Path: utzoo!attcan!uunet!mcvax!kth!enea!sommar
From: sommar@enea.se (Erland Sommarskog)
Newsgroups: comp.lang.eiffel
Subject: Re: Eiffel on VAX/VMS
Message-ID: <4342@enea.se>
Date: 25 Feb 89 22:47:12 GMT
Organization: ENEA DATA AB, Sweden
Lines: 29

Bertrand Meyer (bertrand@hub.ucsb.edu) writes:
>The VMS version uses the same architecture as the Unix version, and
>still relies on C as intermediate language. No specific support is
>offered for the VMS debugger since we feel it is better to have specific
>tools working in terms of the Eiffel source.

I can understand why ISE used the same technique on VMS as on 
Unix - there is a lot of work to be saved that way. This is not
an entirely bad thing - it has given us an Eiffel on implementation
earlier.

However, as an every-day VMS user, I like to point out that a
system that uses C as an intermediate language is not so attractive
on VMS as on Unix. Unix does always come with a C compiler, VMS
does not. The machine code, on the other hand, is always available
and in difference to Unix always the same.
  And then I guess there are performance gains as well...

As for debugger support, it is of course not possible to provide
support for the VMS debugger with the C solution. The question 
does not arise until someone implements a direct-to-machine-code
system. If someone is DEC, the answer is simple. For ISE or any
other I like to point out that if you are an experienced VMS
user, you probably prefer if you can use the debugger you are used
too. Particulary if you mix Eiffel with other langauges.
-- 
Erland Sommarskog
ENEA Data, Stockholm              This signature is not to be quoted.
sommar@enea.se