From lieber@ccs.neu.edu  Thu Mar 15 09:52:25 2001
Return-Path: <lieber@ccs.neu.edu>
Received: from gomeisa.ccs.neu.edu (lieber@gomeisa.ccs.neu.edu [129.10.117.106])
	by amber.ccs.neu.edu (8.10.0.Beta10/8.10.0.Beta10) with ESMTP id f2FEqN915320;
	Thu, 15 Mar 2001 09:52:23 -0500 (EST)
Received: (from lieber@localhost)
	by gomeisa.ccs.neu.edu (8.10.0.Beta10/8.10.0.Beta10) id f2FEqMT28152;
	Thu, 15 Mar 2001 09:52:22 -0500 (EST)
Date: Thu, 15 Mar 2001 09:52:22 -0500 (EST)
From: Karl Lieberherr <lieber@ccs.neu.edu>
Message-Id: <200103151452.f2FEqMT28152@gomeisa.ccs.neu.edu>
To: bwydaegh@info.vub.ac.be
Subject: April 23
Cc: dougo@ccs.neu.edu, johan@ccs.neu.edu, lieber@ccs.neu.edu, wupc@ccs.neu.edu
Status: R
Content-Length: 2961

Hi Bart:

April 23 is good day for your visit.

Please could you give a talk about how you use the AP library
in your work? Please can you send me an abstract and even better,
a draft of a paper or a thesis chapter.

You wrote earlier:

>> http://liinwww.ira.uka.de/~reussner/coconuts/coconuts.html
>>> . You generate adaptors out of the interface information. Currently
>>> I do not.
>>> . I use the interface information to generate adapted components
>>> with restricted functionality (in case not all external services are
>>> available which are required to offer the full functionality).

Yesterday, Jens Palsberg gave a talk related to generating adapters:

Jens Palsberg (Purdue University) will present

Efficient and Flexible Matching of Recursive Types

(Joint work with Tian Zhao; presented at LICS 2000)

Abstract:

Equality and subtyping of recursive types have  been  studied  in
the   1990s   by   Amadio  and  Cardelli;  Kozen,  Palsberg,  and
Schwartzbach;  Brandt  and  Henglein;  and   others.    Potential
applications  include  automatic  generation  of  bridge code for
multi-language  systems  and  type-based  retrieval  of  software
modules  from  libraries.   Auerbach,  Barton,  and  Raghavachari
advocate a highly flexible  combination  of  matching  rules  for
which  there, until now, are no efficient algorithmic techniques.
In this paper, we present an efficient decision procedure  for  a
notion  of  type  equality  that  includes unfolding of recursive
types, and associativity and commutativity of product  types,  as
advocated  by  Auerbach  et al.  For two types of size at most n,
our algorithm decides equality in  O(n^2)  time.   The  algorithm
iteratively  prunes  a  set  of  type  pairs,  and  eventually it
produces a set of pairs of equal types.  In each  iteration,  the
algorithm  exploits  a so-called coherence property of the set of
type pairs produced in the preceding  iteration.   The  algorithm
takes  O(n) iterations each of which takes O(n) time, for a total
of O(n^2) time.

Maybe that is of interest to you.

-- Karl

Hello,

it has been a while that I mailed you. I hope that you had a great start of
the new Millenium. I have been very busy writing my Phd and a couple of
papers. For one of these papers I need to be in Washington D.C. from 17
April 2001 till 20 April 2001 (conference ECBS 2001). That is the reason of
this mail. I have some questions concerning the adaptive library and I also
have a couple of suggestions to improve efficiency (these were a side effect
of the visit of Ralf Reussner in the VUB). If you still reside in Boston I
can easily hire a car and make a short visit to your lab. This would be
after the conference in Washington so on monday 23 April 2001 or later (if
need to be I can also come before the conference). Is this possible at all?
And if so on what day?

Kind Regards and hopefully till soon!
Bart Wydaeghe

E-mail:bwydaegh@info.vub.ac.be
Tel:+(32) 2 6292975