0

I think we should allow more questions. Specifically we currently have a very broad definition of "Asking for recommendations"

It seems to me that questions will be downvoted and close voted if they mention technologies by name, where it falls into "We are not tech support for X". Or if they try to avoid mentioning names, where it falls into "Asking for recommendations".

I suggest that we specifically allow questions which might be loosely termed "State of the Art?"

For example;

  • "Is technology X suitable for purpose Y? (or should I use an alternative)"

  • "What is the current 'best' technology to solve problem X"

  • "Are there any packages/libraries/software which can do X? (or do I have to do it myself)"

It's my belief that with these type of questions the OP is not asking for opinions on whats peoples favorite tech is. Or a list of different products and their pros and cons.

They have come across a real life problem and are asking if other people have experienced the problem and how they solved it.

We have to admit that most software is built on various 3rd party tech stacks. Asking about and comparing those technologies is a major part of day to day software development.

For example:

Windows based user interfaces that use vector graphics

Technically they are asking for a product recommendation true. But the underlying question which is answerable is:

"What's the current state of play when it comes to using SVG in desktop applications? Is it possible, what approaches work etc"

The answer could be very clear

"No, windows has very bad support for this for reason X. your only option are third party components like Y, but they tend to be slow"

OR

"Yes, there is very good support. but you must ensure X and Y"

  • 1
    For the 3rd point, we already have Software Recommendations. – Andrew T. Apr 5 at 9:14
  • I would say that the assumed "(or do i have to do it myself)" is critical there. The question does not want a software recommendation, even though the phrasing suggest that. They want to know about possible solution architectures – Ewan Apr 5 at 9:27
  • You make a good point. Obvs I am far too thick skinned to notice, but new questioners are put off time and again – Ewan Apr 5 at 16:26
  • @DocBrown: not on meta: softwareengineering.stackexchange.com/help/whats-meta - "voting indicates agreement or disagreement with the proposed change rather than just the quality or usefulness of the post itself." – whatsisname Apr 6 at 3:10
  • @whatsisname: ok, that was helpful - I really was not aware of this special rule for meta. Maybe we should make a feature request to show a different tooltip for the up- and downvote buttons on meta sites? Currently, it does not reflect the information from the link you gave. – Doc Brown Apr 6 at 6:49
  • ... unfortunately, IMHO still too many people here seem to think the same voting rules apply to the main site. – Doc Brown Apr 6 at 6:57
4

"Is technology X suitable for purpose Y? (or should I use an alternative)"

I think this could work, as long as purpose Y was specific enough and there was enough context around the situation. There's some risk of it being out of date as technology X either matures or is replaced by technology X2, but it's probably lower compared to other time-specific question.

"What is the current 'best' technology to solve problem X"

This type of question is very time-specific and may become out-of-date without curation. It also requires a lot of context around exactly what problem X is. I think these nuances become hard to explain, so it's best to simply discourage this type of question.

"Are there any packages/libraries/software which can do X? (or do I have to do it myself)"

This type of question will likely get people posting their specific choices. You get a large number of answers which may be simply links to a package or library or tool and voting based on popularity. I would recommend Software Recommendations for this - and if the answer is that none exist or the ones that do exist aren't suitable, the default answer for the asker becomes to do it themselves. The recommendations, however, may be useful for someone else in the future.

3

I disagree on allowing 3rd party resource questions for the "State of the Art" criterion you suggested. These are often just "best practice" questions in disguise, with no clear specification of the problem or what they mean by "best". Moreover, I think todays "State of the Art" is often deprecated next year, and good software engineering questions and answers should be valid for a longer period.

However, there are definitely good questions where askers specify a conceptual, on-topic problem, and it is not inherently clear if solutions should incorporate

  • a tool-agnostic approach they have to integrate into their current way of working

  • available tools, libaries or other 3rd party resources

(or a combination of both). These are often questions I would like to keep here.

Frequently, the asker has an idea such a tool or library could exist and makes the "error" of mentioning this idea. Maybe they also mention some of the tools they found, which I think is fine, since it shows some reasearch effort. Unfortunately, this immediately triggers downvotes and close votes.

To be straight - I really dislike this voting behaviour - it shows many voters fail to read and understand questions in full. They use the site like a computer game, where it is the goal to scan a new question as quickly as possible, find buzzwords in it which correlate to the predefined close reasons, and then stop reading and vote for closing:

  • "is there a library or tool ..." - somewhere at the end of the question - probably leads to a behaviour like "hey, lets shoot at it with the tool request close reason, regardless how good the former parts are. No further thinking required, great!"

What I even more dislike is, if such questions get downvotes, but no close votes and no comments.

My personal approach here is to ask myself,

  • "would the question be on-topic and focussed enough without the 3rd party resource request"?

If yes, I sometimes try to edit those 3rd-party request out, to make "this buzzword-triggered voting" less likely.

I would be really happy

  • if that would not be necessary, because it takes time I could invest into better things,

  • if others with enough rep for editing questions would follow my example.

