6

After reading through some of the questions on Software Recommendations Meta (SRM), most notably:

I'm concerned about how as a community we will be addressing software recommendation questions and answers as the software we are recommending changes and as users ask similar but different questions.

What will we do when users do not take into account the date of a post and answer into their reasoning when upvoting or downvoting? This can ultimately lead to a false recommendation (upvoting) of a piece of software that no longer provides the feature(s) requested, or an inaccurate assertion (downvoting) that the software is a poor match for the feature(s) requested.

To prevent this from occurring, should we always be requesting users to specify the version number, and/or links to a feature list of said version to remove any ambiguity of whether said features are available at the time of the recommendation? Should we also encourage users to update their posts as new software versions are released to ensure that the poster of the question, and other users, understand that the recommended software has changed since the original answer was posted? And if we do this, will users become upset if their answers become community wiki posts/answers that they can no longer earn reputation for after too many edits, even though it is to help the community?

Perhaps I'm reading into this too much, but I want to hear what you guys think about this issue. I hope I have explained it clear, but if I have not, please let me know and I will do my best to clarify my viewpoint. Cheers.

  • 1
    The Firefox Quantum update is an example event that breaks nearly all answers on Firefox addon questions. – Nicolas Raoul Nov 22 '17 at 7:08
5

This can ultimately lead to a false recommendation (upvoting) of a piece of software that no longer provides the feature(s) requested

It is extremely unusual for features to be removed in new version. This shouldn't be a big concern.


To prevent this from occurring, should we always be requesting users to specify the version number,

Unless otherwised mentioned we can just look at the edit history to find a date, and can assume it referred to the latest version as of time of posting

and/or links to a feature list of said version to remove any ambiguity of whether said features are available at the time of the recommendation?

Features Lists are always part of a good answer

Should we also encourage users to update their posts as new software versions are released to ensure that the poster of the question

We can encourage with a comment, (I know on SE I tend to often go back and review my old answers). We can also post a new answer, refering to the updated version.

4

At some point, people who are sensible stop caring about reputation.

However, I'd doubt that a software would constantly change, to the point where it would go CW within a reasonable amount of time. If you're worried about that in your own question, you might choose to write your answer with terms like "As of $date". If its someone elses question, you can choose to comment with specifics - even if the OP dosen't change his answer, there's a clear record of it.

For example "As of $date software X supports bogodyne sumulation." or "Until $date, the free version of software Y, release 1.2 supported bogodyne simulation. The feature was moved to the premium version as of release 1.3, and as of $date is only available there" would be a nice way to avoid this ambiguity to start with.

2

Let's assume on day X someone asking for recommendation on software gets "Software Z" v1 suggested.

After a while, on day Y "Software Z" v2 comes out, and it has become incompatible with v1 (according to the OP's needs). Moreover, v1 is no longer usable, as it always asks for v2 update and does not start any more (.NET Reflector being a real life example I know of).

So, here we have a few possible situations:

  • The optimistic scenario.
    The OP is still using "Software Z", now s/he is not satisfied as of v2 and decides to look for an alternative. Then s/he un-accepts the answer recommending "Software Z", looks trough the other answers and picks a different suggestion. If there are no suitable answers, after un-accepting the OP might as well edit the post, saying that v2 of "Software Z" no longer meets his requirements, and s/he is looking again for a closer alternative. This would also boost the question as active unanswered recently modified one, and hopefully will get to the attentions of users.
  • The realistic scenario. The OP does not care for "Software Z" anymore, or s/he has replaced it from one of the other suggestions without reflecting this by changing the accepted answer. I think this will be more frequently observed, as not all users are that self-conscious, or when day Y arrives with v2 of "Software Z", they no longer use such software. This will render the whole question and answers "orphaned" in a way, that they are useful for reference only, or as a source of alternative suggestions to the accepted answer for other users. The other suggestions will probably be outdated too, so the entire question and answers to it might serve no purpose and have only historical value. Therefore, a new question is likely a better solution, and a naturally occurring one.

The second point inevitably implies that users or moderators may flag the new question as a duplicate. This has the potential to revive the original post if it is not entirely outdated, so flagging will serve its purpose. If the original question is indeed outdated, then the user must explicitly state that his question is not the same, as the old one has now become useless.

So, to summarize, both scenarios imply a form of engagement - either maintenance of an existing question or adequate reasoning when asking new question as to why it is important to be asked again. This will preserve the historical value of the old question and render a new one with fresh and clean answers that will be listed among the active posts, and well differentiated from the old post. In addition, the scenarios I described above are unavoidable due to the software's specific nature, so they will naturally occur regardless of us liking it or not.

A solution that I suggest is tolerating Software Recommendations users to adopt the traits of proper maintenance and asking. We could describe these in the "How to ask" section in the help center, as well as the reasoning behind natural software outdating. Additional flagging options and hinting comments in the form of advice from committed users when such a question is encountered is also a good way to go.

You must log in to answer this question.

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