A question of mine was closed because it was signalled a "shopping question. Even if it is a "comparison" (or a "which is better") question, it seems to me a good answered one. I asked for "facts, references, or specific expertise", and I asked for "what I need to learn". Indeed it has no answer :D

How can I improve a question like that?

  • See blog.stackoverflow.com/2010/09/good-subjective-bad-subjective That said, looking at the question I don't see how you can turn that into an on-topic one.
    – Bart
    Jun 6, 2013 at 7:23
  • @Bart: But I've not asked for opinions, I asked for facts. Excuse me if I ask, but have you read the question? You answered me in 10 seconds... Jun 6, 2013 at 7:25
  • 2
    When asking for a comparison of software products, you automatically ask for (educated, fact-based) opinions, even though you claim you want "facts".
    – Pekka
    Jun 6, 2013 at 7:45
  • Your question is also extremely broad. You might get away with comparisons of specific features, based on a precise scenario you have. But as it is, you'll need pages and pages' worth of comparison data to address all your point.
    – Mat
    Jun 6, 2013 at 7:49
  • Yes I have. Quick reader here. You can ask for facts all you want. What your question boils down to however is something close to a recommendation. Or even a large list of comparisons between the platforms. Very interesting information, but simply not what Stack Overflow is about. There is another site that comes up here often as an alternative recommendation for such questions. But I can't for the life of me remember what it's called. Anyone?
    – Bart
    Jun 6, 2013 at 7:50
  • @Bart: Even if I'm asking for something subjective in my answer, I think it satisfies the section "Guidelines for Great Subjective Questions": it invites to get me explanations, I shared my actual experience, I wrote it without flaming tones, and I asked for direct experience of answerers and external references. Can you tell me what parts of the answer do not render it a "good" subjective answer? Jun 6, 2013 at 7:52
  • @Mat: so do you suggest me to split the question? Jun 6, 2013 at 7:55
  • Perhaps it would help if you would imagine answering your own question. I wouldn't be surprised if this ended up in a lot of "I find that", "in my experience", etc. Then, as @Mat said, it's very broad. There is not a single question to answer. There are many. Some of them seem to ask for personal experiences. Those experiences may have factually taken place, but that does not make them the type of "facts" we refer to. Different users will have had different experiences. They are all fair enough questions to have, but SO is not the platform for them.
    – Bart
    Jun 6, 2013 at 7:58
  • @Bart: in the link you posted me there's a short list of requirements for a "good" subjective question. So for what I understand you can ask subjective questions if they are well written. Jun 6, 2013 at 8:01
  • @LucasMalor The idea indeed seems to be that if you manage to write such a question with clear boundaries, two which only one or two answers are ultimately possible, you might not have a bad question. That said, even in such cases the community often decides otherwise. We have had such questions here on Meta where a moderator said "this should be fine, I'll reopen it", only for it to be closed again. It's tricky. Those guidelines should help, but I can offer no guarantees.
    – Bart
    Jun 6, 2013 at 8:04
  • Odd that nobody has stated the fact that even if it was acceptable in a good-constructive way it would still be off topic. We don't select, install and manage server software here, we write server software.
    – user1228
    Jun 6, 2013 at 12:53

1 Answer 1


Let's try to break down the several sub questions in your question:

Apache, well... is Apache. Inertia pushes me to it. But the other servers are very interesting, since they seems to be more high-performing. Anyway I have not found a good benchmark test.

The reason you have not found a good benchmark test is that there isn't one. And there isn't one because performance depends on project specifics. We can't possibly tell you which of the four servers would be more suitable for your project performance wise, if we don't know quite a few technical details about your project.

Each of the three servers (I haven't worked with Cherokee) outperforms the others in certain tasks. What are the tasks you are interested in?

The other servers use their own modules, but is there some significant module available for Apache and not for the other servers?

Define significant. What's significant for your project will probably not be for mine, and vice versa.

Is this a big problem in a production server?

Depends on the project, and your skills. Also, define "big problem".

I would see if the vulnerabilities were patched quickly, but where I can find these data?

Check each server's issue tracker. I'm afraid you'll have to do your own research on this one. Also, whatever's true today won't necessarily be true six months from now. Open source communities aren't exactly stable in that regard.

Furthermore, how much is simple to switch to / from Apache? I mean, if I now decide to install Apache for inertia but in a future I want to try another server, how much is simple to do? On the contrary, if I decide to try one of the other servers, how much is difficult to come back to Apache?

Depends on your project, on how exactly you utilize your web server, and on your skills.

What are the main differences between Apache and the other servers (httpd.conf, .htaccess, modules), and how much they are simple to learn?

Again, you'll have to do your own research. Start by checking each server's documentation. "how simple" is unanswerable, depends on your skills and your definition of simple.

Are there some other aspects I'm missing?

Yes, there probably are. However, you've already asked quite a few questions in one, even if your sub-questions didn't have other problems, the overall question is already an extremely broad one. One question per question would be both easier to answer and a tad more constructive.

I'm afraid the only answer to your question is "depends", and that's not really what Stack Overflow is for. We prefer questions that can be answered, preferably definitely, and not just discussed.

The only way I can think of improving the question is replacing all of it with a thorough technical overview of your project. We can't comment on which web server would be more suitable for you until we have a very good idea of what your project. That said, since your project seems to be a fairly small one, chances are that there's no objective reason to pick one web server over the other. If that's the case, just pick Apache (the more popular one) and be done with it.

  • Well, I did not write my specific requirements because I thought the question would be too much localized this way. It's not simple to ask a question on SO :-P Jun 6, 2013 at 8:22
  • @LucasMalor It certainly isn't simple, and it does get harder over time. That's the price of popularity, when you have literally tons of new questions coming in every day, you'll have to focus a bit and raise your standards. Otherwise this place will turn into a madhouse. Your questions are reasonable, and are more or less the same questions I had when I was starting out, however we simply can't answer them (in their current form) in a way that would actually be beneficial to you.
    – yannis
    Jun 6, 2013 at 8:28
  • 1
    @LucasMalor (cont...) To give you a (rough) idea: I could write a very long thesis on how Apache sucks in every possible way. Its given me tons of grief over the years, and I hate the damn thing with a passion. However, it also happens to be the more suitable choice for a variety of projects. A lot less projects than 10 years ago when there wasn't any other option, but still a lot of projects. What use will my rant be if your project happens to be one of those?
    – yannis
    Jun 6, 2013 at 8:29

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