W3C home > Mailing lists > Public > public-media-fragment@w3.org > May 2010

Re: Test Cases: moving forward

From: Silvia Pfeiffer <silviapfeiffer1@gmail.com>
Date: Wed, 19 May 2010 19:27:17 +1000
Message-ID: <AANLkTimxyo_q00WPHKxYMtic3ZurF6_15dJIbf7IvaBl@mail.gmail.com>
To: raphael.troncy@eurecom.fr
Cc: Media Fragment <public-media-fragment@w3.org>
Hi Raphael,

2010/5/19 RaphaŽl Troncy <raphael.troncy@cwi.nl>:
> Hi Silvia, all,
>
> I think we have now a very solid base with the test cases resolution
> described at
> http://www.w3.org/2008/WebVideo/Fragments/wiki/TemporalDimension.
>
> I have some minor comments:
> †- For #02: originally, we had two test cases to test: #t=0,0 and #t=a,a
> where a>0. Silvia has merged them into a single one, which makes sense but I
> wonder if we should not split them again in order to have explicit test
> cases for the boundary case!

a=0 is not a boundary case; a=s is a boundary case. And no, I don't
think it's necessary - it doesn't change any of the activities.


> †- For #10: this can happen only if s > 0 and the entire fragment requested
> is before the start position of the media. We could add this in comment ?

Isn't that obvious?

> †- For #12: you don't want to say in comment that this is equivalent to
> #t=s,e ?

It's there in writing for the special case. I'd rather say that t=a,
is equal to t=a,e - which is what I wrote.

> †- For #11, #12, #14 and #16: I suggest we force a 206 answer, with the
> rationale that a fragment has been requested (that might be equal to the
> entire resource but this is similar to case #01!).

It's not really up to us, but it's up to the UA. As the UA will have
to retrieve the full resource, the question is whether it wants to
retrieve the resource in byte ranges or in one go. I have formulated
this better in the spec actually.


> In the sub-section "Temporal dimension syntax", should we have more test
> cases when the unit is %-encoded? Currently, we have only one for the 'n' of
> 'npt', but what about the 's' for 'smpte', etc. ? This could lead to a very
> big number of ugly test cases :-(

I think they are all the same type and therefore don't need to be
listed explicitly.


> What to do next? Should we have someone (Jack?) to review them ? Should
> we put the test cases in the spec or in a separate document ?

Jack reviewing the spec/wiki page would be good.

> The Spatial dimension is easy, there will be no test cases, no request is
> issued.
> What about the Trach dimension test cases ?

Don't know.

Cheers,
Silvia.
Received on Wednesday, 19 May 2010 09:28:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 21 September 2011 12:13:38 GMT