Re: Brain Dump of Towards FIT 1.0 OpenSpace session

----- Original Message ----- 
From: "Ward Cunningham" <[email protected]>
To: "FIT Developers" <[email protected]>
Sent: Monday, August 25, 2003 11:03 AM
Subject: Re: [Fit-dev] Brain Dump of Towards FIT 1.0 OpenSpace session


> 
> On Friday, August 22, 2003, at 07:13  AM, [email protected] wrote:
> 
> > at one point I thought of pushing type conversion to the fixtures
> > themselves; the fixtures, after all, know what's needed better than 
> > some
> > global framework. It's still an idea that has a bit of an appeal.
> 
> Type conversion is the responsibility of the fixtures.

OK. What I had in mind, though, was quite different from the
current mechanism. All of the methods in the fixture
would take a single string argument (that is, whatever was in the
cell, without any conversion), and return a
result object. There wouldn't be any distinctions between
getters and setters, or any type determination at all on the
level of the Fixture or the supplied subclasses. 

The result object would have attributes that distinguished
between a setter and a getter (that is, the fixture knows,
and needs to return that information so the result can be
colored appropriately), and provide a result that
distinguished between OK, Failed (with the actual result,)
and Error (with an error message or a traceback.)

John Roth

> 
> --
> Ward Cunningham