|         |      1 <?xml version="1.0" encoding="utf-8"?> | 
|         |      2 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> | 
|         |      3 <HEAD> | 
|         |      4 <TITLE>2018/19 BSc Projects</TITLE> | 
|         |      5 <BASE HREF="https://nms.kcl.ac.uk/christian.urban/"> | 
|         |      6 <script type="text/javascript" src="striper.js"></script> | 
|         |      7 <link rel="stylesheet" href="nominal.css"> | 
|         |      8 <script type="text/javascript" src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/MathJax.js?config=TeX-MML-AM_CHTML"> | 
|         |      9 </script> | 
|         |     10 </HEAD> | 
|         |     11 <BODY TEXT="#000000"  | 
|         |     12       BGCOLOR="#4169E1"  | 
|         |     13       LINK="#0000EF"  | 
|         |     14       VLINK="#51188E"  | 
|         |     15       ALINK="#FF0000" | 
|         |     16       ONLOAD="striper('ul','striped','li','first,second')"> | 
|         |     17  | 
|         |     18  | 
|         |     19  | 
|         |     20 <TABLE WIDTH="100%"  | 
|         |     21        BGCOLOR="#4169E1"  | 
|         |     22        BORDER="0"    | 
|         |     23        FRAME="border"   | 
|         |     24        CELLPADDING="10"      | 
|         |     25        CELLSPACING="2" | 
|         |     26        RULES="all"> | 
|         |     27  | 
|         |     28 <TR> | 
|         |     29 <TD BGCOLOR="#FFFFFF"  | 
|         |     30     WIDTH="75%"  | 
|         |     31     VALIGN="TOP"> | 
|         |     32  | 
|         |     33 <H2>BSc Projects</H2> | 
|         |     34 <H4>Supervisor: Christian Urban</H4>  | 
|         |     35 <H4>Email: christian dot urban at kcl dot ac dot uk,  Office: Bush House N7.07</H4> | 
|         |     36  | 
|         |     37 <H4>Note that besides being a lecturer at the theoretical end of Computer Science, I am also a passionate | 
|         |     38     <A HREF="http://en.wikipedia.org/wiki/Hacker_(programmer_subculture)">hacker</A> … | 
|         |     39     defined as “a person who enjoys exploring the details of programmable systems and  | 
|         |     40     stretching their capabilities, as opposed to most users, who prefer to learn only the minimum  | 
|         |     41     necessary.” I am always happy to supervise like-minded students. | 
|         |     42 </H4> | 
|         |     43  | 
|         |     44 <H4>In 2013/14, I was nominated by the students | 
|         |     45     for the best BSc project supervisor and best MSc project supervisor awards in the NMS | 
|         |     46     faculty. Somehow I won both. In 2014/15 I was nominated again for the best MSc | 
|         |     47     project supervisor, but did not win it. ;o) | 
|         |     48 </H4>   | 
|         |     49  | 
|         |     50 <ul class="striped"> | 
|         |     51 <li> <H4 id="regex">[CU1] Regular Expressions, Lexing and Derivatives</H4> | 
|         |     52  | 
|         |     53   <p> | 
|         |     54   <B>Description:</b>   | 
|         |     55   <A HREF="http://en.wikipedia.org/wiki/Regular_expression">Regular expressions</A>  | 
|         |     56   are extremely useful for many text-processing tasks, such as finding patterns in hostile  | 
|         |     57   <A HREF="https://www.snort.org">network traffic</A>, | 
|         |     58   lexing programs, syntax highlighting and so on. Given that regular expressions were | 
|         |     59   introduced in 1950 by <A HREF="http://en.wikipedia.org/wiki/Stephen_Cole_Kleene">Stephen Kleene</A>, | 
|         |     60   you might think regular expressions have since been studied and implemented to death. But you would definitely be | 
|         |     61   mistaken: in fact they are still an active research area. On the top of my head, I can give | 
|         |     62   you at least ten research papers that appeared in the last few years. | 
|         |     63   For example | 
|         |     64   <A HREF="http://www.home.hs-karlsruhe.de/~suma0002/publications/regex-parsing-derivatives.pdf">this paper</A>  | 
|         |     65   about regular expression matching and derivatives was presented in 2014 at the international  | 
|         |     66   FLOPS conference. Another <A HREF="https://nms.kcl.ac.uk/christian.urban/Publications/posix.pdf">paper</A> by my PhD student and me was presented in 2016 | 
|         |     67   at the international ITP conference. | 
|         |     68   The task in this project is to implement these results and use them for lexing.</p> | 
|         |     69  | 
|         |     70   <p>The background for this project is that some regular expressions are  | 
|         |     71   “<A HREF="http://en.wikipedia.org/wiki/ReDoS#Examples">evil</A>” | 
|         |     72   and can “stab you in the back” according to | 
|         |     73   this <A HREF="http://peterscott.github.io/2013/01/17/regular-expressions-will-stab-you-in-the-back/">blog post</A>. | 
|         |     74   For example, if you use in <A HREF="http://www.python.org">Python</A> or  | 
|         |     75   in <A HREF="http://www.ruby-lang.org/en/">Ruby</A> (or also in a number of other mainstream programming languages) the  | 
|         |     76   innocently looking regular expression <code>a?{28}a{28}</code> and match it, say, against the string  | 
|         |     77   <code>aaaaaaaaaaaaaaaaaaaaaaaaaaaa</code> (that is 28 <code>a</code>s), you will soon notice that your CPU usage goes to 100%. In fact, | 
|         |     78   Python and Ruby need approximately 30 seconds of hard work for matching this string. You can try it for yourself: | 
|         |     79   <A HREF="http://talisker.inf.kcl.ac.uk/cgi-bin/repos.cgi/afl-material/raw-file/tip/progs/catastrophic.py">catastrophic.py</A> (Python version) and  | 
|         |     80   <A HREF="http://talisker.inf.kcl.ac.uk/cgi-bin/repos.cgi/afl-material/raw-file/tip/progs/catastrophic.rb">catastrophic.rb</A>  | 
|         |     81   (Ruby version). Here is a similar problem with the regular expression <code>(a*)*b</code> in Java:  | 
|         |     82   <A HREF="http://talisker.inf.kcl.ac.uk/cgi-bin/repos.cgi/afl-material/raw-file/tip/progs/catastrophic.java">catastrophic.java</A>  | 
|         |     83   </p>  | 
|         |     84  | 
|         |     85   <p> | 
|         |     86   You can imagine an attacker | 
|         |     87   mounting a nice <A HREF="http://en.wikipedia.org/wiki/Denial-of-service_attack">DoS attack</A> against  | 
