Summary of Lisp meeting Dec. 11th in Berlin
Andreas Hinze
here a short summary of our LISP meeting yesterday in Berlin
(I tried to translate my german summary. Please apologise possible odds)
James Anderson, Benno Biewer, Niels Gösche, Joerg-Cyril Hoehle,
Sven Köhler and i met at 5pm.
James had prepared two short talks about XML and a JAVA->LISP translator.
We start with a discussion about the advantages and disadvantages of XML.
James shows a nice example which uses XML to allow independent development
of content and look of a HTML page. He had organized a G4 powerbook and so
we are able to look how it works in practice.
The LISP & XML theme poped up again over the whole evening. This was obvious
since James wrote CL-XML, a XML parser that simplyfies the handling of XML remarkable.
Our second major topic was a JAVA->Lisp translator developed by James (a nice
point: both tools are based on an ATN parser originally developed by Benno).
James demonstrated the goods and odds of automatic translation with a short
example. It shows that while the translation itself works remarkable good, the
result still needs some reworks due to semantic differences and (obviously) missing
Java compatible libraries.
The rest of the evening was spend to a lot of small themes like "Does feature X
from MCL is also available in LWW" over "What are you doing with LISP" up to
"In Smalltalk it is possible to ...". In summary all themes that can better be
discussed in a bar than in front of a computer monitor.
Overall, it was a great evening and we will shure meet again.
Special thanks to James who spend a lot of energy to give two excellent
presentations of excellent LISP tools.
Regards
AHz
Benno Biewer
news:3FD9E22E...@smi.de...
It is nice to see that James could use my concepts and program codes for the
generation of atn parsers on the basis of extended BNF grammars so much
(CL-XML, xpath, xquery, Java-Lisp Translator, ...).
The history has several remarkable turns:
(1) Ca. 1986 i'd developed and used ATN compilers for my natural-language
interface to spss (statistical package for the social sciences).
(2) More than 10 years later, i had to develop parsers for a medical
laboratory equipment. The interfaces were described virtually as BNF
grammars. Therefore i've written a parser generator which translates
extended BNF grammars in ATNs. The ATN formalism serves both as means and as
target: at first the BNF grammar is parsed with an ATN parser (means), an
object model of the BNF grammar is then build (BNF-OM), which is transformed
into an ATN object model (ATN-OM, means), finally the source code of an ATN
parser (target) is derived from the ATN-OM. (Till now, i've used Common
Lisp, Java and Dylan as target languages.)
(3) It's amusing that the initial and recent use of the parser generator
follows in a reversed way of a similar motivation: i.e. to be able to
program in Lisp instead of programming in Java: Initially the parser
generator served me to produce Java parsers without having to write Java
code by hand. I preferred to write a parser generator in Common Lisp which
generates Java ATN parser from BNF grammars. Now, James uses the (modified)
parser generator, to translate java programs into Common Lisp, so it's quiet
possible to use them without writing java code.
Benno Biewer