There are a lot of very minimal things that we can rewrite with the same effect >99% of the time, two of which come to mind are

  • Non-standard syntaxes for Identity Columns for databases that support them (like PostgreSQL which has serial that predates Identity Column syntax)
  • Things like LIMIT x [OFFSET y] which are now standardized to FETCH { FIRST | NEXT } [ count ] { ROW | ROWS } ONLY

I'm just wondering if we shouldn't be more aggressive about putting the best foot forward and showing people the standard-sytnax rather than having a corpus of things that represent syntax only there because it predates the standard.

Should we perhaps as a policy matter give editors free reign to upgrade to the most-standard-compliant syntax so long as there is no adverse effect in doing so?

  • 2
    My short answer: No, not yet. Maybe in 5-10 years. Jun 2 '18 at 20:48
  • 8
    I think the big problem you'll face is that an answer that was written for a version that is 10 years old can still service people using that version. If you go and have free reign to edit the question to update it to today's syntax, you not only change the context of the question, but you also make the answer unusable to people it used to be usable for. If you feel strongly that people on newer versions should be using newer syntax, add a new answer, but don't disrupt the existing answers because they may still be used for future readers who can't or won't upgrade. Jun 2 '18 at 22:29

I think it would be acceptable to add a footnote to an existing answer to note the availability of updated syntax, but please be sensitive about it, and preserve the content that is applicable to users of older versions.

If it cannot be done minimally, consider adding a separate answer instead.

Also be prepared to accept (with good grace) a rollback by the original author, or anyone else that objects to the edit.


You must log in to answer this question.

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