Second Announcement
VSTTE Competition 2013
20-22 April 2013 (NOTE: The first announcement posted on 4 April had the wrong
dates!)
Organizers: Joseph Kiniry, Hannes Mehnert, Dan Zimmerman
This edition of the VSTTE programming contest is an experiment of a different
kind, as it is more about software
engineering than programming. It is not a contest to see who can write and
verify small problems as quickly as
possible, but instead how can a team create a quality piece of code, using any
tools and techniques (not just
verification), in a short period of time.
Quality software is about more than just verified data types and algorithms at
the source code level. Unlike
previous competitions, this year's VSComp will focus on a rigorously
engineered software system. Contestants will be evaluated for all of the
software engineering artifacts that they
produce, not just for verifying their implementations.
Consequently, teams that competed in previous competitions are encouraged to
recruit new team members whose skills
complement those of the existing team members. For example, perhaps the
current team is great at low-level design
and verification, but is weak in writing
requirements or in rigorous validation/testing.
The aims of the competition are:
• to bring together those interested in rigorous software engineering and
formal verification, and to provide an
engaging, hands-on, and fun opportunity for competition and mutual-learning,
• to evaluate the usability of a variety of software engineering tools, not
the least of which are logic-based
program verification tools, in a controlled experiment that could be easily
repeated by others.
After the initial announcement of the contest we were asked the following
question by more than one party. You’ll
find it in our FAQ.
Q. Why attach this style of contest to VSTTE, if we are not focusing
exclusively on verification?
A. Not a single international programming contest rewards good software
engineering behavior. We have tried for
years to influence the big contests and venues (e.g., ACM, ICFP, and TopCoder)
to pay more attention to
engineering and quality, but to little effect. If we are going to see a
contest like this survive and even thrive
in the long run, it seems it has to be an outgrowth of the verification
community, rather than the
implementation-centric hack-fast community.
The contest takes place over a two-day period. The system that contestants
must develop is secret until the
moment the contest starts. The system will be decomposed for the contestants
into an architecture, whose
constituent pieces are the sub-problems of the contest. Thus, by solving all
sub-problems, one writes the entire
application. What's more, the architecture is specified in such a way that
independent solutions to sub-problems
submitted by competing teams should compose into the final system.
The kinds of software engineering concepts mentioned in the contest include:
requirements, domain analysis,
design, architecture, formal specifications, implementation, validation,
verification, and traceability. A
well-prepared team will have a methodology prepared for each of these facets.
The submission of a solution for a
sub-problem need not include any of these facets in particular---i.e., running,
verified code is neither necessary
nor sufficient to win the
contest.
There are no restrictions on concepts, tools, and technologies used. Teams
whose focus in on "early" (i.e.,
requirements or domain analysis) or "late" (validation/testing or evolution)
phases of the software engineering
process are very welcome. There is no limit on team size, but the results will
be normalized by team size.
We particularly encourage participation of:
• student teams (this includes PhD students),
• non-developer teams using a tool someone else developed, and
• several teams using the same tool
A panel of judges will evaluate contest entries to independently score
sub-problems and determine the winner.
Solutions will be judged for correctness, completeness and elegance. The
total score for a sub-problem is the
sum of its scores in the following categories, where the total number of points
in each category available is
indicated in parentheses: domain analysis (3), requirements (3), architecture
(3), design (6), implementation (6),
validation (6), formal verification (12), traceability (3). The maximum number
of points available for each
sub-problem is 42. The verification researcher will note the weight given to
formal verification.
All submitted artifacts will be made public immediately after the contest ends
so that contestants can comment
upon each other's submissions. We expect that a paper will be co-authored by
all interested contestants about the
contest's results, as in several previous contests.
The contest begins at 9:00 GMT on Sat 20 April and ends at 9:00 GMT on Mon 22
April.
Prizes will be awarded in the following categories:
• best team
• best student team
• tool used most effectively by the most teams
The contests website is http://www.vscomp.org/. You will find there
momentarily an outline of the contest, a some
Frequently Asked Questions. The contest problems will go live at this site at
9:00 GMT on Saturday the 20th of
April. (Take note of your daylight savings time GMT offset!) There is no need
to pre-register for the contest,
but you are welcome to warn us that you'll be competing.
Questions or comments about the contest should be sent to Joe Kiniry
([email protected]).
------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
hol-info mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/hol-info
