I have a not totally useless Perl implementation of on CPAN and GitHub now.

It is passing these 6 data driven roundtripping tests, which should be easily grokkable:

Here is the (reasonably small amount of) source code:

It has a dump() and a load() function. dump() takes a Perl object, makes a full clone in the JSYNC/JSON style, and JSON encodes it. load() takes a JSYNC string, JSON decodes it, and makes a full clone to a native Perl object.

The full clone in both directions seems a bit heavy handed, but I think it is too soon to make premature optimizations. I want to get a common test suite passing in several languages, and then think about optimization refactoring.

The tests point out that it might be a good time to revisit the default tags for objects. Having the term 'perl' in there will surely not pass on Python, etc. The Python defaults are also not portable.

Enjoy, Ingy