Re: ISSUE-14: Gathering requirements [System Info & Events]

Hi David,

On Oct 14, 2009, at 13:52 , David Rogers wrote:
> Your comments about prioritisation conflict with the discussion on the
> 16th of September about prioritisation of APIs[1].

Could you clarify how? Because I really don't see where the conflict  
is. That the WG has priorities (so that people who are interested in  
working on any aspect can work together) is one thing, that we get in  
the way of people who are interested in focussing only on a specific  
part is another — and I really fail to see the value in blocking them.

> Let's approach this in an orderly fashion otherwise this group will  
> not
> get anything done.

Could you please clarify what you find disorderly about the group's  
progress? The only thing I see keeping us from getting anything done  
at this point is the volume of meta-discussion versus actual  
proposals. I am personally considering placing a two-week ban on all  
meta-discussion to see how things turn out, and that's not something  
that I consider lightly given that I've never had to reach for it in  
the toolbox in any previous group.

> We have a clear list of inputs, so get those
> delivered as per the preferences stated in [1] and concentrate on the
> nice-to-haves as a secondary objective.

I wholeheartedly and strongly agree. Let's drop the process and  
patterns discussions and talk about technical specifications.

> We also need guidance from legal about IPR.

Legal can only provide guidance on things that are relatively  
concrete. Let's wait until there's at the very least the beginning of  
something that they can look at.

> I do not believe we have consensus on a way forward for sensors.

That is because I don't believe that we need to discuss finding  
consensus on how to reach consensus. People have taken actions to  
produce proposals about how to move forward, about how to define  
something universal or something simple that works for v1. Once we  
have these, we can talk. We can point at problems. We can find counter- 
examples. We can raise issues. Before then, it's very obvious from the  
various threads that people aren't on the same page and could talk  
forever about whether this or that is a good idea or not.

> Clearly, there is a lot of thinking to be done rather than just piling
> into designing an API that could well be redundant within a year  
> because
> it addresses a use-case which is too narrow. Sensors is a huge  
> subject.

That is certainly true — hence the preliminary work that some kind  
participants have committed to. There are many potential bad outcomes;  
another is spending three years working on a "perfect" API that covers  
all cases but that is completely overtaken by the industry and too  
complex to use (e.g. XML Schema). Working from concrete proposals  
rather than on process has the massive advantage that it means we can  
get feedback from users and implementers, as opposed to just from  
standards people. The former are always more important.

--
Robin Berjon
   robineko — setting new standards
   http://robineko.com/

Received on Wednesday, 14 October 2009 13:25:03 UTC