On The Next Move in Programming1
Dave Thomas, Bedarra Research Labs
|
 |
COLUMN
PDF Version |
IN SEARCH OF THE NEW NEW PROGRAMMING THING
We are now emerging from the dark period earlier in the decade and
many in the community are bemoaning the good old days when new languages
appeared with great frequency and new paradigms abounded.
Unfortunately,
we find ourselves in a world which increasingly ignores the experiences
of past successes and failures to claim inventor rights
are the new new thing. Simple techno sound bites are easy to communicate
using the web and academics have become as adept as major marketers
in spin and mindshare capture. Rather than evolving from the best of
what we know to the better, each new thing in our field seems to need
to kill off those that came before it.
A recent PhD student remarked that it was very upsetting to have
to read old papers published in the mid to late 70s that contain immature
versions of their new ideas which they must now read and cite. It sometimes
appears that scholarly work in our field is no longer a badge of honor
but a burden forced on graduate students who have grumpy old supervisors.
We seem to be in denial that our industry is maturing, we are losing
our elders, some of us don’t code any more and many of our new
things are really good ideas from 20 or more years ago. Some good old
ideas reappear because the technology is now available to commercially
exploit them; others become popular because they provide a better approach
to solving a particular problem. Reinvention and improvement of good
old ideas is both scholarly and sound engineering provided the prior
art is cited and explained as opposed to hyped or even patented. This
means we need to look for the important small improvements and not
just the big bang!
We need to find the balance which builds from our best past work
but also doesn’t send a discouraging message that everything
has been done before, or if it was that it was necessarily done correctly!
We need to benefit from the wisdom of our elders without being constrained
by them. We’ve argued for the increased promotion of computational
diversity (http://www.jot.fm/issues/issue_2003_05/column1)
rather than a narrow focus on using the right current industrial platform
or language.
OBJECTS NEXT?
In the OO community there seems to be a concerted search to either
find what is “Beyond Objects?” or to argue that “Objects
Are It!”. To their credit OOPSLA (http://www.oopsla.org/2006/)
and ECOOP (http://www.emn.fr/x-info/ecoop2006/) have been openly searching
for new ideas in their calls for papers and workshops. The OOPSLA Onward!
track deserves special mention in this regard as it provides a platform
for new ideas.
The popular research area of AOP is one new next contender which
has rapidly emerged from our own community and now has its own successful
conference (http://aosd.net/2005/). Recently, there have been papers/proposals
for Feature, Language, Law, DSL/Domain, and Context Oriented Programming
etc. There is also good work in integrated objects and transactions,
unifying objects, info sets and tuples [1]. Closely related to this
is the interesting work on making relationships found in ER and UML
models first class types in a programming language [2]. Continued work
on Actors [3] and more recently Chords [4] seeks to unify concurrency
with object orientation. Object Ownership has always proved a challenging
problem especially when VM implementations refuse to allow objects
to have an associated owner; however, considerable progress has been
made [5, 6]. There is also a resurgence in research on dynamic languages
including prototype based languages and multi-paradigm languages. Finally,
work on using objects to build next generation multi-modal UI toolkits
is very interesting [7].
All of this work seeks to provide tools or language extensions which
allow the OO model to address specific issues and/or broaden the applicability
of OO. While they don’t promise a New Oriented Programming they
provide the critical and difficult research to build better OO languages,
tools and systems. It is clearly a challenge to integrate all of these
powerful concepts into a language/platform which is useable by those
who are not advanced software professionals.
PEOPLE ORIENTED PROGRAMMING
One of the more exciting changes in our field is the rapid and energetic
emergence of Agile Software development initiated by Scrum and XP.
Unlike traditional process artifact centered approaches to development,
Agile emphasizes human and business values. Large, unfortunately unpublished
internal studies have concluded that defects and timely delivery of
software are most high correlated with the individuals and teams of
individuals rather than the programming environments, languages or
platforms.
We have for too long focused on the technical dimension of our field
without considering the importance of the social aspects and especially
the critical need to effectively communicate with our customers and
between ourselves. Even small progress in this area promises huge benefits.
OPEN SOURCE SOFTWARE
Open Source has provided the impetus for small global teams of open
source project developers, some of whom have never met in person, to
collaborate to build software together. Other developers and users
have full view
of the source code they create and in many cases the interactions of
team members. Simple Open Source tools such as Wiki, JUnit and Fit
have a major impact on allowing teams to build better software. Eclipse
has enabled many to contribute to the open source tooling. Recently
MS and Eclipse have initiated efforts to enable open process definitions
which may lead to integration of processes and tools. The availability
of open tooling has allowed customers to innovate in their tools and
processes, often accelerating the transition to things like Agile development.
