On Wed, Mar 20, 2024, at 5:43 PM, Jaap van Otterdijk wrote:
> I do not think we should revisit the context of psr-5 and 19. Phpdoc
> as a standard still makes a lot of sense. Maybe an extension could be
> to allow attributes, but that would enforce us to rethink the existing
> standard, where psr-5 is just about defining a real standard for a
> defacto standard that has been around for more than 20 years now.
>
> But I'm open to talking about this topic.
>
> Regards
> Jaap
Clarification: PSR-5/19 were proposed as a way to standardize and update the old phpdoc convention, which *currently has no official standard*. That effort has stalled numerous times over the last decade, for various reasons.
The landscape has also changed in that time, such that attributes can now replace virtually all library-specific doc tags, and the language's ability to represent types now covers basically everything except array shapes and generics. That makes the relevant scope for PSR-5 much smaller, but also means it should ask questions like "should there be a FIG standard for defining generics types, or is the current status of PHPStan and Psalm having slightly different versions acceptable?" (I am not answering that question now, just noting it as the sort of thing the WG should discuss.)
--Larry Garfield