Agreed. There is one difference with json (and yaml) though: it truly is more human writable (and readable), and by extension provides nicer diffs. I'd be much more happy to maintain my resume in json than in XML, even if I might be more happy exchanging/submitting resumes with (one of) the existing XML standards.
But all that's needed is a script that eats XML resumes and outputs json resumes, and vice versa.
Just like I'd rather write markdown and read (rendered) html+css.
That said, on the subject of XML not being "icky": one resource for having fun with XML: "Program Generators with XML and Java" by J. Craig Cleaveland (Prentice-Hall, 2001):
Hm, looks like there might be bit too many dependencies here for my taste:
npm install resume-cli
du -hs node_modules
103M node_modules
All this, and it doesn't provide line numbers when reporting errors due to an extra comma in the json file? Nor does it (yet) have LaTeX output...
Still, can't complain, 5 minutes and I have a resume draft that doesn't look half bad in html, and is presumably entirely reasonable to customize (not to mention it's all inn a json-file anyway, so one could just use what's in ones actual resume, and shove the json through pretty much any scripting language w/templates and generate something reasonable).
Sorry about the large dependency size, we were mainly rapidly prototyping and didn't expect so much usage. That size will drastically be reduced and we hope to have better error reporting.
Currently have 60 fresh issues on Github for the schema so after we digest those we will be improving the CLI once again.
But all that's needed is a script that eats XML resumes and outputs json resumes, and vice versa.
Just like I'd rather write markdown and read (rendered) html+css.
That said, on the subject of XML not being "icky": one resource for having fun with XML: "Program Generators with XML and Java" by J. Craig Cleaveland (Prentice-Hall, 2001):
http://craigc.com/pg/