16

Going through @Gareth's list of candidates for closing, I find several topics that are asking about the best X, with X being software.

We're usually pretty liberal with product recommendations if it's software. It's difficult not to, because most I'm looking for an X that does Y questions cannot be answered otherwise.

Hardware with its shorter product cycles and usually much larger number of options is different, and many products are very similar, being based on the same components, making recommendations for one but not the other often rather pointless. And it's a lot more difficult to find some obscure piece of software, than some obscure piece of hardware you can actually buy. Just browse computer stores and the shopping sites: if it's not there, you can't get it.


AFAICT, we're happy to answer (and have) questions like:

  • I want a software that does X
  • I want X but for Linux
  • I want something like X but without its annoying characteristics
  • I want a tool like X but free

Several of the topics we are voting to close right now are very similar: I want the best X. Now, of course, it's subjective what's the best X is. But users have always voted on and accepted as answer what they liked most, i.e., according to their criteria, the best. Likewise did other users answer primarily what software they like most that fits the criteria (i.e. being an X).


As an example, consider https://superuser.com/questions/21155/best-quicksilver-clone-on-windows:

Wouldn't we tolerate it, if the user instead asked:

I'm looking for a tool that allows me to launch applications quickly (with hotkeys) using the keyboard? Doing basic calculations, note taking and creating new documents would be a plus.

And what exactly would be the difference? Quicksilver is a well-known program that does this (and more), so it's rather obvious what the user is asking about. In other cases, explaining what something does is far more work (and not any more helpful) than giving a well-known example.

And he doesn't want the tool someone hacked together in half a day, so he's asking for the best. But does this really make any difference? It's no different than asking to please not suggest crap that doesn't work. The user'll still try more than the most up voted tool (unless it's perfect for his needs) and accept the answer best suiting his needs.


In a way, I like questions like the example above even more than some of the more obscure software recommendation questions. Topics like these can help a lot of people and make them aware of competitors of something they're already using. Some of other recommendation questions are really only helpful to the person asking (even though they cannot be considered too localized). OTOH, they suffer from serious up votes for popular answers, giving newcomers little chance to be discovered.


So how is it we consider What's the best X? to be off-topic, but accept questions like I need recommendations for an X? Aren't they the same thing?

  • 5
    Quicksilver is a well-known program that does this - I had no clue about QuickSilver till now. Popular niche program is not always known to all – Sathyajith Bhat Sep 28 '11 at 6:51
  • Since no one else has done so yet, I thought I'd just add a link to the most relevant blog post on this general topic, for reference: Q&A is Hard, Let’s Go Shopping! – DMA57361 Sep 28 '11 at 10:39
  • @Sathya There's a Wikipedia article, so it's somewhat known. From that article, a top 3 Google hit, you could quickly determine what it's about. – Daniel Beck Sep 28 '11 at 18:17
  • 1
    @DMA For some reason, all examples in that blog post are hardware. I mentioned hardware and why questions about it are generally not useful. Also, there's no free hardware. There is free software. – Daniel Beck Sep 28 '11 at 18:38
  • @Daniel Software is a bit of a different ball game, yes. You can't really rephrase most software req's into "how do I tell if software is great a X". I could support hardware questions that meet those guidelines - how to analyse and make a decision without actually making one (and thus not really a recommendation question) - but I'm yet to be convinced software rec's are similarly worthwhile, I just didn't add that to the comment above as it was already an answer (which is to say, I support Diago's point at this stage). – DMA57361 Sep 28 '11 at 18:56
  • 1
    @slhck In addition to Diago's comment, I'd add that there's sometimes been resistance to deleting old-but-once-acceptable questions of "historic value", which is the reason for the new post annotations appearing (admittedly the problem is much stronger on SO). Community driven efforts, like Gareth's recent post, seem to be more successful at this - they get more constructive input and less people complaining than if a few ♦ just went on a massive clean up. – DMA57361 Sep 28 '11 at 21:48
  • @DMA57361 My point is that it all seems like there was a rule introduced I haven't seen yet. I'll write a new meta post, it distracts from my issue. – slhck Sep 28 '11 at 21:53
  • This covers some of the same ground as How literally does SU take "not about a shopping or buying recommendation?" – Pops Sep 29 '11 at 12:29
8

As noted in

http://blog.stackoverflow.com/2010/11/qa-is-hard-lets-go-shopping/

Such posts have a lot of pitfalls. In general they are bad news.

Is it possible to write a recommendation question that can survive? It's difficult, but to even begin, they need to be very clear, very specific, and narrow enough that they won't get an infinite list of answers over time.

  • 2
    Software isn't as quickly outdated, a major reason to not allow them in the blog post. All examples in that post are hardware. The FAQ clearly states shopping/buying, neither of which applies to free software/freeware. If this is really policy and, as a rule, we allow no recommendation questions at all, you really need to update the FAQvor properly document the rules somewhere. I managed to participate for over a year and be unaware of that, and sent my share of users to read both your blog post and the FAQ... (I'm stupid, i know, but probably not the only one) – Daniel Beck Sep 30 '11 at 5:54
  • Regarding very specific requirements for questions: there are no variants or products based on basically the same components that behave quite similarly. The issue of needing to know more about a user's situation before being able to recommend a specific program comes up rarely, in my experience. – Daniel Beck Sep 30 '11 at 5:58
  • Nope, you're not the only one, @Daniel -- at least the FAQ should somehow reflect that, which it does not as of now. – slhck Sep 30 '11 at 8:40
  • Also, I think it's very important to take into account the context of the problem that needs to be solved, as mentioned in my answer here. – slhck Sep 30 '11 at 8:42
  • 2
    I always thought the FAQ did reflect this, and that the community/mods here just ignored the rule. To be honest it's always kind of discouraged me from using SU. – Pops Sep 30 '11 at 14:54
5

Any recommendation question creates a slippery slope. If we allow one type of question, which borders closely on another type, we open up the list of, why was my question closed but the others left open issues.

