Hi Jim:

That is going to be an excellent use of RIDL.
I would like to use your project as a case study for how to
employ RIDL.

You may want to take a look at Crista's thesis for futher background
and experiences of other RIDL users.

ftp://ftp.ccs.neu.edu/pub/people/lieber/theses/lopes/

>I wonder if it is possible to instantiate the objects directly on the
>remote systems, taking a remote reference to them during the parse of
>the input. That would be something to look at too.

I am not sure about this. Johan please can you answer?

-- Karl


>From jtsilla@ccs.neu.edu Fri Nov  6 13:05:17 1998
>CC: johan@ccs.neu.edu
>
>
>Dear Prof. Lieberher,
>
>I read the RIDL document you refered to and I am very interested.
>I will try a few things with it to learn it. Iwas hoping to be able to
>design and build a sort of distributed data base for my project. This
>database will have some class definition like the library system example
>but will distribute the information about the books and users among
>several host.
>
>I will design a traversal to distribute the infromation based on hints
>in the class definition. A new field called "site" will indicate which
>host to distribute the object to.
>
>I will then design some traversals (similar to the example) which
>operate over this distributed object sctructure.
>
>I wonder if it is possible to instantiate the objects directly on the
>remote systems, taking a remote reference to them during the parse of
>the input. That would be something to look at too.
>
>My project will start off with a simple non-distrbuted part. This will
>then be made distributed with a RIDL aspect. I will then extend the
>class and add behaviors while keeping the RIDL aspect fixed. I will then
>add more distribution (user infromation will be distributed now, for
>instance) while keeping all other parts mostly-fixed. Output logs on the
>remote systems will trace events as the program operates.
>
>Does this sound like a reasonable project?
>
>reagrds,
>-Jim.
>