W3C home > Mailing lists > Public > public-mwts@w3.org > April 2009

Re: http://www.w3.org/2005/MWI/Tests/track/actions/23

From: Dominique Hazael-Massieux <dom@w3.org>
Date: Tue, 28 Apr 2009 10:06:11 +0200
To: Kai Hendry <hendry@iki.fi>
Cc: public-mwts <public-mwts@w3.org>
Message-Id: <1240905971.20865.443.camel@localhost>
sorry it took me so long to reply... Per my ACTION-30:

Le mardi 31 mars 2009 à 11:52 +0100, Kai Hendry a écrit :
> There is quite a few bits I didn't like. For example:
> Since new standardised Web features could be implemented by plugins, I
> thought it was totally unnecessary to say tests should avoid using
> them. For example SVG implementations often rely on a plugin. Perhaps
> I am little biased as I do work for company who hopes to compete by
> implementing standardised device APIs (W3C widget spec) via a plugin.
> http://webvm.net

While I can see why you'd think plugin usage shouldn't be discouraged,
in practice, if you ask your testers to install a plugin, fonts, etc.,
you're likely to not have testers, or not have reliable test results.

I don't think we should stay "you must not require plugins", but calling
attentions of testers on that topic sounds rather important to me.

> Keyboard and pointing devices can be collapsed into one section.

Works for me.

> Prerequisites - I hate this word. :-) But really, why does one have
> explicitly state the multitude of features a test might require or
> rather depend (slightly better word) on. Perhaps in a meta tag
> http://wiki.csswg.org/test/css2.1/format#requirement-flags , but not
> as explicit text.

I agree with Wilhelm's point, here:
> This is the one type of meta data a tester would be interested in 
> seeing when running through a set of tests. If there's a bunch of tests
> testing Geolocation, and you either don't have this feature or have it
> disabled, you'd want to know. They would be a different class of
> failures.

That said, maybe it doesn't need to be in the test itself, but in some
documentation that is part of the testing process?

> I really didn't like the "Target devices" section. At first you say
> that it's impossible to account for all possible constraints, and then
> the next steps seems to be about assessing which technologies are
> widely deployed?? That doesn't make sense to me.

I'm not sure what is the contradiction you're seeing?

>  A tester should not
> have to care about such decisions! Testers should be writing tests
> against standards to verify conformance.

In an ideal world, possibly, but if you're writing tests for mobile
devices in practice, you have to take some of these into account...

> I didn't understand "Take care when triggering DOM operations that
> they will not require downloading DTDs". I don't think it's important
> for testers to include doctypes.

Yeah, that one can probably be dumped.

Received on Tuesday, 28 April 2009 08:06:38 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:08:03 UTC