[antlr-interest] Re: Software Transformation Systems Workshop

matthew ford Matthew.Ford at forward.com.au
Fri Jun 11 18:59:42 PDT 2004


Hi Monty,

I have written an Algorithmic Differentiation program for GAUSS
(www.aptech.com  see below).
It uses Antlr to parse a GAUSS procedure and then three tree walkers to
translate and optimize the
Algorithmic derivative procedure.

Due to the limitations of AST tree creation syntax (see my previous posts on
a suggested revised syntax),
 most of the tree building (and rebuilding) is done by hand coded
proceedures, having used
the tree walker to identify a particular type of branch.

The parser is a little unusual because GAUSS changes to being white space
significant inside matrix
indexing expressions, [ .. ].

I have also used Antlr to build a translator between a user friendly
language and a time dependent (non-SQL) data base language.
(http://www.beaconit.com.au/Beacon_timeframe.htm).

As for other methods, I have built a number of translators by hand
previously (in C), for example
SimGauss uses a hand build translator to turn pseudo GAUSS code into
derivative procedures for the
integration procedures.
(http://www.forward.com.au/products/simgauss/index.htm and
http://www.aptech.com/AS_prodMF.html  see below)

If I was doing this again now, I would used Antlr because it is faster to
develope with and easier to maintain.

All my Antlr work is done in Java.
matthew

AD 1.0 (Algorithmic Derivatives)

The GAUSS AD 1.0 module is an application program for generating GAUSS
procedures for computing algorithmic derivatives. A major achievement of AD
is improved accuracy for optimization. Numerical derivatives invariably
produce a loss of precision. The loss of precision is greater for standard
errors than it is for estimates. At the default tolerance, Constrained
Maximum Likelihood (CML) and Maximum Likelihood (Maxlik) can be expected
generally to have four or five places of accuracy, whereas standard errors
will have about two places. Accuracy essentially doubles with AD.

For some types of optimization problems, convergence is accelerated.
Iterations are faster and fewer of them are needed to achieve convergence.
The types of problems that will see the most improvement are those with a
large amount of computation.


SimGauss v2.1
Nonlinear simulation
Forward Computing and Control Pty. Ltd.


A fully interactive nonlinear simulation module written in GAUSS, SimGauss
provides a fast and easy way to simulate nonlinear differential equations
and state-space systems, such as vehicle dynamics, biological systems and
economic models. The module features extensive user control. GAUSS's
Publication Quality Graphics provide exceptional ways to visualize your
results. Comprehensive documentation and on-line help complete the package.



The model simulation code is written in GAUSS. You can use GAUSS' high level
mathematical functions such as probability density functions, FFTs, matrix
inverse, eigenvalue/eigenvector and SVD functions to quickly simulate
complex models and control algorithms.




> --- In antlr-interest at yahoogroups.com, Monty Zukowski <monty at c...> wrote:
> > Hi folks,
> >
> > I'm writing a position paper for this conference about real-world
> > translation projects.  I get lots of inquiries about doing such
> > projects, so I've learned a bit about what people think they can and
> > can't do, etc.  But I could use more tales too, so---
> >
> > Anyone do a real translator with ANTLR out there?
> >
> > Anyone consider it but decided against it?

> >
> > For instance, Larry White recently posted asking about doing C++ to
> > Java translation.  In the end he and his comrades decided that the easy
> > stuff to translate was trivial, the hard stuff was too hard to
> > automate, and they decided to do it by hand.
> >
> > Thanks for any responses!
> >
> > Monty Zukowski
> >
> > ANTLR & Java Consultant -- http://www.codetransform.com
> > ANSI C/GCC transformation toolkit -- 
> > http://www.codetransform.com/gcc.html
> > Embrace the Decay -- http://www.codetransform.com/EmbraceDecay.html
> >
> >
> > On May 26, 2004, at 10:14 AM, Terence Parr wrote:
> >
> > > Hi Gang,
> > >
> > > Co-located with OOPSLA this year (October in Vancouver I believe), is
> > > the  Generative Programming and Component Engineering (GPCE'04)
> > > conference.  Within that is a workshop:  Software Transformation
> > > Systems Workshop:
> > >
> > > http://www.program-transformation.org/Gpce/STS/
> > >
> > >  From the website:
> > >
> > > "The workshop will have a small number of participants, around 20,
> > > selected on the basis of short position papers submitted to the
> > > organisers. The aim is to let people with different perspectives meet
> > > in order to allow fruitful interaction. The workshop will start with
> > > short presentations of the background of the participants. Then
> pointed
> > > presentations of the various generative techniques, followed up with
> > > discussions on how software transformation systems may aid in
> > > supporting the techniques. If the discussions converge, a small
> > > collection of joint papers may be written as a result of the
> workshop."
> > >
> > > I will be there for sure.
> > >
> > > Also, don't forget we'll be settling on an ANTLR workshop date
> (here in
> > > San Francisco) fairly soon.
> > >
> > > Ter
> > > --
> > > Professor Comp. Sci., University of San Francisco
> > > Creator, ANTLR Parser Generator, http://www.antlr.org
> > > Cofounder, http://www.jguru.com
> > > Cofounder, http://www.knowspam.net enjoy email again!
> > > Cofounder, http://www.peerscope.com pure link sharing
> > >



 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
     http://groups.yahoo.com/group/antlr-interest/

<*> To unsubscribe from this group, send an email to:
     antlr-interest-unsubscribe at yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
     http://docs.yahoo.com/info/terms/
 



More information about the antlr-interest mailing list