We've been very clear defining what is and isn't allowed, and any type of recommendation supposed to be off-topic. Hardware one's date quickly, but software one's also have issues with version upgrades, and requirements being specific to a scenario, so do we need one for every type of requirement each individual user has?

As Jeff pointed out before, if the question is well thought out, and proves to be off high quality, it's worth reviewing, however in general I feel these should be closed as not constructive.

  • 1
    I'd tend to agree with this stance. Some of those questions might have high views and high-quality answers, in which case they could be edited to better fit our format. But overall, we need to be careful about specifically granting exceptions which might set a precedent for subjective and unhelpful questions. – nhinkle Sep 28 '11 at 7:00
  • Software rec questions aren't as easily outdated as hardware rec questions. What are you referring to by Both? – Daniel Beck Sep 28 '11 at 7:39
  • @DanielBeck I have updated the answer to state any recommendations should not be allowed, period. – BinaryMisfit Sep 28 '11 at 7:43
  • 1
    @nhinkle I disagree with regards to view count, since it could be driving the wrong type of traffic to the site. However quality is a much higher value, and if the user spent time asking a good question, it is worth reviewing. I have seen a lot of bad questions in the past generate huge view counts, without deserving them. – BinaryMisfit Sep 28 '11 at 7:46
  • @Diago I totally agree - I'm not saying that view count alone should be seen as an indicator of questions to keep. I just think that questions which have high quality answers and have a lot of views indicate that it's useful to a lot of people. It should have both, not just high views. – nhinkle Sep 28 '11 at 17:22
  • Is it a shopping recommendation if it's asking for freeware or "open source" programs? Is that really where we'd draw the line? – Daniel Beck Sep 28 '11 at 18:19
  • 2
    @DanielBeck The key is recommendation, not the shopping part. If your asking to recommend something to perform a task, your stepping on dangerous ground. Also, open source or freeware is rarely ever, well, free. Thing is, we can debate it on meta forever, but the policy is in place network wide, and I can't see it changing anytime soon. – BinaryMisfit Sep 28 '11 at 20:47
3

Update: This was written under the impression that we actually allowed software-rec questions. Hey, there are over 1000 of them as of now. I am talking about how I think a good software recommendation question can still exist.


In fact you are right that there is a sort of ambiguity.

As an example, consider Best QuickSilver clone on Windows:

Wouldn't we tolerate it, if the user instead asked: […]

Exactly. At least from what I've seen in the last months, software recommendations were actually fine as long as they were based on facts (i.e. certain features, measurable speed, user interface constraints, versions, open-source, price). And since software doesn't rapidly change as much as hardware does, those were fine at least at the point of writing and the following few months.

However …

For most of the old software recommendation questions, it's just the usual problem of clutter, lack of consistency, missing context, et cetera. This is a problem with the answers, not the question per se. The question might be driven by a problem many people have.

Sooner or later, this will happen to all of these, if we don't take care of them.

What to do in the future?

So, in the first instance, whenever you see a subjective recommendation-type question …

  • Don't close as not constructive immediately.
  • Try to salvage it. Ask the OP what their criteria are or emphasize them (maybe even in the title).
  • Edit the question accordingly.

Even more so, I think software-recommendations (as long as they're not "What do you use?") can be worth keeping. Maybe we should set up a few constraints for answers, like for example:

  • Including name, official website
  • Including estimated price, license
  • Including context. Let the OP know why this meets their criteria. This is the most important part.
  • Including a screenshot, if possible.

So, I don't think all of these should be closed. There is still a marginal difference between "What's the best XY?" and "What's the best XY for this context", if and only if the answers also reflect that.

  • 1
    In the end I refer back to my point 1. If we keep 1 open, we create a moderation issue every time we close a bad question. The policy about recommendations are in place network wide, and debating whether it is software vs hardware is irrelevant. These questions draw the wrong attention, and end up polluting more then they do good. Finding one gem in a hundred questions is not a good guideline at all. – BinaryMisfit Sep 28 '11 at 20:49
  • 2
    But why are still there so many around? For as long as I've been here, I personally haven't seen the major part of them closed as long as they were based on somehow measurable criteria (as Daniel mentioned). Don't you think it would make more sense to actually enforce such a rule rather then letting any software-rec question slip through (if this is the "official" policy)? – slhck Sep 28 '11 at 20:59
  • 1
    Cleaning up is a tedious job, and generally left to the bigger community. To be honest, they should be closed, but we have not brought to ban hammer down hard, generally, because there are other more important duties to focus on. Good questions should stay, but we don't always apply policies retroactively. If we did, well we would never be doing anything else. There is a reason there are community moderators and there are diamond moderators. Cleanup should be left to the community. – BinaryMisfit Sep 28 '11 at 21:20
  • 2
    Well, to be fair, I've just never seen it like that (like hardware-rec questions which are more "aggressively" shut down for obvious reasons). Mentioning "retroactively", that implies that there was a policy introduced at some point, right? Somehow I must have missed that, since I never heard anyone say "Software recommendations are off-topic!". – slhck Sep 28 '11 at 21:31
  • 1
    @Diago If the policy is in place network wide, why does Apple.SE only prohibit hardware recommendations? Quoting the Apple.SE FAQ: a shopping or buying recommendation for hardware – Daniel Beck Oct 18 '11 at 20:13

You must log in to answer this question.

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