Re: observation framework page updated


As you can see on the pages, we decided to use tasks rather than
giving the abilty to choose several observers among a big list. It
looks handier, easier to use and more general than only giving a list
of possible observations (the tasks can describe each observation
separately). We think that this task list is rather powerful,
providing the ability to include tasks in other ones, and an
"advanced" parameters mapping that should be enough flexible to allow
strange mappings.

We added an information in the mapping tags: an inputmethod attribute
that can contain "file", "direct" or "uri" (or a combination of them)
to allow specific mappings for each input method.
We also changed a constraint for the contract : there can be up to
three 'method' tags (in the WADL file), each of them describing one
input method. Combined to this is the RDF file which contains
information on which WADL method represents each input method.
These changes are not on the "unicorn" pages, but I think this is
important to mention them.

Finally, in the grammars section of the contract files, we decided to
allow only complex types containing a single choice element, as you
can see in the sample file
This might change in the future, but we think this is enough to
describe parameters.


On 6/10/06, olivier Thereaux <> wrote:
> Hello,
> The "unicorn" observation framework pages have been updated this week.
> I'll let Jean-Gui and Damien give more details if necessary, but in a
> nutshell, the main update has documentation on the "contract", the
> "tasks" and observation output, with schemas and samples. These will
> probably be refined and improved as the project moves to
> implementation phase, but your early feedback would certainly be useful.
> Thanks,
> --
> olivier

Received on Monday, 12 June 2006 09:05:48 UTC