---

Hi folks.



Brief Bio

After spending nineteen years in the industry as a software engineer (notably, as a lead developer at the American Board of Family Medicine, a consultant for the American Board of Radiology, and a software architect at Hewlett-Packard), I returned to academia in 2011, earning a BS in CS summa cum laude from the University of Kentucky in 2012.

"I" have been on CNN — in this Anderson Cooper story, while I'm not shown physically, I was responsible for the architecture and much of the implementation of most of the exam and imaging software shown in the segment. And as far as I'm concerned, that counts!

At present, I teach the University of Kentucky's Intro to Software Engineering course — you can find general notes about my teaching here or the course web site here. My students continue to, as they have in the past, kick ass. All assignments in the course are handled through the UKCS Source Control System, which I designed and implemented.

Research Interests

Logic programming, AI, networking, software engineering.

More specifically, my primary interests are computer decision-making (in its various forms) and the design and development of tools to bring AI/logic programming/knowledge representation techniques to a wider audience — with the conga project being the major manifestation of the latter.

---

Typically, the best way to reach me is via email — david.b.brown@uky.edu.

My office phone is (859)257-6745, and while I pick it up when here, do not count on me listening to voice mail; please send email instead.

Formal office hours: M 2:30-4; TR 3-4.

I can generally be found in my office, room 234, in the Hardymon Building. Hardymon is unfortunately located quite a bit away from main campus:


(I won't spend a huge amount of time pretending to feel regret at the location, as the distance from main campus means the parking lot is actually usable...)

I have an open door policy (and, in fact, encourage students to visit), and am regularly available outside of normal office hours. Don't be confused by the closed door; it's typically closed only to protect my colleagues from my poor taste in music. Check the status screen on the door, which, through the magic of the Internet, currently reads:


(Updates every 15 seconds)

---

I currently teach CS 216 - Introduction to Software Engineering. The course is taught in C++, and focuses on object-oriented design and software engineering methodologies applicable to all modern object-oriented languages.

C++ is, fundamentally, a mathematical notation. And like any notation, one expresses ideas with it — and as we move from simply programming to software engineering, we learn that the clarity of expression of those ideas is vital. The first important insight to a software engineer is that code being correct does not guarantee that it is well-engineered. During the course, students learn the difference between the two concepts, techniques that lead to well-engineered code, and how to prove that code is correct in the first place.

The core project for the class is based, loosely, on the classic computer game NetHack.



The game basis of the project brings out creativity in the students, and critically for a course heavy in object-oriented design, allows the definition of a large, understandable, and non-contrived object hierarchy.



Students in the class learn techniques that allow large software projects to work. And, after completing three to nine thousand lines of code for the course's four interconnected programming assignments, develop a sense of scale in software development.

All submissions for the course are handled through the UKCS Source Control System, which I designed and implemented, and is typically used by three classes (each under separate faculty) each semester.

---

Assessing Portal Design Skills in the Radiation Oncology Interactive Case Management Examination
Evanoff, M., Hussey, D., Rovinelli, R., Brown, D.B.; Journal of Digital Imaging, Vol 14, No 2, Suppl 1 (June), 2001: pp 104-107

Temporal Action Language
Li, W., Brown, D.B., Truszczyński, M., Huffman Hayes, J.; University of Kentucky Computer Science Department Technical Report (2012-521-12) (2012)

Answer-Set Programming in Requirements Engineering
Li, W., Brown, D.B., Huffman Hayes, J., Truszczyński, M.; REFSQ 2014

ChoiceNet Gaming: Changing the Gaming Experience with Economics
D. Brown, H. Nasir, C. Carpenter, O. Ascigil, J. Griffioen, and K. Calvert; CGAMES14

Designing A GENI Experimenter Tool To Support The ChoiceNet Internet Architecture
D. Brown, O. Ascigil, H. Nasir, C. Carpenter, J. Griffioen, K. Calvert; CNERT 14

---

Conga — a portmanteu of "convergent grounder" — is a new answer set programming language and grounder. The language itself — heavily influenced by KU Leuven's IDP² exposes answer set programming via first order logic semantics (with extensions to make the actual programming a bit more pleasant, such as a full suite of arithmetic operations, aggregation of atoms, and an extended existential quantifier).

I began working on conga in 2012 under the direction of Mirek Truszczyński, designing and implementing a grounder for the IDP² answer set programming language. In 2013 (inspired by a conversation with Marc Denecker), the project expanded to its current focus — the conga answer set programming language and the conga C++ class library, both designed and implemented by me.

In addition to the logical connectives shown to the right, conga follows C-style arithmetic operators (and thus == for comparison) and follows C's rules for operator precedence.

Thus far, the conga command line tool implements the conga language, as well as IDP² at near 100% functionality. Work is ongoing in implementing the ASP-Core-2 standard.

Conga can be a bit more verbose than other answer set programming grounders, as it requires explicit declaration of variables, but otherwise conga programs tend to be of similar size of programs in other answer set programming languages.

So, without further ado, here's an example conga program which solves a sudoku board:


However, in the grand scheme of things, this isn't all that exciting; I'll be the first to admit that the syntax of IDP³ is more elegant, and the field as a whole appears to be moving away from the concept of new languages (or, at least, new divergent languages) inasmuch as the acceptance of the ASP-Core-2 project is receiving support.

But! This isn't the interesting part of conga. In fact, the inelegance of the syntax of the stand-alone language is a result of the language's main design goal: In addition to existing as a standalone grounder, it also exists as a C++ class library which uses the exact syntax in C++. So, with the proper headers included and libraries linked, we can convert the above logic program into the following C++ function:


Which gives us a perfectly normal C++ function which takes a vector of ints (representing a sudoku board), builds a theory, grounds it, solves it, and returns the results in a standard C++ collection class.

Of particular interest in the above code is the line:


Which demonstrates what is likely the most powerful feature of the conga class library: the ability to embed "live" C++ code into a logic programming rule, and have that code executed during the grounding step. In this example the code performs a lookup in the vector passed to the solvePuzzle function, which then, during grounding, gets executed to evaluate the semantic value of the element.

An end result is the daily puzzle to the right — generated daily by the example source code here.

A simpler example — solving a single puzzle given by the user as opposed to generating a whole puzzle — can be found here.

---

Conga has been tested with and is fully functional compiled with GCC 4.7 or greater and Visual Studio 2012 (with the November 2012 Compiler CTP) or Visual Studio 2013. Other C++ compilers with robust C++11 support should work as well.

A public release is expected in the first quarter of 2014.

---

2014 marks the fourth year that the UKCS Source Control System has been used by students at the university.

As the designer and implementor of the system, I have the advantage of using the system on a daily basis over the past two years of teaching. As a result, the system is functional, easy to use, and stable.


Originally started in 2011 at the request of Jerzy Jaromczyk, the system has now been utilized for four distinct courses and served over a thousand students.

The system manages a SVN source control server, automatically handling student, faculty, and grader accounts and privileges on the source control server, while automating assignments and providing quick access to submitted files for grading.


As the feature set (and the software itself) has stabilized towads the end of 2013, we hope to publically release the system in 2014 pending finalization of admin tools.

---

From 2011-2013, I was involed in the Temporal Action Language project with Wenbin Li, Mirek Truszczyński, and Jane Hayes. Melding my experience with software engineering with my interest in logic programming, I was responsible for the design and development of the translator from TeAL to clingcon.

---

Since 2013, I've been a member of the ChoiceNet team at the UK Laboratory for Advanced Networking.

While the project is still in its early stages, I've been active in developing IPV6 routers supporting ChoiceNet extensions.