|         |     88   your program if it contains such an “evil” regular expression. But it can also happen by accident: | 
|         |     89   on 20 July 2016 the website <A HREF="http://stackstatus.net/post/147710624694/outage-postmortem-july-20-2016">Stack Exchange</A> | 
|         |     90   was knocked offline because of an evil regular expression. One of their engineers talks about this in this | 
|         |     91   <A HREF="https://vimeo.com/112065252">video</A>. A similar problem needed to be fixed in the | 
|         |     92   <A HREF="http://davidvgalbraith.com/how-i-fixed-atom/">Atom</A> editor. | 
|         |     93   A few implementations of regular expression matchers are almost immune from such problems. | 
|         |     94   For example, <A HREF="http://www.scala-lang.org/">Scala</A> can deal with strings of up to 4,300 <code>a</code>s in less than a second. But if you scale | 
|         |     95   the regular expression and string further to, say, 4,600 <code>a</code>s, then you get a <code>StackOverflowError</code>  | 
|         |     96   potentially crashing your program. Moreover (beside the "minor" problem of being painfully slow) according to this | 
|         |     97   <A HREF="http://www.haskell.org/haskellwiki/Regex_Posix">report</A> | 
|         |     98   nearly all regular expression matchers using the POSIX rules are actually buggy. | 
|         |     99   </p> | 
|         |    100  | 
|         |    101   <p> | 
|         |    102   On a rainy afternoon, I implemented  | 
|         |    103   <A HREF="http://talisker.inf.kcl.ac.uk/cgi-bin/repos.cgi/afl-material/raw-file/tip/progs/re3.scala">this</A>  | 
|         |    104   regular expression matcher in Scala. It is not as fast as the official one in Scala, but | 
|         |    105   it can match up to 11,000 <code>a</code>s in less than 5 seconds  without raising any exception | 
|         |    106   (remember Python and Ruby both need nearly 30 seconds to process 28(!) <code>a</code>s, and Scala's | 
|         |    107   official matcher maxes out at 4,600 <code>a</code>s). My matcher is approximately | 
|         |    108   85 lines of code and based on the concept of  | 
|         |    109   <A HREF="http://lambda-the-ultimate.org/node/2293">derivatives of regular expressions</A>. | 
|         |    110   These derivatives were introduced in 1964 by <A HREF="http://en.wikipedia.org/wiki/Janusz_Brzozowski_(computer_scientist)"> | 
|         |    111   Janusz Brzozowski</A>, but according to this | 
|         |    112   <A HREF="https://www.cs.kent.ac.uk/people/staff/sao/documents/jfp09.pdf">paper</A> had been lost in the “sands of time”. | 
|         |    113   The advantage of derivatives is that they side-step completely the usual  | 
|         |    114   <A HREF="http://hackingoff.com/compilers/regular-expression-to-nfa-dfa">translations</A> of regular expressions | 
|         |    115   into NFAs or DFAs, which can introduce the exponential behaviour exhibited by the regular | 
|         |    116   expression matchers in Python, Java and Ruby. | 
|         |    117   </p> | 
|         |    118  | 
|         |    119   <p> | 
|         |    120   Now the authors from the  | 
|         |    121   <A HREF="http://www.home.hs-karlsruhe.de/~suma0002/publications/regex-parsing-derivatives.pdf">FLOPS'14-paper</A> mentioned  | 
|         |    122   above claim they are even faster than me and can deal with even more features of regular expressions | 
|         |    123   (for example subexpression matching, which my rainy-afternoon matcher cannot). I am sure they thought | 
|         |    124   about the problem much longer than a single afternoon. The task  | 
|         |    125   in this project is to find out how good they actually are by implementing the results from their paper.  | 
|         |    126   Their approach to regular expression matching is also based on the concept of derivatives. | 
|         |    127   I used derivatives very successfully once for something completely different in a | 
|         |    128   <A HREF="https://nms.kcl.ac.uk/christian.urban/Publications/rexp.pdf">paper</A>  | 
|         |    129   about the <A HREF="http://en.wikipedia.org/wiki/Myhill–Nerode_theorem">Myhill-Nerode theorem</A>. | 
|         |    130   So I know they are worth their money. Still, it would be interesting to actually compare their results | 
|         |    131   with my simple rainy-afternoon matcher and potentially “blow away” the regular expression matchers  | 
|         |    132   in Python, Ruby and Java (and possibly in Scala too). The application would be to implement a fast lexer for | 
|         |    133   programming languages, or improve the network traffic analysers in the tools <A HREF="https://www.snort.org">Snort</A> and | 
|         |    134   <A HREF="https://www.bro.org">Bro</A>. | 
|         |    135   </p> | 
|         |    136  | 
|         |    137   <p> | 
|         |    138   <B>Literature:</B>  | 
|         |    139   The place to start with this project is obviously this | 
|         |    140   <A HREF="http://www.home.hs-karlsruhe.de/~suma0002/publications/regex-parsing-derivatives.pdf">paper</A> | 
|         |    141   and this <A HREF="https://nms.kcl.ac.uk/christian.urban/Publications/posix.pdf">one</A>. | 
|         |    142   Traditional methods for regular expression matching are explained | 
|         |    143   in the Wikipedia articles  | 
|         |    144   <A HREF="http://en.wikipedia.org/wiki/DFA_minimization">here</A> and  | 
|         |    145   <A HREF="http://en.wikipedia.org/wiki/Powerset_construction">here</A>. | 
|         |    146   The authoritative <A HREF="http://infolab.stanford.edu/~ullman/ialc.html">book</A> | 
|         |    147   on automata and regular expressions is by John Hopcroft and Jeffrey Ullmann (available in the library).  | 
|         |    148   There is also an online course about this topic by Ullman at  | 
|         |    149   <A HREF="https://www.coursera.org/course/automata">Coursera</A>, though IMHO not  | 
|         |    150   done with love.  | 
|         |    151   There are millions of other pointers about regular expression | 
|         |    152   matching on the Web. I found the chapter on Lexing in this | 
|         |    153   <A HREF="http://www.diku.dk/~torbenm/Basics/">online book</A> very helpful. Finally, it will | 
|         |    154   be of great help for this project to take part in my Compiler and Formal Language module (6CCS3CFL). | 
