Lutz Prechelt wrote:
> Dear Denis,
>
> great message, thank you very much!
>
> Indeed: Rules that are not being followed
> are worse than no rules at all, because they
> undermine the value of the remaining rules
> (as the result in people getting used to
> not following rules).
>
> What to do in the specific cases you mention?
> I suggest to simply drop the first two of the
> hardly-followed rules:
> - listeners should always reside in an "events"
> subpackage
> - Empty abstract classes for listeners should be
> called FooAdapter
> because their added value is low.
Agreed.
> In contrast, the tags, in particular those related
> to thread-handling, are valuable help for clarifying
> what one is doing (when writing new code)
> or for understanding what is there (when modifying
> code), so I propose to make them a constant area
> of attention in the code reviews and gradually
> add them into existing code as well.
The question is who does the work to apply the tags we choose consistently
throughout the codebase. Like with the conventions in general, having the tags
and used partially diminishes their usefulness, because their absence carries
no useful information ("doesn't have this property" and "noone didn't apply any
tag here yet" are equally likely). So if we have no one to do the work,
dropping the tags that are not or almost not used (and maybe adding a sentence
to the method or class summary instead) might be less damaging.
Regards,
Denis
------------------------------------------------------------------------------
_______________________________________________
DPP-Devel mailing list
DPP-Devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dpp-devel
