Re: Type-free Circles and Ellipses [T]

On Fri, 31 Aug 2001 00:41:07 GMT, topmind_at_technologist.com (Topmind) wrote:

>Is that language documented on-line anywhere by chance?

The report describing the ANSA computational mode is still available here: <http://www.ansa.co.uk/ANSATech/94/Primary/100101.pdf>. The language was really just a vehicle for exploring the model. There is a whole lot more at the site <http://www.ansa.co.uk/> for those interested in object-based distributed computing.

>I see no distinct boundary between "sub-types" and "attributes".

"attributes" is another of those words I try to avoid. I remember many fruitless debates about the difference between attributes and properties with neither term being properly defined. There was also a tendency for those asserting the existence of attributes as something fundamentally different from operations to appeal to some sort of underlying and inexplicable magic to make them work (at least in a distributed system).

>One can be exchanged for the other. Thus, the choices seem
>to be:
>
>1. use only subtypes to represent variation
>
>2. use only attributes to represent variation
>
>3. mixed
>
>
>I don't think anybody supports #1, so I will ignore
>it.

In CORBA IDL, attributes are merely syntactic sugar that saves you having to write out one or two operations explicitly. The attributes are as much part of the type as are the operations.

Since this does not appear to match your concept of "attribute", I do not know what you mean by "attribute" and so cannot comment further.

-- 
Owen Rees - opinions expressed here are mine; for the full disclaimer
visit <http://www.users.waitrose.com/~owenrees/index.html#disclaimer>