0740-7459/10/$26.00 © 2010 IEEE September/October 2010 IEEE SOFTWARE 61
focus
Dean Wampler and Tony Clark: One criticism of
MPP is that the average development team can’t
handle more than one “thing” at a time. Is that
true? If so, should we give up the idea of MPP?
Or are there mitigation strategies worth pursuing?
As trends seem to be moving in this direction, if
development teams cannot handle MPP, then will
they be relegated to using “second class” develop-
ment processes? Will this put these organizations
at a serious disadvantage?
Neal Ford: Yes! But this question seems to pose
it as a binary proposition. Most teams are by
necessity MPP teams now. No one writes in a
single language anymore. Even trivial applica-
tions have a general-purpose language, SQL, Ja-
vaScript, CSS, and dozens of frameworks, each
of which includes an external DSL (usually in
XML) that is its own mini language (the syntax
is XML, but the XMLSchema defnes the se-
mantics). For a long time, teams that refused to
use SQL (preferring instead some sort of fat fle
“database” tuned to a particular 4GL language)
or that refused to use Ajax would be considered
behind the mainstream. Increasingly, teams that
ignore MPP will fnd themselves falling behind
their competitors.
MPP is really just a realization that we’ve been
U
sing multiparadigm programming (MPP) has costs as well as benefts.
Over email, guest editors Dean Wampler and Tony Clark discussed with
Neal Ford and Brian Goetz the practical issues for MPP in industrial soft-
ware development teams. What follows is a transcript.
Dean Wampler, DRW Holdings
Tony Clark, Middlesex University
Neal Ford, ThoughtWorks
Brian Goetz, Oracle
Multiparadigm
Programming
in Industry:
A Discussion with Neal Ford
and Brian Goetz
multiparadigm programming