260
|
1 |
|
|
2 |
|
|
3 |
Dear Professor ???
|
|
4 |
|
|
5 |
I have recently been appointed as a University Lecturer
|
|
6 |
at King's College London. I know Chunhan Wu from my earlier
|
|
7 |
work with Professor Xingyuan Zhang on regular languages and
|
|
8 |
the Isabelle theorem prover. This work led to a highly
|
|
9 |
regarded publication at an international conference and
|
|
10 |
we also submitted a journal version of this work. In
|
|
11 |
both publications, Chunhan did much of the work and
|
|
12 |
has been the first author of the papers.
|
|
13 |
|
261
|
14 |
I like to invite Chunhan to work with me in London for 11 months.
|
|
15 |
His living expenses for the stay will be covered by
|
260
|
16 |
my grant at King's College London. He would work with me
|
|
17 |
on improving the reasoning in the Isabelle theorem prover.
|
|
18 |
I can teach him the internals of this theorem prover so
|
|
19 |
that he can independently implement his own work. To do this
|
|
20 |
effectively it would be most convenient for us if Chunhan
|
|
21 |
can be at King's College London. I can also introduce him
|
|
22 |
there to the research community in the field of automated
|
|
23 |
reasoning. I have worked and got to know
|
|
24 |
well researchers in Cambridge and in Munich, where the
|
|
25 |
theorem prover Isabelle is mainly developed and which
|
|
26 |
are centres of excellence in the field of automated
|
261
|
27 |
reasoning. During these 11 months, Chunhan and I would
|
260
|
28 |
work on projects which we hope can be published at
|
|
29 |
top-tier conferences.
|
|
30 |
|
261
|
31 |
I kindly request that you grant 11 months leave to
|
260
|
32 |
Chunhan. I am sure this scientific visit will be
|
|
33 |
invaluable for his career. He is a bright researcher
|
|
34 |
who will benefit greatly from collaboration with
|
261
|
35 |
established researchers in the field of automated reasoning.
|
260
|
36 |
In due course when all bureaucratic hurdles are taken
|
|
37 |
the Head of the Informatics Department of King's College
|
|
38 |
will also write to you supporting the request to
|
261
|
39 |
invite Chunhan to work with me in London.
|
260
|
40 |
|
|
41 |
|
|
42 |
Yours sincerely,
|
|
43 |
Dr Christian Urban
|
|
44 |
|
|
45 |
|
|
46 |
PS: Please find attach an outline of the working plan
|
|
47 |
between Chunhan Wu and me.
|
|
48 |
|
|
49 |
|
|
50 |
Work Plan
|
|
51 |
=========
|
|
52 |
|
|
53 |
- We first want to investigate how the datatype package
|
|
54 |
of the Isabelle theorem prover can be made more expressive
|
|
55 |
and more efficient. Chunhan in his current work already
|
|
56 |
went to the limits of the capabilities of this package.
|
|
57 |
I have some preliminary results how this state of affairs
|
|
58 |
can be improved and I would work with Chunhan to implement
|
|
59 |
these results. This work is on the cutting edge of research
|
|
60 |
in this field and the hope is it will be published in
|
|
61 |
top-tier conferences and journals.
|
|
62 |
|
|
63 |
- The second project is about making the Isabelle theorem
|
|
64 |
prover more like a programming environment: programmers should
|
|
65 |
be able to write their code inside a theorem prover, verify
|
|
66 |
properties about the code and then compile it to efficient
|
|
67 |
machine code preserving the guaranties obtained by the verification.
|
|
68 |
These systems are desirable because they can make it cost-effective
|
|
69 |
to verify the functional behaviour of software and allow errors to
|
|
70 |
be detected early during program development. However, to
|
|
71 |
complete this project we will have to make changes to the
|
|
72 |
HOL-logic underlying the Isabelle theorem provers. This
|
|
73 |
is a long-term project which will stretch beyond the time
|
|
74 |
Chunhan can stay in London. But the hope is that preliminary
|
|
75 |
results can already be submitted for publication.
|
|
76 |
|
|
77 |
- Finally, I like to continue with Chunhan on our work
|
|
78 |
about formalising regular expressions and the Myhill-Nerode theorem.
|
|
79 |
There are still a number of question open, for example how this
|
|
80 |
work can be extended to parsing algorithms.
|
|
81 |
|