Favorite questions and answers of 2015

Question

I think we spend a bit too much time talking about what's wrong with Programmers.SE, and who is or is not to blame for it. In particular, it's probably fair to say that some of the curmudgeons active close voters like me have a better understanding of what's off-topic here than we do of what's on-topic.

So let's try talking about this site's good questions and answers for a change.

You may include as much or as little commentary on your choices as you wish, and present them in any format you like, but try to stick to constructive, positive commentary. No tangential rants about how terrible other questions are or whatever.

_{For the record, the specific idea of a "favorites poll" comes from SFF.SE.}

Answer 1

Lord Loh.'s "How is spinlock different from polling?" was one of those questions that made me think "Huh, he's right, what is the difference?" when I first saw it. But after reading Jorg's answer, it makes perfect sense to me; I feel like I learned something.

pduersteler's "Git branching from a feature branch to work on a subfeature" Most git branching questions strike me as people confusing themselves by overcomplicating things, and the answers often seem to only make things more complicated, but MichaelT's answer was surprisingly clear, to the point that it "feels obvious" in retrospect.

David Moles's "Why does(/did) Bertrand Meyer think subclassing is the only way to extend a “closed” module?", answered by gnat is a rare example of a question about a very abstract, high-level design principle that manages to be interesting, answerable and show genuine research effort.

Here are two questions where a (low-rep!) OP did everything right in asking their question, and as a direct result they got exactly the answer they needed. This is the kind of Q&A that makes me think "the system does work!"

• TBridges42's "Is a large static initializer a code smell?", answered by amon
• James Ko's "Is it possible to avoid abstraction inversion in this scenario?", also answered by amon

Finally, some high-quality answers that convinced me not to close-vote questions which initially seemed too broad/opinion-based to get a good answer:

• Thomas Owens's answer on BЈовић's "Where to describe architectural problems?"
• MichaelT's answer on Kimble's "Estimating in groups - How to extract useful statistics / estimates?"
• MichaelT's answer on Carcigenicate's "How does a skip list work?"

Answer 2

Some good questions or answers that I thought were astounding this year:

Just downright interesting: Which language is a BIOS written in?

Great question, felt compelled to write an answer myself amongst other great answers: Where to describe architectural problems?

Are there any significant disadvantages to depending upon abstractions?

I really liked this question because it garnered some very thought provoking answers. This is one of those questions where I now think about the persistence of data and highly sensitive data in a different way than I have before. One of those questions that attract answers where you realize that you might have a tool in your toolbelt to handle this kind of situation, but you think pretty deeply about why that is great tool and what features of that tool make it great. I liked the one answer so much that I put a bounty on the question just to reward the one author. No central database

Not a high view count question but I think it was a good question with a great answer: How to store (and find) updated values in a Distributed Hash Table?

More self promotion but everybody loves a good Agile question: Is it possible to take a flexible agile approach to projects that require estimates of both time taken and time saved?

Interesting thought exercise: Is a memory of all possible permutations of a kilobyte block and pointers possible?

Dates, calendars, durations, intervals, timestamps, epoch, timezones, all the fun stuff that Programmers since the dawn of computers have underestimated the complexity of only to be driven mad by the unaccounted complexity of doing any application development where Time is involved: Why isn't there a true "Date-Only" data type?

Good explanation of Merge Sort: Why is mergesort O(log n)?

Best way to consume REST service within Servlet on same server?

Simple question, simple excellent answer: what the best practice in REST to append or delete to a list

Another simple question, simple answer: UML : What is an Activity in an Activity diagram?

Interesting and thought provoking question about SOA principles with an excellent easy to follow answer (Yes this question wasn't posted this year but this answer was): Does SOA service composition actually work in practice?

Thomas Owens wrote a great answer here: Notifications in Use-case Diagrams

This borders on "maybe too opinionated?" but it is certainly a controversial question that attracted answers with up and downvotes. Thought it was interesting: Is this breaking SOA?

A good Q&A on the atomicity of unit tests: How to make one test depend on the results of another test?

Answer 3

I'm going to blow my own trumpet here but as the question was good, I feel ok with it. (parrrrrp)

Should You Write Your Back End as an API

this is a question hat encouraged some discussion (in a good way, not the usual "you suck", "no you suck" kind of discussion) to build up answers that took the questioner from a somewhat entrenched position to a better understanding where he accepted what we were telling him. I think this provided less of an answer and more of a guided education.