|         |    155   Test cases for “<A HREF="http://en.wikipedia.org/wiki/ReDoS#Examples">evil</A>” | 
|         |    156   regular expressions can be obtained from <A HREF="http://www.haskell.org/haskellwiki/Regex_Posix">here</A>. | 
|         |    157   </p> | 
|         |    158  | 
|         |    159   <p> | 
|         |    160   <B>Skills:</B>  | 
|         |    161   This is a project for a student with an interest in theory and with | 
|         |    162   good programming skills. The project can be easily implemented | 
|         |    163   in functional languages like | 
|         |    164   <A HREF="http://www.scala-lang.org/">Scala</A>, | 
|         |    165   <A HREF="http://fsharp.org">F#</A>,  | 
|         |    166   <A HREF="http://en.wikipedia.org/wiki/Standard_ML">ML</A>,   | 
|         |    167   <A HREF="http://haskell.org/haskellwiki/Haskell">Haskell</A>, etc. Python and other non-functional languages | 
|         |    168   can be also used, but seem much less convenient. If you do attend my Compilers and Formal Languages | 
|         |    169   module, that would obviously give you a head-start with this project. | 
|         |    170   </p> | 
|         |    171  | 
|         |    172 <li> <H4>[CU2] Grammars and Derivative-Based Parsing Algorithms</H4> | 
|         |    173  | 
|         |    174 <p> | 
|         |    175 Parsing is an old nut. Generations of software developers need to do parsing of data or text. | 
|         |    176 There are zillions of links, tools, papers and textbooks about parsing. One particular | 
|         |    177 <A HREF="https://dickgrune.com/Books/PTAPG_1st_Edition/BookBody.pdf">book</A> contains something | 
|         |    178 like 700 different algorithm, nicely analysed and described. Surely, parsing must be a solved problem. Or is it?  | 
|         |    179 Laurie Tratt has a blog <A HREF="https://tratt.net/laurie/blog/entries/parsing_the_solved_problem_that_isnt.html">post</A> | 
|         |    180 about <i>Parsing: The Solved Problem That Isn't</i>. IMHO parsing is still a wide open field and not solved at all. | 
|         |    181 PEG parsing, error reporting, error correction, runtime to name just a few are aspects that seem to cause headaches | 
|         |    182 to developers, and to researchers.</p>    | 
|         |    183  | 
|         |    184 <p> | 
|         |    185 A recent <A HREF="https://www.cl.cam.ac.uk/~jdy22/papers/a-typed-algebraic-approach-to-parsing.pdf">paper</A>  | 
|         |    186 follows an idea for regular expressions: it adapts the notion of  | 
|         |    187 derivatives of regular expressions to grammars. The idea is to implement in a functional programming language | 
|         |    188 the parsing algorithm proposed in this paper and to try it out with some sample data. There is also  | 
|         |    189 a recent PhD thesis about derivative-based parsing  | 
|         |    190 <A HREF="https://infoscience.epfl.ch/record/287059?ln=en">Efficient Parsing with Derivatives and Zippers</A>. | 
|         |    191 </p> | 
|         |    192  | 
|         |    193 <p> | 
|         |    194 <B>Literature:</B> <A HREF="https://www.cl.cam.ac.uk/~jdy22/papers/a-typed-algebraic-approach-to-parsing.pdf">paper</A>, | 
|         |    195 <A HREF="https://lara.epfl.ch/~kuncak/papers/EdelmannETAL20ZippyLLParsingDerivatives.pdf">paper</A>   | 
|         |    196 </p> | 
|         |    197  | 
|         |    198 <p> | 
|         |    199 <B>Skills:</B> See [CU1]. | 
|         |    200 </p> | 
|         |    201  | 
|         |    202  | 
|         |    203 <li> <H4>[CU3] A Compiler for a small Programming Language</H4> | 
|         |    204  | 
|         |    205   <p> | 
|         |    206   <b>Description:</b>  | 
|         |    207   Compilers translate high-level programs that humans can read and write into | 
|         |    208   efficient machine code that can be run on a CPU or virtual machine. | 
|         |    209   A compiler for a simple functional language generating X86 code is described | 
|         |    210   <A HREF="https://libraries.io/github/chameco/Shade">here</A>. | 
|         |    211   I recently implemented a very simple compiler for an even simpler functional | 
|         |    212   programming language following this  | 
|         |    213   <A HREF="https://www.cs.princeton.edu/~dpw/papers/tal-toplas.pdf">paper</A>  | 
|         |    214   (also described <A HREF="https://www.cs.princeton.edu/~dpw/papers/tal-tr.pdf">here</A>). | 
|         |    215   My code, written in <A HREF="http://www.scala-lang.org/">Scala</A>, of this compiler is  | 
|         |    216   <A HREF="https://nms.kcl.ac.uk/christian.urban/compiler.scala">here</A>. | 
|         |    217   The compiler can deal with simple programs involving natural numbers, such | 
|         |    218   as Fibonacci numbers or factorial (but it can be easily extended - that is not the point). | 
|         |    219   </p> | 
|         |    220  | 
|         |    221   <p> | 
|         |    222   While the hard work has been done (understanding the two papers above), | 
|         |    223   my compiler only produces some idealised machine code. For example I | 
|         |    224   assume there are infinitely many registers. The goal of this | 
|         |    225   project is to generate machine code that is more realistic and can | 
|         |    226   run on a CPU, like X86, or run on a virtual machine, say the JVM.  | 
|         |    227   This gives probably a speedup of thousand times in comparison to | 
|         |    228   my naive machine code and virtual machine. The project | 
|         |    229   requires to dig into the literature about real CPUs and generating  | 
|         |    230   real machine code.  | 
|         |    231   </p> | 
|         |    232   <p> | 
|         |    233   An alternative is to not generate machine code, but build a compiler that compiles to | 
|         |    234   <A HREF="http://www.w3schools.com/js/">JavaScript</A>. This is the language that is supported by most | 
|         |    235   browsers and therefore is a favourite | 
|         |    236   vehicle for Web-programming. Some call it <B>the</B> scripting language of the Web. | 
|         |    237   Unfortunately, JavaScript is also probably one of the worst | 
|         |    238   languages to program in (being designed and released in a hurry). <B>But</B> it can be used as a convenient target | 
