Re: Ichibah flames, and flames out over, Ada 9X

[asuvax!ennews!enuxhb.eas.asu.edu!koehnema@g.ms.uky.edu (Harry Koehnemann]

In article <1993Feb17.065421.12021@sei.cmu.edu> wellerd@ajpo.sei.cmu.edu (David
 Weller) writes:
>Speaking of tagged types,
>it was my impression that this was Mr. Ichbiah's most sour grape.
>He was rather insistent on creating a class-based language, rather
>than the tagged types that we have now.  Although I agreed with him
>on quite a few points, Tucker's counterpoints were far more
>compelling.  

Now that's interesting, if it is indeed true.  Technical merits are
not the sole consideration for any design.  WHile the package/type
extension approach may have advantages over a class based approach,
it also has disadvantages - perhaps the largest being understanding
by developers.  Class based langauges are common, their usage well
understood (well, as understood as things get in CS), and most
importantly, are currently embrassed by OO developers.  Type
extensions (tagged types), while by no means a new concept, are not
seen in the popular OO langauges.

I think it will be interesting to see how the OO community that Ada
seeks to crack will embrace extensions as a construct for inheritance.
Particularly considering their past acceptance of the construct.  I
also think the Ada community has a big sell ahead of itself.  Not that
it can't be done, but "technical merit" does not directly lead to
"popular" or "used" (ask FORTRAN and COBOL - and unfortunately Ada83).
--
Harry Koehnemann			Arizona State University
koehnema@enuxha.eas.asu.edu		Computer Science Department