Review of deleted answer where a generic rather than specific model may be appropriate

Question

I posted this answer (now deleted) to this question, which was deleted as "question asks for a router, this is a hardware not a software recommendation site". I'd like to ask for review, but I also think there are a couple of sensible discussion points.

1. In this question, the user may have been looking at their issue in a suboptimal way. They may not actually need a new router to do what they wanted (even though they assumed they would and asked for one), and it would be a disservice to not let them know if that could be the case. It would be perverse if we could recommend hardware but not tell them we think they could actually be looking at their issue wrongly and might want to look at it a different way.
   
   A non-technical user might well think "my laptop is slow, what new laptop should I buy", where it's likely from their description it's their network speed or disk that's the issue, or the version of the OS they are running is known to have speed issues fixed in a later release, and they don't need a new laptop at all - I think an answer should be able to say if the user sounds like they may actually be looking in the wrong place, or for the wrong kind of solution.

2. Sometimes there are so many choices available, because the features needed by the OP can be met by generic hardware. In that case I feel it would be unfair to point to a specific product, but a list may be unreasonably long and unhelpful. Sometimes the best answer for buying hardware will be "look for these criteria in the spec, and anything that has them will meet your needs"_. So for example "I want a computer I can do X on" might be best answered as "any computer that has an Intel i3, i5 or i7 processor (of any generation) and 8GB or more of RAM will be more than enough for that". The best answer for the user can sometimes be to give them an idea how they can find a suitable model for themselves if they truly don't have other criteria. Can such answers be acceptable here?

Turning to the specific question, the user clearly believes they may need a new router of some kind, and yes, that's one option. Routers with native 3G/4G and support for specific VPN software might only come with premium models, or packaged with expensive features they don't need. But it's quite likely they don't need an expensive router (or perhaps any new router at all, depending on their situation), and that if they knew soft routers existed, they could look on a hardware compatibility list and choose a far cheaper hardware model which meets all their needs. It feels misleading for me to know this could well be useful to them, and not say so.

The answer itself couldn't recommend a specific router, because the features they want aren't built in to most routers unless expensive, but they are available very cheaply via a generic router with FOSS firmware. The router and 3G/4G hardware I would recommend - the actual hardware recommendations - would be "any generic router off the OpenWRT compatibility list that supports 3G/4G USB dongles, and any 3G/4G USB dongle supported by them" - and that answer is a hardware recommendation that would meet all their requirements and probably cost a lot less than other solutions. If OpenWrt is inappropriate for them, they would then say so, and the discussion would continue, but they (and future users with the same question) would at least have the knowledge to make a hardware choice that is best for them.

So I think it's a valid response to the OP. Can this be undeleted, or at least, the principle discussed?

Answer 1

I've provided an answer to a similar question before. A software recommendation should be allowed if:

• It's not link only/"I Googled this for you"/Answer without explanation
• It is verbose and explains how/why this meets their requirements
• It's nicely formatted

Your answer is very close to meeting those, but I don't think it quiet gets there. You provide a high level overview of what to look for - two pieces of hardware: a router and a 3/4G dongle.

I think you're answer would be much better if you took that extra step and made a recommendation of those two pieces of hardware. With links to those two devices and an explanation of how using OpenWrt will solve the problem, the answer would meet the criteria above.

It also keeps the spirit of your post - use OpenWrt instead of spending hundreds/thousands more to get a single device that meets the requirements.

Answer 2

I deleted this answer because it didn't recommend any hardware to the specifications of the question. At the moment, our policy is that answers must recommend hardware that matches the request in the question to at least some extent. This answer instead recommended a piece of software for use with existing hardware; it didn't answer "what router should I buy?".

I'm loath to change that policy, though I am open to it (and of course, my opinion is by no means final - that's a community decision). I suspect it'll result in a lot more rules-lawyering to try to sneak in answers that don't really address the problem, purely because "it could be a different problem".

Take your example - "my laptop is slow, it takes hours to copy a few files, please recommend me a new laptop" (with all the necessary details, of course). Sure, it might be the case that I'm a non-technical user and have just misidentified the cause of the problem when it is in fact just a disk issue. However, it might also be the case that this is the last straw in a string of problems with my laptop, and I really do need a new one (but just haven't mentioned that, because the copying issue affects me the most). Or, of course, it might be the case that I am a technical user, have troubleshooted the problem, and identified some other component as the cause - but, this being a laptop, can't easily replace that component, so I've made the choice to get a new laptop instead.

My point is that knowing the exact situation of a question's author is really difficult. Without additional context that we can't get, we have no other information to go on than what they've given us in the question. For that reason, I think answers here should just address the question. If, going back to that example, I'm a technical user and have done my homework, and someone comes along with an answer saying "this is probably your disk, I recommend this disk instead"... I'm not going to be pleased, and I'm going to flag that as not an answer.

I do see the arguments for allowing alternative answers based on an answerer's technical expertise: allowing answers like these potentially makes us more helpful to some people by correctly identifying a problem they've missed. However, I don't think the extra value provided by allowing that is worth it - we'll probably see an equal number of cases where an answerer misidentifies the problem because they don't have enough information, and the question's author gets annoyed because "that stupid answerer didn't read my question, grumble mutter".

There's also the question of technical knowledge. If, say, you identify an issue that the question's author missed, like "that's a disk issue, you only need a new disk", then sure, you could recommend a new disk. But if the author missed identifying that issue, what's the likelihood of them having the expertise to replace a disk in a laptop, without losing any of the data? I'd say it's fairly low.

For this reason, I think answers should stick to answering the question as stated. If you have a recommendation that answers the question (say, a new laptop), and you want to provide extra advice, that's awesome - you could say "[this] laptop will suit your needs, but be aware that you might not need a new laptop, as it could be a disk issue..." (paraphrased, but you get the point). However, if you don't have something to answer the question, I'd advocate just leaving it in a comment.