On Nov 23, 2009, at 1:22 PM, Jason Grout wrote: > Florent Hivert wrote: >>>>> Of course, when writing code, we have no idea what date the next >>>>> release >>>>> would be, and sometimes are mistaken about the release that the >>>>> patch >>>>> will be included in as well. How should we take care of this >>>>> patch-updating problem? >>>> I've been lazy to raise this problem on the lists. Sorry for that. >>>> >>>> Is this really a problem a warning tells that a function is >>>> deprecated since >>>> version 3.1 whereas it was really deprecated in version 3.2 ? If >>>> yes a >>>> possible solution is the following: >>> It's deprecated as soon as a decision is made. >> >> So should I understand that you even don't want to mention any sage >> version in >> the deprecation warning. IE you'd rather have >> >> >> sage: Partition([2,1]).boxes() >> DeprecationWarning: (Since 2009-09-10) The function bar is removed. > > > From the point of view of the user, having a date that is not the date > of a sage release seems a bit insulting; "you mean to tell me that two > weeks ago, you deprecated some functionality I use, but didn't tell me > until the next Sage release?". I think I'd rather have the > deprecation > based on Sage version number (and corresponding date). I don't know > of > any other software that backdates deprecations to the time the patch > was > merged. > > In fact, most software that I've seen dates deprecation warnings > *only* > by version, and leaves the dates up to the user.
Looking at both options, I agree--there's no need to include the date-- just the version number is sufficient. What might be more useful to a user is a link to the ticket/email thread where the rational was explained, alternatives presented, and decision made. When we go to remove deprecated functions, whatever script that detects that can have the logic that translates version numbers to dates/ages. - Robert -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
