5

When I search for queries on data.SE, I often get a large number of results with the same query name, which are, I assume, forks and variations of the same query.

I believe something needs to be done to help users distinguish between these. Now, it's true that there are view (or execution) counts and star markers, but they're actually not very useful, for at least two reasons:

  • Often the execution counts are on the same order of magnitude for a large number of variants, with little or no starring.
  • The queries might differ in some aspect of functionality such that both aspects may be interesting, i.e. there's no inherent well-ordering of quality, either subjectively or objectively.
  • If someone recently improved an old popular query (say by avoiding some corner case which should not be counted), his/her query will have much lower execution counts despite being better. In fact, this situation is likely to persist, since people will naturally be attracted to the popular query, which also comes up first when searching.

Now, one could well consider a much more involved mechanism for representing forks/variants and their relations - but that would require lots of coding. I suggest that, at the least, people creating a query will be able to add some text explaining how theirs differs from an existing query (as it existed at a specific time); and that this be visible somehow in the query's page (with a link to the original?), and perhaps as a tooltip on the search page.

I'm not saying that's the best thing to do, but it's some thing, and like I said, I believe something should be done about this.

  • They kind of could do already add some text explaining how theirs differs from an existing query, in the description field, right? – rene Feb 4 '17 at 12:31
  • 2
    Agreed & noted – Tim Stone Feb 4 '17 at 14:27

You must log in to answer this question.

Browse other questions tagged .