[Grace-core] Early experience with traits not overriding superobject

Andrew P Black black at cs.pdx.edu
Thu Mar 10 14:05:54 PST 2016


> On 9 Mar 2016, at 23:15 , James Noble <kjx at ecs.vuw.ac.nz> wrote:
> 
> Right. This will only work if traits don't get the default methods - if the overall model discourages trait conflicts rather than encourages and mitigates them.  

Apparently, my post wasn’t clear.  The problem has nothing to do with traits getting default methods.  It arises because classes inherit default methods.

Yes, if we get rid of default methods altogether, then this problem won’t be there — we will instead have another problem.   Experience has shown that giving all objects minimal behavior is a good idea.

> part of me wants to say: if it hurts, perhaps don't do it that way.
> If you only used traits (Fortress style, right?)
> then you wouldn't have the "override for free" option:
> ideally there would only be one trait defining each method in the composition.

I did only use traits.  That’s the problem.

	Andrew

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailhost.cecs.pdx.edu/pipermail/grace-core/attachments/20160310/ba4a3dd7/attachment.html>


More information about the Grace-core mailing list