[xquery-talk] [xml-dev] Query 3.1 vs. JSONiq WAS Re: MarkLogic using JSONiq for processing JSON ?

Adam Retter adam.retter at googlemail.com
Sat May 9 16:42:09 PDT 2015

> Botton line is: as a result of what W3C did with XQuery 3.1, they created
> more harm them good overall for the industry.
> And this:  for both the XML community  AND  the JSON community.
> For the XML community: they’ll be hated and avoided even more they used to
> be, and more and more isolated, and

I don't understand your perspective at all.
I don't believe that XQuery is perfect, but then I don't believe that
any other programming or query language is either. Significantly
however we do have real XQuery 3.1 users (that were previously using
XQuery 3.0 and XQuery 1.0) publicly thanking us for the new features
of XQuery 3.1 that they are enjoying, here is one such recent thank
you - http://exist.markmail.org/thread/mb7jdspx5h3d67kj

> For the JSON community:  they’ll avoid anything related to XQuery like scary
> evil, which means that they’ll design silly query languages
> by themselves (see Cassandra, see Mongo, see BigTable….)  for 15 years
> before finding some decent solution.

JSON is important sure, but I don't believe it is the beginning and
the end of the Web and/or NoSQL. You mention Cassandra, but their
query language CQL appears to me to be inspired by SQL rather than
anything like JSONiq.

I really like JSONiq, I even started an implementation (unfinished) a
few years back. However, I have no sympathy for people or communities
that want to ignore a technology base because it is `scary evil`, I
don't buy into that as an argument, it just sounds like FUD; Serious
implementers of any language will always do their homework and learn
about the best and worst of their predecessors.

Regards Mongo, the only JSONiq implementation for that seems to be
from 28msec which you were heavily involved in I believe. Outside of
28msec and their partner work (IBM), apart from Xidel, I have not seen
any implementations of JSONiq. Certainly the NoSQL databases that you
mention, don't require a W3C stamped query language for them to
produce an implementation. I would be genuinely interested to know why
JSONiq was not more widely adopted? I really believed that JSONiq
would be snapped up very quickly by NoSQL JSON/BSON stores, Node.js
and others.

I think that if people want just a JavaScript query language for JSON
then why don't they just get/create an implementation of JSONiq in
JavaScript? Sure it could have been XQuery 3.1, but it's not... and
well... I think that is okay. XQuery 3.1 has its own use-cases and
purpose, it might not be as popular as JSON, but I don't see that as
an issue, they solve different (and sometimes similar) problems.

Adam Retter

skype: adam.retter
tweet: adamretter

More information about the talk mailing list