Re: getCapabilities() as both static and object method?

On 5/16/14 1:40 AM, Harald Alvestrand wrote:
> On 05/16/2014 06:21 AM, Jan-Ivar Bruaroey wrote:
>> On 5/15/14 9:39 AM, Harald Alvestrand wrote:
>>> I ignored the UA's possible prompting in the descriptions; the
>>> "Sorry" messages would come from the error handlers in the
>>> Javascript, handling the failure callback from getUserMedia.
>> Your example is a strawman.
> Please, let's talk about the arguments instead of trying to make
> accusations about their quality. This was my attempt to distil down into
> an example the high points of a discussion with a couple of my
> development team; I may have done a bad job, but that doesn't mean the
> concerns aren't real.

Sorry, I didn't mean to offend, but I stand by my criticism of the 
example. I apologize if 'strawman' is the wrong word, as I see some 
definitions online imply intent by the user, which is not what I meant 
(I think strawmen often appear on their own will, unintentionally. I've 
certainly conjured strawmen without realizing it, many times).

What I meant to criticize is that the example sets up a separate problem 
as a representation of the actual problem, and that the representational 
problem happens to be easier to deduce to a particular conclusion. I 
don't think it follows that the original problem benefits equally or (If 
someone could pm me with a better word for that I would appreciate it). 
Because:

And here is my rationale. By mixing UI with API, we get this: By mapping 
the API (which is a programmer's interface) directly to a UI (which is 
usually a user's interface), the burden of having software repeatedly 
make calls is equated with the burden of a user repeatedly having to 
click buttons. The burden is not the same because software is written 
once - and executes really really fast thereafter - and the users' 
burden is repeated for every interaction, a repetitious UI, which is 
bad. In contrast, a may people prefer APIs that are imperative.

>> Mandatory constraints let software probe without bothering the user
>> until there's something to bother the user about. If there's a
>> difference here, it's not this.
>>
>> Why did we invent a whole language to solve this prescriptively -
>> "Here's what I want" - to just ask "what do you have?"
> That argument doesn't contain enough detail for me to evaluate it.

Sorry, I can elaborate more (see my pm), but can you help show me the 
benefit your team is seeking from this new feature first, either with a 
different example or different words?

.: Jan-Ivar :.

Received on Friday, 16 May 2014 14:07:26 UTC