I currently see no better solution. I don't think we should weaken the rules in general, and deciding if a "3rd party request" in a question is a clear rule violation, or only a minor, unimportant issue, is definitely to some degree opinionated. Moreover, we cannot educate the community here to change their behaviour how the read and reflect questions, that's just a tilt against windmills.

Or if they try to avoid mentioning names, where it falls into "Asking for recommendations".

"Asking for recommendations" - for a clearly specified problem - is and was never a close reason, and I don't think the community treats it that way. "Asking for (a list of) recommendations" should stay a close reason, these kind of questions are almost always too broad. And "Asking for 3rd party resource recommendations" should IMHO stay a close reason as well - when this is the primary focus of the question. Our site is not a replacement for Google.

To your example

Windows based user interfaces that use vector graphics

This question may fall indeed into my initial category of 3rd party resource questions, where it is not clear to the askers if a programming approach is required, or if there is an existing framework for the job. However, the problem I see here is, the OP did not show any research effort. Someone wo claims that they actively use Winforms should already know the canonical alternative WPF, and could have easily informed themselves that it solves exactly the problems mentioned in the question.

  • hmm see I interpret the question as not a request for "use software X, its great!" but a question about general approaches to the problem and what works. No-one would suggest writing your own SVG library to solve this problem, so the question is framed in terms of "what software can i use?" – Ewan Apr 7 at 12:11
  • @Ewan: if someone asks "what software can i use", the expectation of an answer like "use software X, its great!" does not sound unreasonable to me. But however one interprets the question, it does not change the fact OP did not do their homework. – Doc Brown Apr 7 at 13:29
  • are you sure? this SO answer for example got 100 up votes stackoverflow.com/questions/3526366/… – Ewan Apr 7 at 13:38
  • @Ewan: the very minimum I had expected from the asker is a statement like "I tried WPF for doing this, but I found the following limitations - <include limitations here>..." And I would not treat SO questions from 2010 as a representative example for a good SE question. – Doc Brown Apr 7 at 13:43
  • I would challenge you on that, that SO question for example is a single sentence. Why do we put such high bars in front of potentially good question/answers when other sites upvote? not a single downvote on SO for that one line question – Ewan Apr 7 at 13:49
  • @Ewan: honestly, I think you are comparing apples with oranges. That other question looks like a very focussed programming question to me, and though it has only one sentence, it gives me the impression the askers had already tried out what they were asking about. It is definitely not a question "which library should I use for this". The SE question, however, did not even mention the existence of WPF, which means the OP did not even put 10 minutes into any thoughts of solving the problem by themselves. – Doc Brown Apr 7 at 13:57
  • ... or do you think we should start to weaken the rule the on this site that askers are expected to do some minimum research on their own beforehand? – Doc Brown Apr 7 at 13:59
  • I think its an assumption that the OP didn't do any research. The answer obviously isn't as simple as "Use WPF, it just works" – Ewan Apr 7 at 14:11
1

The "Request for Recommendations" close reason has a fairly specific purpose: to close "I need a good book or blog, do you know of any..." questions.

The problem with technology recommendation questions is not really that they become a big list of things, or that they encourage bikeshedding and "me too" answers. Those are just side-effects of the real problem:

Recommendation questions are often under-specified.

Which is to say that they're too broad. They are too broad, because not enough relevant detail has been provided in the question to make these questions answerable in a meaningful way that is relevant to both the asker and to other community members.

Often, such questions are asked in the context of a much larger picture: Which technology should I use to [do this thing?] Nobody argues that such a question is too broad: "It depends." And so you wind up in a protracted discussion, trying to figure out what it depends on.

Many of these questions lack the necessary prior research and analysis to make them good questions here. How big is your team? How large is your project? What technologies are you currently using? Has your team evaluated any of the available technological options to determine if they are a good fit for your specific needs?

As you can see, the problem is not that these are recommendation questions; they are closeable for all of the other usual reasons: Too Broad, Unclear What you are Asking, etc.

The answer to many such questions is "do some research. Find the two or three most likely technological candidates, and stand up some simple prototypes. When you're done with that exercise, it should become very clear which option is the best one for your particular project."


All that said, if a question is clearly asking for a technology recommendation, but is otherwise squarely on-topic for our site, well-researched, and has enough detail in the question to identify an adequate answer, I am inclined to leave it open.

A well-written recommendation question will encourage answers that specify a technology recommendation and explain why it is the best choice for the asker's specific situation, not merely an answer stating "here's my favorite."

  • ..so are you saying that the referenced question should not have been closed for that reason? – Ewan Apr 5 at 15:23
  • No, I'm saying that most poor technology recommendation questions don't need the "recommendations" reason to be closed anyway. – Robert Harvey Apr 5 at 15:41
  • well, then we can alter the definition without problem then – Ewan Apr 5 at 15:42
  • It's not about the close reason definition. That's the point I'm trying to make. – Robert Harvey Apr 5 at 16:13

You must log in to answer this question.

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