|         |    239   for translating programs from other languages. In particular there are two | 
|         |    240   very optimised subsets of JavaScript that can be used for this purpose: | 
|         |    241   one is <A HREF="http://asmjs.org">asm.js</A> and the other is | 
|         |    242   <A HREF="https://github.com/kripken/emscripten/wiki">emscripten</A>. Since | 
|         |    243   last year there is even the official <A HREF="http://webassembly.org">Webassembly</A> | 
|         |    244   There is a <A HREF="http://kripken.github.io/emscripten-site/docs/getting_started/Tutorial.html">tutorial</A> for emscripten | 
|         |    245   and an impressive <A HREF="https://youtu.be/c2uNDlP4RiE">demo</A> which runs the | 
|         |    246   <A HREF="http://en.wikipedia.org/wiki/Unreal_Engine">Unreal Engine 3</A> | 
|         |    247   in a browser with spectacular speed. This was achieved by compiling the | 
|         |    248   C-code of the Unreal Engine to the LLVM intermediate language and then translating the LLVM | 
|         |    249   code to JavaScript. | 
|         |    250   </p> | 
|         |    251  | 
|         |    252   <p> | 
|         |    253   <B>Literature:</B> | 
|         |    254   There is a lot of literature about compilers  | 
|         |    255   (for example <A HREF="http://www.cs.princeton.edu/~appel/papers/cwc.html">this book</A> - | 
|         |    256   I can lend you my copy for the duration of the project, or this | 
|         |    257   <A HREF="http://www.diku.dk/~torbenm/Basics/">online book</A>). A very good overview article | 
|         |    258   about implementing compilers by  | 
|         |    259   <A HREF="http://tratt.net/laurie/">Laurie Tratt</A> is  | 
|         |    260   <A HREF="http://tratt.net/laurie/tech_articles/articles/how_difficult_is_it_to_write_a_compiler">here</A>. | 
|         |    261   An online book about the Art of Assembly Language is | 
|         |    262   <A HREF="http://flint.cs.yale.edu/cs422/doc/art-of-asm/pdf/">here</A>. | 
|         |    263   An introduction into x86 machine code is <A HREF="http://ianseyler.github.com/easy_x86-64/">here</A>. | 
|         |    264   Intel's official manual for the x86 instruction is  | 
|         |    265   <A HREF="http://download.intel.com/design/intarch/manuals/24319101.pdf">here</A>.  | 
|         |    266   Two assemblers for the JVM are described <A HREF="http://jasmin.sourceforge.net">here</A> | 
|         |    267   and <A HREF="https://github.com/Storyyeller/Krakatau">here</A>. | 
|         |    268   An interesting twist of this project is to not generate code for a CPU, but | 
|         |    269   for the intermediate language of the <A HREF="http://llvm.org">LLVM</A> compiler | 
|         |    270   (also described <A HREF="http://llvm.org/docs/LangRef.html">here</A>). If you want to see | 
|         |    271   what machine code looks like you can compile your C-program using gcc -S. | 
|         |    272   </p> | 
|         |    273   <p> | 
|         |    274   If JavaScript is chosen as a target instead, then there are plenty of <A HREF="http://www.w3schools.com/js/">tutorials</A> on the Web. | 
|         |    275   <A HREF="http://jsbooks.revolunet.com">Here</A> is a list of free books on JavaScript. | 
|         |    276   A project from which you can draw inspiration is this | 
|         |    277   <A HREF="http://jlongster.com/Outlet--My-Lisp-to-Javascript-Experiment">Lisp-to-JavaScript</A> | 
|         |    278   translator. <A HREF="https://bitbucket.org/ktg/parenjs/overview">Here</A> is another such project. | 
|         |    279   And <A HREF="https://github.com/viclib/liscript">another</A> in less than 100 lines of code. | 
|         |    280   <A HREF="http://en.wikipedia.org/wiki/CoffeeScript">Coffeescript</A> is a similar project | 
|         |    281   except that it is already quite <A HREF="http://coffeescript.org">mature</A>. And finally not to | 
|         |    282   forget <A HREF="http://www.typescriptlang.org">TypeScript</A> developed by Microsoft. The main | 
|         |    283   difference between these projects and this one is that they translate into relatively high-level | 
|         |    284   JavaScript code; none of them use the much lower levels <A HREF="http://asmjs.org">asm.js</A> and  | 
|         |    285   <A HREF="https://github.com/kripken/emscripten/wiki">emscripten</A>. | 
|         |    286   </p> | 
|         |    287   <p> | 
|         |    288   <B>Skills:</B>  | 
|         |    289   This is a project for a student with a deep interest in programming languages and | 
|         |    290   compilers. Since my compiler is implemented in <A HREF="http://www.scala-lang.org/">Scala</A>, | 
|         |    291   it would make sense to continue this project in this language. I can be | 
|         |    292   of help with questions and books about <A HREF="http://www.scala-lang.org/">Scala</A>. | 
|         |    293   But if Scala is a problem, my code can also be translated quickly into any other functional | 
|         |    294   language. Again,  it will be of great help for this project to take part in | 
|         |    295   my Compiler and Formal Language module (6CCS3CFL). | 
|         |    296   </p> | 
|         |    297  | 
|         |    298   <p> | 
|         |    299   <B>PS:</B> Compiler projects consistently received high marks in the past. | 
|         |    300   I have supervised eight so far and most of them received a mark above 70% - one even was awarded a prize. | 
|         |    301   </p> | 
|         |    302  | 
|         |    303 <li> <H4>[CU4] Webassembly Interpreter / Compiler</H4> | 
|         |    304  | 
|         |    305 <p> | 
|         |    306 Webassembly is a recently agreed standard for speeding up web applications in browsers. In this  | 
|         |    307 project the aim is to implement an interpreter or compiler for webassembly. There are already | 
|         |    308 <A HREF="https://github.com/WebAssembly/spec/tree/master/interpreter">reference interpreters</A>, | 
|         |    309 but people take different views, for example implement a  | 
|         |    310 <A HREF="https://groups.google.com/forum/#!topic/comp.lang.forth/CvNrP_AOmmw">Forth</A> language on top of webassembly.  | 
|         |    311 What is good about webassembly is that is a rather simple format, which can be generated quite | 
|         |    312 easily, unlike Java class files, which need some head-standing when you generate them.  | 
