How should we treat answers that don't fit all requirements?

Question

On the question proposed by @juergen d where he asks about XML inspection tool, the only answer (so far) is one answer that doesn't fit 100% of the requirements, but provides good information and may be helpful for another person that is looking for something similar.

The answer itself has good information about the software, that proposes as an answer to the question.

There were a brief discussion in the chat about this subject, but I think we should talk through on this matter, since this may occur with a certain frequency on the SoftRec SE.

Answer 1

Robert and Flyk have it right, together. Robert is quite correct in his concerns surrounding spam, guerrilla marketing and the like. However, our worst possibility is also our best when it comes to that, there will be times when people paid to market software outright nail a question with a good, comprehensive and quality answer. Moderating this site is going to be interesting.

The stuff we want to get rid of is typically spotted by noticing a lack of engagement with the question itself, a lack of adherence to our quality guidelines, or both. Acceptable answers on this site will address a bullet list of requirements and constraints defined in the question. Great answers will meet or exceed every single one of them. But what do with the possibility of nothing meeting the question author's expectations? Should we simply leave those questions unanswered?

I tend to think no. If you know of something that meets almost every requirement and clearly state that what you're recommending does everything but something, and sufficiently engage the question, it might just be worth the mention. Now, this isn't to say "Hey my thing kind of does that so I'll put it here" - any time you don't meet the constraints of the question, you need to start your answer with "I don't think such a thing exists yet, but this does most of the job."

The example answer, while obviously written with good intentions has some problems. It doesn't address at least two of the must-have points, and it's clear that the author of the answer wasn't clear on some terminology being used. It appears that some clarification should have taken place prior to the answer being posted. I'm not sure that it comes close enough.

We have to come to terms with the fact that sometimes there will be no feature-complete solution. This is particularly important when it comes to free and open source software, where someone conceivably could write in a missing feature. Note, I said a missing feature, not 'roll your own Excel'.

I found this answer to my question about monitoring my Internet connection to be extremely helpful, in fact, I'm using the solution as prescribed right now. The thought of seeing such a valuable contribution deleted just because it won't run native on Windows 7 makes me very nervous.

Answer 2

An answer that meets our quality requirements while not completely satisfying the requirements stated in the question are still answers. I say this for a number of reasons:

• Sometimes an asker may have been too specific, there may be simply be no software that fully meets their requirements, as stated in the question
• After some period of time, the asker may decide that they're not fully in need of all of the requirements stated in their question
• The recommendation that doesn't meet all of the requirements may in fact turn out to be a higher quality piece of software
• An answer that meets most of the requirements may still be useful for future visitors, even if the person who asked the question is missing some functionality that is key to their decision

The answer should be kept simply because the person who answered the question both made a valid recommendation and met our quality guidelines.

This, of course, is where down votes come in. An acceptable answer is still an acceptable answer even if it is down-voted. A down vote merely means that the answer wasn't useful for the person that chose to down vote it.

Answer 3

I don't mean to speak in absolutes, but unless the author is clearly offering a persuasive argument that their software is "close enough" to actually answer the question, we should flag it as

Not an answer

We have to be really, really careful here. Folks become very energized when you start talking about their favorite software. That will likely lead to a lot of "hey, that reminds me to about {this} software, too."

But my bigger concern is when the bad actors start to arrive. If we allow a broad range of "somewhat related answers", this opens door to software vendors and enthusiasts trawling this site, looking for every opportunity to post/advertise their products on every question that has even a modicum of plausible deniability of feigned interest.

Overall, I think we should err towards requiring that answers actually answer to the question… as stated. After all, we ask question authors to be very specific about their problem statement; answers should be just as rigorous.

It has to be so, or this site is going to be a guerrilla marketers' haven.

Hey, you asked about kid's learning tools.
I think you will be interested in our nanny software. It's really great!

---

• @user Thank you for the answer, but unfortunately your solution doesn't quite fulfill the specific requirements outlined by the author. While your answer is well-meaning and somewhat related, we have to remain extra vigilant that we don't allow these recommendations to expand outward where folks will be posting "other" software under the premise: "you might also be interested in..." — Robert Cartaino♦ Feb 17 at 23:02

<post deleted on Feb 17 at 23:03>

Yes, that's a really obvious example, but it's a realllllly long continuum between blatant astroturfing and "I think people will be interested in <my thing>, too."