Re: CACM article (Feb 2009): "Compiler research: the next 50 years"

On Thu, 12 Feb 2009 09:02:06 -0800 (PST), Jason Evans
<joevans@gmail.com> wrote:


>On Feb 11, 6:46 am, Max Hailperin <m...@gustavus.edu> wrote:
>> > [For those of us whose ACM memberships expired a decade ago, what else
>> > does it say? -John]
>>
>> For those of you, I would point out that you should really take a look
>> at the revitalized CACM and decide whether it merits reactivating your
>> membership. The CACM has not been so good since the 1970s, and
>> depending on your perspective, perhaps not even then.
>
>As far as the compiler research article is concerned, I have mixed
>feelings about its message. The basic tenet is that we should be
>working together on large systems, and stop wasting so much time
>implementing the same infrastructure over and over. The problem with
>that is similar to the problem with object-oriented programming and
>class reuse: the existing components are rarely a perfect fit for the
>project at hand. In each case, the cost/benefit of retrofitting has
>to be weighed against the cost/benefit of reimplementation.


More to the point, the article implies that developing and maintaining
academic compiler projects is a waste of time and effort. The authors
suggest that researchers should not develop their own compilers but
rather should team with established open source efforts.


Specifically, it says:


"The compiler community must change its current research model, which
emphasizes small-scale individual investigator activities on one-off
infrastructures. Complete compiler infrastructures are just too
complex to develop and maintain in the academic research environment.
However, integration of new compiler research into established
infrastructures is required to ensure the migration of research into
practice. This conundrum can be solved only through a new partnership
between academia and industry to produce shared open source
infrastructure ..."




Apart from this, the article says nothing particularly new. The other
points they raise we've all heard before.


George

Return to the comp.compilers page.
Search the comp.compilers archives again.