|         |    313 </p> | 
|         |    314  | 
|         |    315 <p> | 
|         |    316 A <A HREF="https://github.com/WebAssembly/spec/tree/master/interpreter">reference interpreter</A> for webassembly.   | 
|         |    317 </p> | 
|         |    318 <p> | 
|         |    319 <B>Skills:</B> See [CU1]. | 
|         |    320 </p> | 
|         |    321  | 
|         |    322 <li> <H4>[CU5] Slide-Making in the Web-Age</H4> | 
|         |    323  | 
|         |    324   <p> | 
|         |    325   The standard technology for writing scientific papers in Computer Science  is to use | 
|         |    326   <A HREF="http://en.wikipedia.org/wiki/LaTeX">LaTeX</A>, a document preparation | 
|         |    327   system originally implemented by <A HREF="http://en.wikipedia.org/wiki/Donald_Knuth">Donald Knuth</A> | 
|         |    328   and <A HREF="http://en.wikipedia.org/wiki/Leslie_Lamport">Leslie Lamport</A>. | 
|         |    329   LaTeX produces very pleasantly looking documents, can deal nicely with mathematical | 
|         |    330   formulas and is very flexible. If you are interested, <A HREF="http://openwetware.org/wiki/Word_vs._LaTeX">here</A> | 
|         |    331   is a side-by-side comparison between Word and LaTeX (which LaTeX “wins” with 18 out of 21 points). | 
|         |    332   Computer scientists not only use LaTeX for documents, | 
|         |    333   but also for slides (really, nobody who wants to be cool uses Keynote or Powerpoint). | 
|         |    334   </p> | 
|         |    335  | 
|         |    336   <p> | 
|         |    337   Although used widely, LaTeX seems nowadays a bit dated for producing | 
|         |    338   slides. Unlike documents, which are typically “static” and published in a book or journal, | 
|         |    339   slides often contain changing contents that might first only be partially visible and | 
|         |    340   only later be revealed as the “story” of a talk or lecture demands. | 
|         |    341   Also slides often contain animated algorithms where each state in the | 
|         |    342   calculation is best explained by highlighting the changing data. | 
|         |    343   </p> | 
|         |    344  | 
|         |    345   <p> | 
|         |    346   It seems HTML and JavaScript are much better suited for generating | 
|         |    347   such animated slides. This <A HREF="http://www.impressivewebs.com/html-slidedeck-toolkits/">page</A> | 
|         |    348   links to slide-generating programs using this combination of technologies.  | 
|         |    349   However, the problem with all of these project is that they depend heavily on the users being | 
|         |    350   able to write JavaScript, CCS or HTML...not something one would like to depend on given that | 
|         |    351   “normal” users likely only have a LaTeX background. The aim of this project is to invent a | 
|         |    352   very simple language that is inspired by LaTeX and then generate from code written in this language | 
|         |    353   slides that can be displayed in a web-browser. An example would be the | 
|         |    354   <A HREF="https://www.madoko.net">Madoko</A> project. | 
|         |    355   </p> | 
|         |    356  | 
|         |    357  <p> | 
|         |    358  This sounds complicated, but there is already some help available: | 
|         |    359  <A HREF="http://www.mathjax.org">Mathjax</A> is a JavaScript library that can | 
|         |    360  be used to display mathematical text, for example</p> | 
|         |    361  | 
|         |    362  <blockquote> | 
|         |    363  <p>When \(a \ne 0\), there are two solutions to \(ax^2 + bx + c = 0\) and they are | 
|         |    364  \(x = {-b \pm \sqrt{b^2-4ac} \over 2a}\).</p> | 
|         |    365  </blockquote> | 
|         |    366  | 
|         |    367  <p>  | 
|         |    368  by writing code in the familiar LaTeX-way. This can be reused. | 
|         |    369  Another such library is <A HREF="http://khan.github.io/KaTeX/">KaTeX</A>. | 
|         |    370  There are also plenty of JavaScript | 
|         |    371  libraries for graphical animations (for example | 
|         |    372  <A HREF="http://raphaeljs.com">Raphael</A>, | 
|         |    373  <A HREF="http://svgjs.com">SVG.JS</A>, | 
|         |    374  <A HREF="http://bonsaijs.org">Bonsaijs</A>, | 
|         |    375  <A HREF="http://jsxgraph.uni-bayreuth.de/wp/">JSXGraph</A>). The inspiration for how the user should be able to write | 
|         |    376  slides could come from the LaTeX packages <A HREF="http://en.wikipedia.org/wiki/Beamer_(LaTeX)">Beamer</A> | 
|         |    377  and <A HREF="http://en.wikipedia.org/wiki/PGF/TikZ">PGF/TikZ</A>. A slide-making project from which | 
|         |    378  inspiration can be drawn is <A HREF="http://maciejczyzewski.me/hyhyhy/">hyhyhy</A>. | 
|         |    379  </p> | 
|         |    380  | 
|         |    381   <p> | 
|         |    382   <B>Skills:</B>  | 
|         |    383   This is a project that requires good knowledge of JavaScript. You need to be able to | 
|         |    384   parse a language and translate it to a suitable part of JavaScript using | 
|         |    385   appropriate libraries. Tutorials for JavaScript are <A HREF="http://www.w3schools.com/js/">here</A>. | 
|         |    386   A parser generator for JavaScript is <A HREF="http://pegjs.majda.cz">here</A>. There are probably also | 
|         |    387   others. If you want to avoid JavaScript there are a number of alternatives: for example the | 
|         |    388   <A HREF="http://elm-lang.org">Elm</A> | 
|         |    389   language has been especially designed for implementing interactive animations, which would be | 
|         |    390   very convenient for this project. A nice slide making project done by a previous student is  | 
|         |    391   <A HREF="http://www.markslides.org">MarkSlides</A> by Oleksandr Cherednychenko.  | 
|         |    392   </p> | 
|         |    393  | 
|         |    394 <li> <H4>[CU6] Raspberry Pi's and Arduinos</H4> | 
|         |    395  | 
|         |    396   <p> | 
|         |    397   <B>Description:</B> | 
|         |    398   This project is for true hackers! <A HREF="http://en.wikipedia.org/wiki/Raspberry_Pi">Raspberry Pi's</A> | 
|         |    399   are small Linux computers the size of a credit-card and only cost £26, the | 
|         |    400   simplest version even costs only £5 (see pictures on the left below). They were introduced | 
