6
$\begingroup$

Looking at this question Bitwise flags in different languages - It basically asks

"How does your language facilitate working with bitwise flags?"

But the asker helpfully provides the answer for their own language in the question itself.

I can't help thinking the answer part should be added as an answer rather than to the question. If you do that though the question becomes rather weaker.

I raised a related question here Should we apply the same rules to question and answer pairs that we apply to questions?

In that question I note that Q & A pairs are sometimes voted down and closed for the question not being good enough in its own right - despite the answer being high quality. I was thinking about questions that only have a single answer. This is a different case however as it is soliciting multiple answers from the out. It isn't a self answer because its asking about prior art which I believe is on topic.

Are "How does your language do X?" questions on topic?

I think they should at least be "What are some ways to implement X?" but really that is just a synonym. Your language could be any language you work with.

Also is "do" too vague? Should such questions be narrowed to say syntax or a particular aspect of implementation to be sufficiently focused?

Should we encourage such questions to provide the example from their own implementation experience as this one does? and if so should we encourage it to be made an answer?

Related: What should be the quality standards for list-style questions?

$\endgroup$
5
  • $\begingroup$ I've separated my language case into an answer; guess it makes sense to not keep it in the question. $\endgroup$
    – Hydroper
    Commented Jun 25, 2023 at 13:07
  • $\begingroup$ That might not be the right thing to do! I have done that with some questions and they been closed as a result. You might want to wait for meta to build a consensus. $\endgroup$ Commented Jun 25, 2023 at 13:20
  • $\begingroup$ Right, rollback! $\endgroup$
    – Hydroper
    Commented Jun 25, 2023 at 13:21
  • 1
    $\begingroup$ Your answer hides some subtle hints as to what questions you want answered, for example, section headings about strictness and inference rules. $\endgroup$ Commented Jun 25, 2023 at 13:22
  • 2
    $\begingroup$ Also I should say I am not picking on your question in particular but the general kind of question. Yours may be narrow enough in focus that it is still valid either way. The votes may decide. $\endgroup$ Commented Jun 25, 2023 at 13:25

2 Answers 2

7
$\begingroup$

The answer should be in an answer, and not in the question. Having it in the question is confusing for readers, because it makes it appear that the information about VioletScript is necessary to understand in order to answer the question (when it isn't). Additionally, it prevents the answer about VioletScript from being voted on alongside other answers.

The question should also be reworded to avoid referring to "your" language, since an answer about a given language can be just as useful even if it isn't written by the language designer. People who know enough about the implementation details of languages which aren't "theirs" can write good answers too. "How can X be implemented?" is therefore a more useful question than "How is X implemented in your language?". I do not buy the argument that they are synonyms.

$\endgroup$
1
$\begingroup$

I agree with kaya3's answer but I want to emphasize and elaborate the point about "in your language" phrasing.

First, allowing questions like that to proliferate could create the impression that users here are in some sense expected to have personal languages that they're working on. While I see value generally in "gatekeeping" and don't really care about misguided third-party impressions of "elitism" (as people familiar with my participation on Stack Overflow Meta may infer), it's important when setting expectations that they are actually accurate. In reality, even if we circled the wagons as tightly as possible and only admitted "experts", it would still be perfectly on topic to discuss existing, popular languages. Someone who wants to create a new C++ compiler from scratch is absolutely in the right place here.

Second: language surveys of implementations are not useful insofar as they are about languages. On a conceptual level, there are nowhere near as many reasonable ways to "facilitate working with bitwise flags", as there are languages that do so. The natural underlying purpose of asking the question is to gather inspiration for ways to do it in a new language - i.e. it's really a design question. Any other motivation translates into seeking a list of language trivia - it doesn't make the question practical. The early days of Stack Overflow proved that such questions make for bloated, poor content and encourage treating the site software too much like a discussion forum.

Third, and closely related: because "in your language" is so open-ended (inviting a separate answer for every "language"), such a question invites redundant answers, which make the material harder to use for people who find the page later with a search engine.

Such questions should instead be edited to ask for the methods themselves, not about languages using them. (Aside from that, the original question is really a design question rather than an implementation question.)

In this case, "how can [the syntax of] a language facilitate working with bitwise flags?" Although now I think that still sounds unfocused: it doesn't identify why "facilitation" is necessary. It seems like OP is specifically concerned with issues related to static typing, but perhaps there are multiple such issues. Something like "how should I handle binary bitwise operations between operands of different enum types?" might be closer to the mark.

$\endgroup$

You must log in to answer this question.

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