The large open source code base provides a rich set of research data
for social researchers as well as software engineers who can now finally
look at large programs as they evolve over time. The combination of
sophisticated program analysis, text understanding and simple heuristics
promises new tools for helping us understand how we work.
SOFTWARE WITH POLISH
It is well accepted that it takes time to polish a good essay, research
paper or a program into something that we believe is correct and readable
[8]. Modern languages lack publication syntax and IDEs lack a nice
way to publish/present the code using multiple fonts and indentations2
so that it is readable [9] by others. We need to make the technical
and business case that clean, well-written software is a better investment
and that “send it back, refactor it and make it better” is
a necessity not a luxury.
SAYING MORE WITH LESS – THE NEED FOR A VHSL
Despite the great progress in tools and languages there is still
far too much stuff that gets in the way. The latent and accidental
complexity
of current languages and platforms makes a 4GL of the past look like
a dream. Java and C# are just barely above the machine code of the
VM. VMs have lifted the level of the platform but we are still programming
just above them. There is too much code to be developed and maintained
to keep using low-level languages even if they are supported with fancy
visual tools and generators. Even in old functional programming languages
such as Lisp and APL all developers knew map reduce as a programming
idiom and didn’t need a separate course in design patterns.
Perhaps we should take a tip from our colleagues in hardware design
and think about languages which are simple, expressive and allow
us to describe major applications in a small number of pages of code.
If electrical engineers can move to HDL and VHDL we should be prepared
to move to at least a similar level of abstraction with a VHSL.
We need to move beyond the complexity, limitations and weaknesses
of Smalltalk and Lisp but seek a language with at least as simple
a syntax
and more expressiveness. We have the benefit of considerable research
on advanced type systems which can used to work with the developer
as opposed to against him. Our hardware colleagues can help by
using some of those transistors to support robust and efficient dispatch,
true sandboxing and type and ownership information for every instance.
Small investments in hardware software co-design can have a huge
impact on the robustness and adaptability of software.
Footnote:
1 This column was inspired by
JOT discussions including an article titled “The Next Move In
Programm-ing: A Conversation with Sun's Victoria Livschitz, http://java.sun.com/developer/technicalArticles/Interviews/livschitz_qa.html” which
inspired the title.
2 JavaDoc and Doxygen provide
basic capabilities for publishing.
REFERENCES
[1] Gavin Bierman, Erik Meijer, and Wolfram Schulte, The essence
of data access in Cω ECOOP 2005, http://research.microsoft.com/Users/gmb/Papers/ecoop-corrected.pdf
[2] Gavin Bierman, Alisdair Wren, First-Class relationships in an
objected-oriented language, ECOOP 2005, http://www.cl.cam.ac.uk/~aw345/talks/ecoop05.pdf
[3] Denis Caromel, Ludovic Henrio, Bernard Serpette, Asynchronous
and Deterministic Objects, POPL'04.
[4] Nick Benton, Luca Cardelli, Cedric Fournet, Modern Concurrency
Abstractions for C#, http://research.microsoft.com/Users/luca/Papers/Polyphony%20(TOPLAS).pdf
[5] Chandrasekhar Boyapati, Alexandru Salcianu, William Beebee, Martin
Rinard, PLDI 2003.
[6] David G. Clarke, James Noble, John M. Potter, Simple Ownership
Types for Object Containment, ECOOP 2001.
[7] Eric Lecolinet, http://www.infres.enst.fr/~elc/ and
Stéphane
Huot, Cédric Dumas, http://dastuf.free.fr/projects.html
[8] Richard Gabriel, The Poetry of Programming, http://java.sun.com/features/2002/11/gabriel_qa.html
[9] Ronald M. Baecker, Human Factors and Typography for More Readable
Programs, Addison-Wesley, January 1, 1990
About the author
|
|
Dave Thomas is cofounder/chairman
of Bedarra Research Labs (www.bedarra.com), www.Online-Learning.com
and the Open Augment Consortium (www.openaugment.org) and a founding
director of the Agile Alliance (www.agilealliance.com). He is
an adjunct research professor at Carleton University, Canada
and the University of Queensland, Australia. Dave is the founder
and past CEO of Object Technology International (www.oti.com)
creator of the Eclipse IDE Platform, IBM VisualAge for Smalltalk,
for Java, and MicroEdition for embedded systems. Contact him
at dave@bedarra.com or www.davethomas.net.
|
Cite this column as follows: Dave Thomas, “On the next move
in Programming”, in Journal of Object Technology, vol. 5, no.
2, March-April 2006, pp. 7-11 http://www.jot.fm/issues/issue_2006_03/column1
|