|         |    401   in 2012 and people went crazy...well some of them. There is a | 
|         |    402   <A HREF="https://plus.google.com/communities/113390432655174294208?hl=en">Google+</A> | 
|         |    403   community about Raspberry Pi's that has more | 
|         |    404   than 300k of followers. A similar number follow the corresponding <A HREF="https://www.facebook.com/raspberrypi/">group</A> | 
|         |    405   on Facebook. It is hard to keep up with what people do with these small computers. The possibilities | 
|         |    406   seem to be limitless. The main resource for Raspberry Pi's is <A HREF="http://www.raspberrypi.org">here</A>. | 
|         |    407   There are <A HREF="https://www.raspberrypi.org/magpi/">magazines</A> dedicated to them and tons of | 
|         |    408   <A HREF="http://www.raspberrypi.org/phpBB3/viewforum.php?f=39">books</A> (not to mention | 
|         |    409   floods of <A HREF="https://www.google.co.uk/search?q=raspberry+pi">online</A> material, | 
|         |    410   such as the <A HREF="https://www.raspberrypi.org/magpi-issues/Projects_Book_v1.pdf">RPi projects book</A>). | 
|         |    411   Google just released a | 
|         |    412   <A HREF="http://googlecreativelab.github.io/coder/">framework</A> | 
|         |    413   for web-programming on Raspberry Pi's turning them into webservers. | 
|         |    414   In my home one Raspberry Pi has the very important task of automatically filtering out | 
|         |    415   nearly all advertisments using the  | 
|         |    416   <A HREF="https://github.com/pi-hole/pi-hole">Pi-Hole</A> software | 
|         |    417   (you cannot imagine what difference this does to your web experience). | 
|         |    418   </p> | 
|         |    419  | 
|         |    420   <p> | 
|         |    421   <A HREF="http://en.wikipedia.org/wiki/Arduino">Arduinos</A> are slightly older (from 2005) but still very cool (see picture on the right below). They | 
|         |    422   are small single-board micro-controllers that can talk to various external gadgets (sensors, motors, etc). Since Arduinos | 
|         |    423   are open-software and open-hardware there are many clones and add-on boards. Like for the Raspberry Pi, there | 
|         |    424   is a lot of material <A HREF="https://www.google.co.uk/search?q=arduino">available</A> about Arduinos. | 
|         |    425   The main reference is <A HREF="http://www.arduino.cc">here</A>. Like the Raspberry Pi's, the good thing about | 
|         |    426   Arduinos is that they can be powered with simple AA-batteries. | 
|         |    427   </p> | 
|         |    428  | 
|         |    429   <p> | 
|         |    430   I have several Raspberry Pi's including wifi-connectors and two <A HREF="http://www.raspberrypi.org/camera">cameras</A>. | 
|         |    431   I also have two <A HREF="http://www.freaklabs.org/index.php/Blog/Store/Introducing-the-Freakduino-Chibi-An-Arduino-based-Board-For-Wireless-Sensor-Networking.html">Freakduino Boards</A> that are Arduinos extended with wireless communication. I can lend them to responsible | 
|         |    432   students for one or two projects. However, the aim is to first come up with an idea for a project. Popular projects are | 
|         |    433   automated temperature sensors, network servers, robots, web-cams (<A HREF="http://www.secretbatcave.co.uk/electronics/shard-rain-cam/">here</A> | 
|         |    434   is a <A HREF="http://www.raspberrypi.org/archives/3547">web-cam</A> directed at the Shard that can | 
|         |    435   <A HREF="http://www.secretbatcave.co.uk/software/shard-rain-cam-quantifying-cloudy/">tell</A> | 
|         |    436   you whether it is raining or cloudy). There are plenty more ideas listed | 
|         |    437   <A HREF="http://www.raspberrypi.org/phpBB3/viewforum.php?f=15">here</A> for Raspberry Pi's and | 
|         |    438   <A HREF="http://playground.arduino.cc/projects/ideas">here</A> for Arduinos. | 
|         |    439   </p> | 
|         |    440  | 
|         |    441   <p> | 
|         |    442   There are essentially two kinds of projects: One is purely software-based. Software projects for Raspberry Pi's are often | 
|         |    443   written in <A HREF="http://www.python.org">Python</A>, but since these are Linux-capable computers any other | 
|         |    444   language would do as well. You can also write your own operating system as done | 
|         |    445   <A HREF="http://www.cl.cam.ac.uk/projects/raspberrypi/tutorials/os/">here</A>. For example the students | 
|         |    446   <A HREF="http://www.recantha.co.uk/blog/?p=4918">here</A> developed their own bare-metal OS and then implemented | 
|         |    447   a chess-program on top of it (have a look at their very impressive | 
|         |    448   <A HREF="http://www.youtube.com/watch?v=-03bouPsfEQ&feature=player_embedded">youtube</A> video). | 
|         |    449   The other kind of project is a combination of hardware and software; usually attaching some sensors | 
|         |    450   or motors to the Raspberry Pi or Arduino. This might require some soldering or what is called | 
|         |    451   a <A HREF="http://en.wikipedia.org/wiki/Breadboard">bread-board</A>. But be careful before choosing a project | 
|         |    452   involving new hardware: these devices | 
|         |    453   can be destroyed (if “Vin connected to GND” or “drawing more than 30mA from a GPIO” | 
|         |    454   does not make sense to you, you should probably stay away from such a project).  | 
|         |    455   </p> | 
|         |    456  | 
|         |    457   <center> | 
|         |    458     <img style="-webkit-user-select: none; cursor: -webkit-zoom-in;" | 
|         |    459          src="http://upload.wikimedia.org/wikipedia/commons/3/3d/RaspberryPi.jpg" | 
|         |    460          alt="Raspberry Pi" | 
|         |    461          width="313" height="209"> | 
|         |    462  | 
|         |    463     <img style="-webkit-user-select: none; cursor: -webkit-zoom-in;" | 
|         |    464          src="https://upload.wikimedia.org/wikipedia/commons/7/7e/Raspberry-Pi-Zero-FL.jpg" | 
|         |    465          alt="Raspberry Pi Zero" | 
|         |    466          width="313" height="209">   | 
|         |    467  | 
|         |    468     <img style="-webkit-user-select: none; cursor: -webkit-zoom-in;" | 
