Re: application facing lifecycle, data and interaction model for things

> On 28 Apr 2016, at 22:45, Carsten Bormann <cabo@tzi.org> wrote:
> 
> I'm going to stop here, but you may get what I'm after here:
> A design from scratch is going to repeat all the mistakes that have been
> made in distributed systems and programming environments and then some.
> I'd rather start with the Web and evolve from there.

On that last point, we agree - the Web of Things is based upon the fundamentals of the Web architecture:

   a.  URIs for addressing resources
   b.  declarative formats for describing resources
   c.  protocols for accessing resources

See: http://www.w3.org/standards/webarch/

For (b) we want to know that when you and I talk about X we are referring to the same idea/thing. Given that different platforms can be expected to have different formats, we need a common way to relate the meanings. This is where the Resource Description Framework (RDF) is a good solution.  Declarative formats with links also allow for compilation of indexes as a basis for search.  For (c) the requirements vary considerably across domains, so no single protocol will be appropriate in all cases. This means that (b) must contain sufficient descriptive power to enable platforms to know how to interoperate.

The Web simplifies application development by providing a standard interface independent of the platform. Web page scripts work the same whether you have a Mac, a Windows PC, a Linux box or so forth.   The same is true for the POSIX APIs for IP, enabling application developers to write networked services independently of the underlying network technology.  We need the analog of the POSIX APIs for the Web of Things.  The event driven programming paradigm has been very successful and is very familiar to Web developers. It also lends itself to a wide range of programming languages, e.g. JavaScript, C#, state machines and so forth.

The idea is thus to define a common framework for specifying the application facing interface. Different platforms could use different formats for this framework, although it would be preferable for us to converge on a small number of formats. The type system is key to the framework and needs to be flexible enough to address the needs of a broad range of application domains. We are not starting from scratch and can build upon existing work, e.g. the schema.org core types.

Best regards,
—
   Dave Raggett <dsr@w3.org <mailto:dsr@w3.org>>