Re: [xml-dev] Towards XML 2.0

[Gavin Thomas Nicol <gtn@rbii.com>]

> JSON is actually richer in its infoset: it has numbers, arrays, structs, strings, boolean, etc.   It does not have simplicity in its infoset, it has convenience and familiarity to programmers: alignment with JavaScript and C-ish languages.
> 
> So saying "lets remove some syntax, because that will make us simpler and more like JSON" doesn't accord with reality: people could equally be saying "lets add some syntax, because that will make us richer and more like JSON."  Seeing nothing but syntactic "simplicity" in JSON seems deeply superficial to me IYSWIM.

This is actually an important point, and can be leveled at wiki markup and other things too: they replace something claiming they're 'less complex' but many times they end up being at least as, if not more complex, in many cases more constrained (mixed content), and often less regular syntactically. JSON is great for slinging around to build simple services, depending on your needs, so is hessian, but you wouldn't write a book using it.

Convenience is an important design point, but convenience and simplicity in one domain does not guarantee applicability to other domains.