+1 I can't agree more with this:
"Perhaps the world would be a better place if standards wouldn't be about "do as you wish and pray that the global network doesn't get even more fractured because of that", but they would be _standards_ which are actually explicitly enforced. While it's great to have liberal views, when it comes to a standard, I want conformance to be a question of Yes or No, not "well, sort of"." Example of a standard in disarray: SDP, for the stated reason. -- Bala -----Original Message----- From: standards-boun...@xmpp.org [mailto:standards-boun...@xmpp.org] On Behalf Of Adam Nemeth Sent: Friday, December 09, 2011 1:37 PM To: XMPP Standards Subject: Re: [Standards] Schemas vs Test Suites Hello, Before the anarchy ensues, I'd like to propose the following plan: 1) define every XEP in a suite of (unit) tests - unit tests should be written in a way that equals to formal checking(!!!) - the test framework can be anything, but it's likely that a custom test language would be needed - optional parts should be defined as "config" variables for XEP test suites on a per XEP basis (as, MAY (default false), SHOULD (default true), etc) - whatever passes the test is said to be conform to the given XEP 2) Find sponsors to finance building this testsuite and find test developers who will actually do it - vendor-independency should be ensured - fosdem summit is a perfect time to actually start buildi the test suite - it's a lot of work, people won't volunteer to do it properly enough (as it never happened here) 3) Ask major XSF-related services to collect statistics on which XEPs are used most of the time - Ask major server vendors to build this possibility in a way that doesn't disrupt network traffic and is privacy-compliant - Ask service providers to enable this on their most used services - Implement these XEP test suites first Perhaps the world would be a better place if standards wouldn't be about "do as you wish and pray that the global network doesn't get even more fractured because of that", but they would be _standards_ which are actually explicitly enforced. While it's great to have liberal views, when it comes to a standard, I want conformance to be a question of Yes or No, not "well, sort of". Udv: Aadaam On 2011.12.09., at 19:13, Kevin Smith <ke...@kismith.co.uk> wrote: > On Fri, Dec 9, 2011 at 6:05 PM, Bala Pitchandi <b...@vidyo.com> wrote: >> I strongly disagree that XSD (RelaxNG or any other way of specifying the >> structure of the XML messages) is a "waste of time". If you don't have a >> schema, you'll either end up implementing all the validation manually or >> worse crash on invalid input. XSD parser-validators (they exist) do all that >> work for you, and are optimized to do that. > > While this may be true, the schemas are not normative, and so can't be > reliably used in this way. > >> Without a schema, all we have is the text and examples which could be >> interpreted in many ways by how they are written and how they are understood >> by the reader. > > Yes - but these are what's normative, not the schemas. > >> Requiring a schema also forces the XEP authors to think hard and come up >> with a design that's structured, extensible (Yes, XSD schemas can be made >> extensible). > > I'll ponder this argument. I'm not entirely convinced one way or the > other at the moment. > > /K