@adrian I was just checking your fm3.mse file... first of all; all files should end in a newline so maybe you want your endDocument to print a newline at the end...
Then I see that for opposites you do not specify derived anymore either? Thats already good. The container property you have, that is for identifying back links right? If so... aren't all One's of a One->Many relationship back links? Does anybody know of a metamodel where this is not the case? If not ... maybe we could throw that out and infer this too?
cheers
On 14 Dec 2007, at 14:04 , Toon Verwaest wrote:
@adrian I was just checking your fm3.mse file... first of all; all files should end in a newline so maybe you want your endDocument to print a newline at the end...
Whitespace is to be ignored.
Then I see that for opposites you do not specify derived anymore either? Thats already good. The container property you have, that is for identifying back links right? If so... aren't all One's of a One->Many relationship back links? Does anybody know of a metamodel where this is not the case? If not ... maybe we could throw that out and infer this too?
There are many, see our discussion
cheers _______________________________________________ Moose-dev mailing list Moose-dev@iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev
Adrian Kuhn wrote:
On 14 Dec 2007, at 14:04 , Toon Verwaest wrote:
@adrian I was just checking your fm3.mse file... first of all; all files should end in a newline so maybe you want your endDocument to print a newline at the end...
Whitespace is to be ignored.
Then I see that for opposites you do not specify derived anymore either? Thats already good. The container property you have, that is for identifying back links right? If so... aren't all One's of a One->Many relationship back links? Does anybody know of a metamodel where this is not the case? If not ... maybe we could throw that out and infer this too?
There are many, see our discussion
Not convinced... as is obvious.
If you really only want 1 tree... maybe you can remove container and replace it by:
(FM3.Class (name "File")
(thatsuperspecialtypeIwanttobetheonlytreetobeusedbyprogramsthatonlywantonetree (ref-to: Folder))..)
which automatically finds which of the One-to-Many attributes links to Folders.
Oh; sure; Whitespace should be ignored... my program doesn't complain; it's just good practice on most operating systems to have files newline-terminated.
Not convinced... as is obvious.
If you really only want 1 tree... maybe you can remove container and replace it by:
(FM3.Class (name "File")
(thatsuperspecialtypeIwanttobetheonlytreetobeusedbyprogramsthatonlywantonetree (ref-to: Folder))..)
which automatically finds which of the One-to-Many attributes links to Folders.
And obviously everybody is still open to give me that one killer example of which I will say; ok... -that's- why we need it. I haven't heard that killer example yet for which I can't give a reply why I would not want that. Holy beliefs are not reasons.
Ok;
I have found one myself which has nothing to do with the things Adrian told me...
Many->Many relations where the -> is modeled as a firstclass entity are Many->One, One-> Many not contained in neither the left nor the right many.
example: Invocations are firstclass relations in Famix. The sender nor the receiver are container, while it would be so if I would derive container in the way I proposed.
-
Adrian Kuhn -
Toon Verwaest