[Python-ideas] Adding jsonschema to the standard library

[Andrew Barnert]

On May 22, 2015, at 09:39, Demian Brecht <demianbrecht at gmail.com> wrote:

> In my mind, the value of bundling anything nowadays really boils down to “this is the suggested de facto standard of solving problem [X] using Python”.

The other way of saying that is to say it explicitly in the stdlib docs, usage docs, and/or tutorial and link to the package. While that used to be pretty rare, that's changed recently. Off the top of my head, there are links to setuptools, requests, nose, py.test, Pillow, PyObjC, py2app, PyWin32, WConio, Console, UniCurses, Urwid, the major alternative GUI frameworks, Twisted, and pexpect.

So, if you wrote something to put in the json module docs, the input/output section of the tutorial, or a howto explaining that if you want structured and validated JSON the usual standard is JSON Schema and the jsonschema library can do it for you in Python, that would get most of the same benefits as adding jsonschema to the stdlib without most of the costs.

> I see two problems with relying on pip and PyPI as an alternative to bundling:

In general, there's a potentially much bigger reason: some projects can't use arbitrary third-party projects without a costly vetting process, or need to work on machines that don't have Internet access or don't have a way to install user site-packages or virtualenvs, etc. Fortunately, those kinds of problems aren't likely to come up for the kinds of projects that need JSON Schema (e.g., Internet servers, client frameworks that are themselves installed via pip, client apps that are distributed by bundling with cx_Freeze/py2app/etc.).

> 1. PyPI is filled with multiple solutions to the same problem. This can be difficult to wade through for the experienced developer, never mind the novice.

Usually this is a strength, not a weakness. Until one project really is good enough to become the de facto standard, you wouldn't want to limit the competition, right? The problem traditionally has been that once something _does_ reach that point, there's no way to make that clear--but now that the stdlib docs link to outside projects, there's a solution.

> 2. You generally won't know about packages that don’t solve problems you’ve solved or are solving. Early on in my adoption of Python, there were a number of times where I just spent time digging through the standard library and was surprised by the offerings that I didn’t even know were a thing. Likewise with jsonschema, I wouldn’t have known it was a thing had a co-worker not introduced me to it a couple years ago.