|
You seem to be splitting historical hairs to avoid a conclusion you never wanted to reach: that SML was a good idea, motivated by a mismatch between what programmers wanted to do with XML and what XML was offering.
It's true that JSON isn't SML, and that YAML didn't feed directly into JSON. Nonetheless, YAML's emergence from SML and JSON's parallel development both seem to be recognition that maybe markup (ala SGML and XML) isn't the best way to share data between programs, and that other structures are often a better idea.
I've long wanted programmers to work with something that fits their needs better than XML, as the efforts to make XML address data transfer have mostly created an incredible bloating. I had hopes for ASN.1 at various times, but JSON and YAML (for heavier-duty work) seem like the right answers to that set of problems in a way that XML, even XML augmented with zany Post-Schema-Validation-Infosets, can never be.
|