvotes:

  • When reviewing, I routinely click on the questions I think are bad and downvote them (if they're not already) so that if they do get closed the auto-delete process will nuke them. But since unanswered closed questions only need -1 for the auto-delete to get them, there's no need to vote them into oblivion.

  • Redundancy between votes and closing is a feature: you want to close the really bad questions, but in the meantime you want them appropriately downvoted to show that they're bad.

  • I just meant to quantify exactly what I do when I vote. I don't actually go through a process. However, for the questions I don't vote on, it's because they are incredibly mediocre. They're not bad enough to downvote, but I don't really like them either. Commented Nov 9, 2013 at 19:39
  • While I agree with your answer Ward, the problem becomes when an OP asks "Why the downvote?" To respond with "I didn't think it was good" ends up being unhelpful and means that voting in general is based on whims/feelings and not on black/white reasoning. It's not a science obviously, and you explain in detail below #2 above, but most will read just #1/#2 in your answer and go "yep...I vote however I feel like voting"...
    – TheCleaner
    Commented May 21, 2014 at 14:02

I tend to view the criteria as somewhat more complex and subjective than that summary.

If a question seems intelligent or interesting, has a purpose, and is on-topic, I will tend to upvote it.

A question is intelligent where the answer isn't something that would be trivially obvious from a google search or a glance at the software's documentation, and where the right answer would explain a concept or technique (as opposed to an item of work).

Questions which I'd also like to know the answer to will also get upvotes from me. These questions are interesting.

I upvote these questions because they provide an opportunity to teach someone about this field, and they will profit professionally from having read and understood the answer. If I don't know the answer, I would also be improved professionally by reading and understanding it. The key thing here is that a good question is a teachable moment.

Such questions as these tend to provide useful resources for later, and are things I would be proud to answer usefully and have people read my answer in the future. They improve the quality of the community and the resource.

If a question is thoughtless, pointless, trivial, impossibly unclear, off-topic, or really a request for someone to do the asker's job, I will tend to downvote it.

A question is trivial where the answer is a non-conceptual item that could easily be found using public resources. So, I don't mind questions about basic networking concepts, or the name of some obscure GPO, but I take umbrage at those asking things such as how to demote a domain controller, or configure a VLAN on a specific switch.

A question is pointless if the thing being asked is something insane that should never be done, and consequently the answer (if it explains how) isn't going to be of use to much anyone. These questions are typified by "How can I beat this horrible kludge into working?".

A question is thoughtless if the asker is mindlessly doing inadvisable things and is not open to alternatives that follow BCP. This is the "My boss told me to bathe all our servers in water; how?" type of question, or the "We will be using windows for workgroups until 2038; how do I secure it?" type of question.

Really unclear questions, where I genuinely can't tell what is being asked, I will downvote. I'll also downvote questions which are unclear by reason that they are composed of a wall of text or excessive amounts of code blocks, where the asker links a tutorial and asks us to fish through and find why it didn't work for them, or questions which require me to make a lot of assumptions and are asked as though I was already familiar with the asker's environment.

As well, if the asker shows a lack of professionalism or common sense, or exudes a sense of entitlement, it is right that their question should accrete downvotes.

Finally, if the question raises inane constraints that prevent a correct solution from being implemented, I'll downvote it. This is the "My budget for the new wireless infrastructure is $0.50" and the "I admin this system but not that one, so I need to fix my problem without touching that one, and I refuse to talk to those other people".

In general, these questions are things that I wouldn't be proud to write an answer to and have people read it later. They are questions that don't really contribute to the community, and in some cases even detract from it. They are also questions where, even if the answer would ever teach anyone anything, the asker probably wouldn't learn it anyway.

  • So to sum up. You see the tooltip as allowing upvoting if it is clear or useful, not necessarily based on the research effort? I don't disagree with your answer, that's how I tend to vote as well. My real question was are we voting correctly?. So maybe it should be put into consideration to change the upvote tooltip to something like "This question is useful and on-topic" and leave the downvote tooltip as is. Since it would be unlikely that someone would upvote a well researched crap/offtopic question.
    – TheCleaner
    Commented Nov 8, 2013 at 20:44
  • 3
    I view the research effort as part of "useful". It's certainly a necessary condition in most cases. Commented Nov 8, 2013 at 20:46
  • 1
    @TheCleaner I think the tooltip is network-wide. I vaguely recall some talk about changing it a year or so ago (not sure if it was here or on mSO) that went down in flames for some reason but if you can come up with language that sucks less throw it in a separate Meta and I'll make a nuisance of myself in the Teacher's Lounge to see if it's changeable :-)
    – voretaq7
    Commented Nov 8, 2013 at 20:46

TL;DR - Yes, we are, but those tooltips also SUCK.

Upvotes are pretty easy - The tooltip of suggested upvote criteria is
This question shows research effort; it is useful and clear

To me the important part of this tooltip is it is useful and clear, which is what I base my voting on -- "Is it conceivable that another professional would have a similar question, and this this question and its answers would help that person?"

