- From: James Graham <jgraham@opera.com>
- Date: Tue, 23 Oct 2012 16:15:36 +0200
- To: public-html-testsuite@w3.org
On 10/23/2012 02:56 PM, Philippe Le Hegaret wrote: > On Tue, 2012-10-23 at 12:40 +0200, James Graham wrote: >> On 10/23/2012 01:56 AM, Philippe Le Hegaret wrote: >>> As part of trying to get an idea of where we are, I generated a view of >>> where we have tests or not (attached). It's only partial and I'll try to >>> improve it by next week. Only covers the HTML5 spec for the moment. >> >> Nice start :) >> >> It would be interesting to hear what your methodology is. Does this >> cover approved tests only, or both submitted and approved? > > It's a partial covering of the submitted ones. I'm assuming that all > approved tests are also in the submission directory so I'm not looking > at those for this report. I see. >> Is it >> automated or does it it involve lots of manual work still? > > All automated so far. It's a bunch a quick and dirty Java code. I can > put the code in the repository if it is of interest. Yeah, I think that would be interesting, and make it easier for others to pick up this idea in the future. >> I suggest we cross-reference this with data about what is implemented in >> browsers; the areas with the biggest delta would be obvious targets for >> short-term testing effort. If you have some code that generates this >> output, I could try adding caniuse.com data. > > An other way is to generate this apart. As long as the data is > associated with ids in the documents, we can mash them up later. Of course.
Received on Tuesday, 23 October 2012 14:16:17 UTC