>The right language is the one chosen by the team at the time.
This is, unfortunately, a truism. It also does not reflect the activities
within the LInux OS environment, where, I believe, pretty much everything
is in C/C++. Languages are ultimately used to communicate with
people. Though programming languages are used to communicate with
machines, their far more important use is to communicate with other
programmers. While it is possible to create "interoperability" layers to
permit machines speaking different languages to communicate with each
other, that is *not* the challenge. The problem (in medical software) is
to get *programmers* to communicate with each other and with domain experts.
Torvald got a group of programmers WHO ALSO HAPPENED TO BE THE DOMAIN
(OPERATING SYSTEM) EXPERTS to communicate with each other using a common
language. That is the entirety of open source.
>If the resources (i.e. people) for a given OS project choose C++, then
>good for
>them - it is no doubt the choice which will result in the best product from
>that team. Another team will choose Java, and others VB, and so on. There will
>always be diversity.
With each new directive from the French authorities banning a certain
percentage of English language films, English words, etc. this position is
given the lie. What's more, while it is rather difficult to make a running
application multi-lingual, it is *already the case* that most programmers
are multi-lingual. Why create interoperability layers when most
programmers can adopt to just about any language? Incidentally, I am not
condemning CORBA or CORBAmed here. It's just that my attitude toward CORBA
is that it is simply another initiative in the Java pantheon. It's Java
disguised as interoperability which is fine with me.
>If they want to make their products usable beyond their own language
>community,
>they don't need to change language, they need to engineer interoperability, by
>providing interfaces to other components in the environment. I.e., they
>need to
>contemplate COM, .net, CORBA, SOAP/XML, etc etc.
As I say, let's come back on March 20, 2002 and see where we are. I
predict that either .NET will carry the day (note please that apparently MS
agrees with me) or Java will. Probably the latter.
John