If I can answer yes to that the question is probably worthy of an upvote (or at least not worthy of a downvote).

I think the Apache sites-available question is certainly useful to other admins - especially Non-Debian folks (though I do think the research there was lacking - Google the title and you get an answer - See below).

Downvotes are harder - The tooltip of suggested downvote criteria is
This question does not show any research effort; it is unclear or not useful
This seems like crap guidance to me.

Certainly a question that could be answered by literally pasting the title into Google and clicking the first link should be downvoted and the asker instructed to, well, paste their title into Google and click the first link.
I'm OK with downvoting someone who is THIS lazy but the question should also be closed.
It's crap, and it will attract more crap.

Questions that are "unclear" should really not be downvoted - they should be closed.
We have a close reason especially for these ("Unclear what you're asking"), and more than one downvote on these questions is superfluous. One is enough to ensure the question reaper will do its job.

Questions that are "not useful"? Well EVERY question is useful to the person asking it. If we apply my "useful to another professional" criteria above this is good guidance, but it's important to remember there are varying levels of experience across the broad swath of "professionals".

We should not be voting based on whether we like the premise of the question.
That said, we do.

The Hyper-V clustering question is getting smacked because of the XY problem: The person asking the question has latched on to a solution and is abjectly refusing to listen to the Right Solution (If anyone is now saying, “wait, what about a SAN or a NAS for the file servers?”, well too bad.)

I'm not sure how I feel about this as a moderator (Mod Hat On it's really NOT what downvotes are for Mod Hat Off), but as a professional if someone is going to take up our time and expect us to give them free advice they should state their problem and listen to our solutions without imposing insane restrictions.

  • The tooltips are fine, they are clear and concise. As with any voting people should vote as they see fit within the bounds of the system. That very few people participate beyond getting their (mostly crappy) question answered means that the voting is left to a relatively small number of people so growing that number is the only solution.
    – user9517
    Commented Nov 12, 2013 at 7:41

As one of the people that downvoted the Hyper-V question, I echo voretaq7. He had asked an earlier question and was told he was approaching the issue with a non-ideal solution and was given various other suggestions on how to better proceed. Rather than taking the advice, he basically posted the followup indicating that he had not taken any of the advice and forged ahead with a bad design and got stuck and asked for help again.

The honest part of me said I couldn't close it as a duplicate because it technically is related but not a duplicate. Had this been his first question on the issue, I would not have downvoted it (as evidence: I didn't down vote his first question) but worked to show up the other alternatives. But his past questions and lack of following the sound advice given led to my downvoting his followup question.


Unlike stack overflow my experience on serverfault seems more like driving thru a slum neighborhood. Your question is likely to be voted down with no reason attached. So post at your own risk and hope you don't suffer from a drive by shoot down. Not much thought being put into that at all. Worrying about the tooltips is like trying to paint a burning house, what's the point?

  • 3
    I went to downvote your post, but unfortunately, it was a drive by vote and I caught @TheCleaner in the cross fire :(
    – Dan
    Commented Nov 22, 2013 at 16:40
  • 6
    You could always try asking better questions.
    – MDMarra
    Commented Nov 22, 2013 at 16:42
  • 1
    Point proved. ServerFault, the place to go when you absolutely don't need any help.
    – boatcoder
    Commented Nov 22, 2013 at 17:23
  • 1
    Mark, I think the issue you are running into here is that SF operates quite differently from SO. While SO has a more "come one come all with your programming questions, regardless of your experience", SF is dedicated to professional sysadmins. As such, it's expected you not only follow the site rules but are yourself a pro sysadmin. That's not why your other question got downvoted though (at least I don't think it is, I wasn't who did it). Most likely it was due lack of research. Simply throwing out a "I don't have time to figure it out...here's the error, what's wrong?" doesn't fly here.
    – TheCleaner
    Commented Nov 22, 2013 at 18:05
  • 3
    And the downvotes here on Meta are different...here in Meta it means "I disagree with your post".
    – TheCleaner
    Commented Nov 22, 2013 at 18:06
  • 4
    I downvoted your most recent question because quite frankly it showed no research and was incredibly sucky. I had to read your code block to discover hos the output had been generated. I used google and was educated second link appears to solve the problem.
    – user9517
    Commented Nov 22, 2013 at 18:30
  • Mark - in the vast majority of cases, drive-by-downvotes are left without comment because they conform to the downvote guidelines laid out in the down-arrow tooltip.
    – EEAA
    Commented Nov 22, 2013 at 18:43
  • 6
    I'd like to also point out that there's nothing wrong with "drive by" downvoting. It's been discussed time and time again on meta.stackoverflow.com among other places. If Stack Exchange didn't want people downvoting without leaving comments, they'd make leaving a comment a requirement to downvote.
    – MDMarra
    Commented Nov 22, 2013 at 19:24

You must log in to answer this question.

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