|         |    469          src="http://upload.wikimedia.org/wikipedia/commons/3/38/Arduino_Uno_-_R3.jpg" | 
|         |    470          alt="Arduino" | 
|         |    471          width="240" height="209"> | 
|         |    472   </center> | 
|         |    473    | 
|         |    474  | 
|         |    475   <p> | 
|         |    476   <B>Skills:</B>  | 
|         |    477   Well, you must be a hacker; happy to make things. Your desk might look like the photo below on the left. | 
|         |    478   The photo below on the middle shows an earlier student project which connects wirelessly a wearable Arduino (packaged | 
|         |    479   in a "self-3d-printed" watch) to a Raspberry Pi seen in the background. The Arduino in the foreground takes | 
|         |    480   measurements of  | 
|         |    481   heart rate and body temperature; the Raspberry Pi collects this data and makes it accessible via a simple | 
|         |    482   web-service. The picture on the right is another project that implements an airmouse using an Arduino. | 
|         |    483  | 
|         |    484   <center> | 
|         |    485     <img style="-webkit-user-select: none; cursor: -webkit-zoom-in;" | 
|         |    486          src="https://nms.kcl.ac.uk/christian.urban/rpi-photo.jpg" | 
|         |    487          alt="Raspberry Pi" | 
|         |    488          width="209" height="313"> | 
|         |    489  | 
|         |    490     <img style="-webkit-user-select: none; cursor: -webkit-zoom-in;" | 
|         |    491          src="https://nms.kcl.ac.uk/christian.urban/rpi-watch.jpg" | 
|         |    492          alt="Raspberry Pi" | 
|         |    493          width="450" height="254"> | 
|         |    494  | 
|         |    495     <img style="-webkit-user-select: none; cursor: -webkit-zoom-in;" | 
|         |    496          src="https://nms.kcl.ac.uk/christian.urban/rpi-airmouse.jpg" | 
|         |    497          alt="Raspberry Pi" | 
|         |    498          width="250" height="254">   | 
|         |    499   </center><p> | 
|         |    500  | 
|         |    501  | 
|         |    502     A really cool project using a toy helicopter and two Raspberry Pi's was done by Nikolaos Kyknas. He transformed | 
|         |    503     an off-the-shelf toy helicopter into an autonomous flying machine. He attached a Raspberry Pi Zero and an ultrasound | 
|         |    504     sensor to the helicopter for measuring the distance from ground. Another Raspberry Pi is attached to the “ground control | 
|         |    505     unit” in order to give instructions to the throttle of the helicopter. Both Raspberry Pi's communicate over WiFi for calculating | 
|         |    506     the next flight instruction. The goal is to find and maintain a steady altitude. Sounds simple? Well, not so fast!  | 
|         |    507     First you need to get the balance of the helicopter plus Raspberry Pi plus its power source just right, | 
|         |    508     otherwise the helicopter will simply take off in random directions. Also the flight instructions need to be just right, | 
|         |    509     otherwise the helicopter would at best “oscillate” around the set altitude, but never be steady. To solve this problem,  | 
|         |    510     Nikolaos used exactly the same algorithm that keeps cars at a steady pace when in cruise control.  | 
|         |    511  | 
|         |    512     <center> | 
|         |    513       <video width="320" height="576" controls> | 
|         |    514         <source src="https://nms.kcl.ac.uk/christian.urban/h1.mp4" type="video/mp4"> | 
|         |    515           Your browser does not support the video tag. | 
|         |    516       </video> | 
|         |    517       <video width="320" height="576" controls> | 
|         |    518         <source src="https://nms.kcl.ac.uk/christian.urban/h3.mp4"> | 
|         |    519           Your browser does not support the video tag. | 
|         |    520       </video> | 
|         |    521     </center>   | 
|         |    522  | 
|         |    523 <li> <H4>[CU7] An Infrastructure for Displaying and Animating Code in a Web-Browser</H4> | 
|         |    524    | 
|         |    525 <p> | 
|         |    526   <B>Description:</B> | 
|         |    527   The project aim is to implement an infrastructure for displaying and | 
|         |    528   animating code in a web-browser. The infrastructure should be agnostic | 
|         |    529   with respect to the programming language, but should be configurable. | 
|         |    530   I envisage something smaller than the projects  | 
|         |    531   <A HREF="http://www.pythontutor.com">here</A> (for Python), | 
|         |    532   <A HREF="http://ideone.com">here</A> (for Java), | 
|         |    533   <A HREF="http://codepad.org">here</A> (for multiple languages), | 
|         |    534   <A HREF="http://www.w3schools.com/html/tryit.asp?filename=tryhtml_intro">here</A> (for HTML) | 
|         |    535   <A HREF="http://repl.it/languages/JavaScript">here</A> (for JavaScript), | 
|         |    536   and <A HREF="http://www.scala-tour.com/#/welcome">here</A> (for Scala). | 
|         |    537   </p> | 
|         |    538  | 
|         |    539   <p> | 
|         |    540   The tasks in this project are being able (1) to lex and parse languages and (2) to write an interpreter. | 
|         |    541   The goal is to implement this as much as possible in a language-agnostic fashion. | 
|         |    542   </p> | 
|         |    543  | 
|         |    544   <p> | 
|         |    545   <B>Skills:</B>  | 
|         |    546   Good skills in lexing and language parsing, as well as being fluent with web programming (for | 
|         |    547   example JavaScript). | 
|         |    548   </p> | 
|         |    549  | 
|         |    550  | 
|         |    551 <li> <H4>[CU8] Proving the Correctness of Programs</H4> | 
|         |    552  | 
|         |    553  <p> | 
|         |    554  I am one of the main developers of the interactive theorem prover | 
|         |    555  <A HREF="http://isabelle.in.tum.de">Isabelle</A>. This theorem prover | 
|         |    556  has been used to establish the correctness of some quite large | 
|         |    557  programs (for example an <A HREF="http://ertos.nicta.com.au/research/l4.verified/">operating system</A>). | 
|         |    558  Together with colleagues from Nanjing, I used this theorem prover to establish the correctness of a | 
|         |    559  scheduling algorithm, called | 
