Re: Unifying testsuite policy and getting rid of CSS exceptions

On 14-Sep-17 14:23, Geoffrey Sneddon wrote:
> There had been some talk about dropping the requirement for <link
> rel=help>, but this requires versioned folders (as you need to require
> somewhere has the version) for the build system's metadata
> requirements, and it would rely on folders within the spec mapping to
> anchors within the spec (and anchors aren't stable across versions,
> because specs get refactored).
>
> Does anyone have any preferences?

Yes. I am strongly against removing direct spec links and replacing them 
with broader, directory based inferred links to an entire section of the 
spec.

This removes a lot of value from the test suite, when using it to meet 
CR exit criteria or checking for spec coverage. It also makes it harder 
to review tests or track down why a test is failing.

If a test author doesn't have the spec open and isn't closely reading 
the spec while writing the test, the test is likely garbage. Thus, the 
incremental cost for adding rel=help spec links is close to zero.

-- 
Chris Lilley
@svgeesus
Technical Director @ W3C
W3C Strategy Team, Core Web Design
W3C Architecture & Technology Team, Core Web & Media