[antlr-interest] De-emphasizing tree grammars?
kferrio at gmail.com
Tue Dec 27 11:31:21 PST 2011
Ter has convinced me that I have nothing to fear from v4 and that it
actually will make life easier for both target maintainers and people using
By pseudo-target, I mean an intermediate representation for which no
compiler or VM exists but which is easily machine-translated into any
target language. In the past, it seemed to me (and I may be wrong) that
keeping a target in sync with antlr development (i.e. the prodigious pace
of Ter) could be a full-time job and that this might explain why some
targets appeared to lag. Of course there might be other reasons. Having a
stable pseudo target would enable (among other things omitted here for
brevity) the antlr core and targets to evolve in parallel rather than
sequentially. It's just another decoupling. Again, based on Ter's
remarks, I think v4 will improve the situation without the added complexity.
A software book written by economists? I must see this, if only to learn
whether it is better than an economics book written by software engineers!
On Dec 26, 2011 1:08 PM, "Jason Osgood" <jason at jasonosgood.com> wrote:
> Hi Kyle Ferrio.
> > A well defined Neutral Implementation Pattern Language, viz. a
> > pseudo-target could be a useful compromise with the following benefits:
> > enabling target-agnostic validation; (2) accelerating simultaneous,
> > synchronous development of both antlr and targets (a major problem
> > currently imho) by providing a thin shim targeted by antlr and known in
> > advance by authors of target generators; and (3) promoting reusuability
> > readability of grammars via abstraction of the implementation.
> I don't understand what a pseudo-target is. Do you have an example?
> As for your stated benefits, I agree enthusiastically. Modularity
> through interfaces is power.
> The book "Design Rules", written by two economists, is the most
> important software architecture book I've read to date.
> It sometimes seems that my entire career has been figuring out ways to
> decouple stuff. :)
> Cheers, Jason
> List: http://www.antlr.org/mailman/listinfo/antlr-interest
More information about the antlr-interest