|         |    560  <A HREF="http://en.wikipedia.org/wiki/Priority_inheritance">Priority Inheritance</A>, | 
|         |    561  for real-time operating systems. This scheduling algorithm is part of the operating | 
|         |    562  system that drives, for example, the  | 
|         |    563  <A HREF="http://en.wikipedia.org/wiki/Mars_Exploration_Rover">Mars rovers</A>. | 
|         |    564  Actually, the very first Mars rover mission in 1997 did not have this | 
|         |    565  algorithm switched on and it almost caused a catastrophic mission failure (see | 
|         |    566  this youtube video <A HREF="http://www.youtube.com/watch?v=lyx7kARrGeM">here</A> | 
|         |    567  for an explanation what happened). | 
|         |    568  We were able to prove the correctness of this algorithm, but were also able to | 
|         |    569  establish the correctness of some optimisations in this | 
|         |    570  <A HREF="https://nms.kcl.ac.uk/christian.urban/Publications/pip.pdf">paper</A>. | 
|         |    571  </p> | 
|         |    572  | 
|         |    573  <p>On a much smaller scale, there are a few small programs and underlying algorithms where it | 
|         |    574  is not really understood whether they always compute a correct result (for example the | 
|         |    575  regular expression matcher by Sulzmann and Lu in project [CU1]). The aim of this | 
|         |    576  project is to completely specify an algorithm in Isabelle and then prove it correct (that is, | 
|         |    577  it always computes the correct result). | 
|         |    578 </p> | 
|         |    579  | 
|         |    580   <p> | 
|         |    581   <B>Skills:</B>  | 
|         |    582   This project is for a very good student with a knack for theoretical things and formal reasoning. | 
|         |    583   </p> | 
|         |    584  | 
|         |    585 <li> <H4>[CU9] Anything Security Related that is Interesting</H4> | 
|         |    586    | 
|         |    587 <p> | 
|         |    588 If you have your own project that is related to security (must be | 
|         |    589 something interesting), please propose it. We can then have a look | 
|         |    590 whether it would be suitable for a project. | 
|         |    591 </p> | 
|         |    592  | 
|         |    593 <li> <H4>[CU10] Anything Interesting in the Areas</H4> | 
|         |    594    | 
|         |    595 <ul> | 
|         |    596 <li><A HREF="http://elm-lang.org">Elm</A> (a reactive functional language for animating webpages; have a look at the cool examples, or <A HREF="http://pragmaticstudio.com/blog/2014/12/19/getting-started-with-elm">here</A> for an introduction) | 
|         |    597 <li><A HREF="http://www.smlserver.org/smltojs/">SMLtoJS</A> (a ML compiler to JavaScript; or anything else related to | 
|         |    598   sane languages that compile to JavaScript) | 
|         |    599 <li>Any statistical data related to Bitcoins (in the spirit of this | 
|         |    600 <A HREF="http://people.csail.mit.edu/spillai/data/papers/bitcoin-transaction-graph-analysis.pdf">paper</A> or | 
|         |    601   this <A HREF="https://eprint.iacr.org/2012/584.pdf">one</A>; this will probably require some extensive C knowledge or any | 
|         |    602   other heavy-duty programming language) | 
|         |    603 <li>Anything related to programming languages and formal methods (like | 
|         |    604   <A HREF="http://matt.might.net/articles/intro-static-analysis/">static program analysis</A>)   | 
|         |    605 <li>Anything related to low-cost, hands-on hardware like Raspberry Pi, Arduino, | 
|         |    606   <A HREF="http://en.wikipedia.org/wiki/Cubieboard">Cubieboard</A> | 
|         |    607 <li>Anything related to unikernel operating systems, like | 
|         |    608   <A HREF="http://www.xenproject.org">Xen</A> or | 
|         |    609   <A HREF="http://www.openmirage.org">Mirage OS</A> | 
|         |    610 <li>Any kind of applied hacking, for example the Arduino-based keylogger described | 
|         |    611    <A HREF="http://samy.pl/keysweeper/">here</A> | 
|         |    612 <li>Anything related to code books, like this | 
|         |    613    <A HREF="http://www.joelotter.com/kajero/">one</A> | 
|         |    614 </ul> | 
|         |    615  | 
|         |    616  | 
|         |    617  | 
|         |    618 <li> <H4>Earlier Projects</H4> | 
|         |    619  | 
|         |    620  I am also open to project suggestions from you. You might find some inspiration from my earlier projects: | 
|         |    621  <A HREF="https://nms.kcl.ac.uk/christian.urban/bsc-projects-12.html">BSc 2012/13</A>,  | 
|         |    622  <A HREF="https://nms.kcl.ac.uk/christian.urban/msc-projects-12.html">MSc 2012/13</A>,  | 
|         |    623  <A HREF="https://nms.kcl.ac.uk/christian.urban/bsc-projects-13.html">BSc 2013/14</A>, | 
|         |    624  <A HREF="https://nms.kcl.ac.uk/christian.urban/msc-projects-13.html">MSc 2013/14</A>,  | 
|         |    625  <A HREF="https://nms.kcl.ac.uk/christian.urban/bsc-projects-14.html">BSc 2014/15</A>, | 
|         |    626  <A HREF="https://nms.kcl.ac.uk/christian.urban/msc-projects-14.html">MSc 2014/15</A>,  | 
|         |    627  <A HREF="https://nms.kcl.ac.uk/christian.urban/bsc-projects-15.html">BSc 2015/16</A>, | 
|         |    628  <A HREF="https://nms.kcl.ac.uk/christian.urban/msc-projects-15.html">MSc 2015/16</A>,  | 
|         |    629  <A HREF="https://nms.kcl.ac.uk/christian.urban/bsc-projects-16.html">BSc 2016/17</A>, | 
|         |    630  <A HREF="https://nms.kcl.ac.uk/christian.urban/msc-projects-16.html">MSc 2016/17</A>, | 
|         |    631  <A HREF="https://nms.kcl.ac.uk/christian.urban/bsc-projects-17.html">BSc 2017/18</A>, | 
|         |    632  <A HREF="https://nms.kcl.ac.uk/christian.urban/msc-projects-17.html">MSc 2017/18</A>, | 
|         |    633  <A HREF="https://nms.kcl.ac.uk/christian.urban/bsc-projects-17.html">BSc 2018/19</A> | 
|         |    634 </ul> | 
|         |    635 </TD> | 
|         |    636 </TR>   | 
|         |    637 </TABLE> | 
|         |    638          | 
|         |    639 <P> | 
|         |    640 2018-09-24 12:12:35 by Christian Urban | 
|         |    641 <a href="https://validator.w3.org/check/referer">[Validate this page.]</a> | 
|         |    642 </P> | 
|         |    643 </BODY> | 
|         |    644 </HTML> | 
|         |    645   | 
|         |    646   | 
|         |    647   |