Re: Consensus Call on Adoption of POSH

Am 30.01.2014 16:28, schrieb Matt Miller:
> [ speaking as a co-chair of the JSON Working Group ]
> 
> The sole item for the JSON WG -- an update to RFC4627 -- is currently
> in the RFC editor's queue.  The WG discussed the topic of comments a
> number of times, but there was no consensus to add comments to the
> JSON format at the time.
> 
> The JSON WG is currently discussing what to do next.
> 
> [ speaking personally ]
> 
> I don't think the lack of a specific syntax for comments is all that
> harmful.  Plenty of systems use JSON today for configuration, and seem
> to be getting by fine; *some* of them use a special syntax, but many
> (from the discussions in the jSON WG, it might even be a majority)
> make use of special name/value pairs within the objects.

When you use it for configuration, you can just use the comment type the
parser of your choice uses. You rely on that parser anyway and will
unlikely switch it. And that workaround to use a special key with the
comment as value is ugly and doesn't give you the possibility to quickly
or temporary disable stuff, nor does it give you the possibility to
comment whole blocks.

Nothing of the above is a solution for standards. They can not rely on
the parser and using a special key as comment doesn't make sense.

And I'm unable to see why they have a problem in defining a character
and it's successors until newline as whitespace. Maybe they can't decide
on the character. They even don't have to enforce it, all what is
necessary is that they standardize one and the same character for the
optional comment-functionality so that all comment-allowing parsers do
use the same character. Something which should have been done right from
the beginning.

Anyway, I assume I'm now really offtopic. Sorry for that, but for me
missing comment functionality is a show blocker for many things and it
smells like going back to stone age.

Regards,

Alexander Holler