[bitc-dev] Explaining inheritance vs. type classes

[Jonathan S. Shapiro]

On Tue, Mar 23, 2010 at 11:10 AM, Rick Richardson <rick.richardson at gmail.com
> wrote:

> ...to translate this into the C++ type system. I would leave (Bruno et al)
> classes out of the resulting object code entirely. They would be used only
> as a compile time construct to enforce interfaces.  Objects, then would
> correspond to structs and classes fairly naturally.


I think that we are talking past each other once again.

Setting aside my continuing distaste for their sealed classes, the
Oliveira/Sulzmann scheme has some advantages from the BitC perspective.
Mainly, it cleans up some of the mess introduced by DEFOBJECT by effectively
merging DEFOBJECT and DEFINSTANCE. This much seems good.

But that still leaves us me with the desire to deal with the following
issues:

  - Type (single) inheritance through C++/Java/C#-style subclassing
  - Virtual methods

I do not claim that either feature is particularly desirable, but I am
concerned that they are a practical necessity for successful interoperation.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.coyotos.org/pipermail/bitc-dev/attachments/20100323/25a0f4d2/attachment.html