Re: [PR Media Fragment] specify how further keywords would be added

On Mon, Apr 23, 2012 at 11:53 AM, Liam R E Quin <liam@w3.org> wrote:
> On Mon, 2012-04-23 at 08:05 +1000, Silvia Pfeiffer wrote:
>
> [...]
>> Both of Liam's proposals add a new parameter,
> [...]
>>
>> If xywh was used together with xySwh to achieve some backwards
>> compatibility,
> It was not clear to me how to do that. What would the syntax look like
> exactly?
>
> I was assuming people would use xySwh _instead_ of xywh when they wanted
> to add an "S" parameter to an isotropic/linear transform.

As a browser/server you have to be able to deal with all
possibilities. If you wanted your URL to just work in old browsers,
too, that would not be supporting xySwh yet, you'd probably add both
parameter, since then at least you get the fragment even without the
scaling. There's of course no problem if you just use a new parameter,
since then you get no action in an old browser and the desired
fragment plus scaling in the new one.


> Note, I'm not asking or a normative change in the spec (not likely given
> its status) but a note to explain how one adds a parameter in such a
> case.
> The example in the spec,
> http://www.example.com/example.ogv#track=audio&t=10,20
> does help, but only when the parameters are orthogonal.

Is http://www.w3.org/TR/media-frags/#general-structure not sufficient?

It might be worth adding a recommendation that new parameters should
be orthogonal to existing ones. OTOH they don't have to - it just
depends on what you're trying to achieve. Both your proposed new
parameters were valid in that regard.

> By the way, the text [[
> "&" is the only primary separator for name-value pairs, but some
> server-side languages also treat ";" as a separator.
> ]]
> suggests to me that implementations are encouraged to accept ; as a
> secondary separator. This is a good thing for the same reason that
> using ; between URI query parameters is better than using & in some
> cases.

Yeah, that will all depend on what browsers (and servers) will
implement, I guess.


> If I want to scale an image by 50% then xywh has a problem, I need to
> use xy, then scale, then use wh. But, maybe it's enough for now.
>
> Thank you for your responses; my concerns are satisfied enough that I do
> not have any objection to the spec moving forward, thank you.

Thanks for the review!

Cheers,
Silvia.

Received on Monday, 23 April 2012 04:15:35 UTC