Lectures 32 & 33 homework

Lecture 32 homework (reversible languages)

1. [7 points] Short paper suggestion.  Read up on the existing reversible high-level languages Janus, PsiLisp, and R.  Write a short paper to compare and contrast them.
2. [10+ points] Small programming projects.  Write a program in the R programming language to do one of the reversible algorithms mentioned during lec. 33 - sorting, searching, division, factoring, matrix multiplication or inversion, all-pairs shortest-path, etc.   You could extend the Schroedinger simulator to 2-D.  You can also make up other problems to solve - anything of interest to you.  Try to make your algorithm as spacetime-efficient as you can.  Use the latest version of RCOMP (the R compiler), which is at http://www.cise.ufl.edu/research/revcomp/users/cclark/rcomp-fall2001/ (see the readme files in there).  It runs under a common lisp such as clisp or gcl (the latter may require some porting).  Compile the programs and run them under PendVM; the latest version of which is at http://www.cise.ufl.edu/research/revcomp/users/cclark/pendvm-fall2001/.  You can get an additional 5 points for each program you write after the first, as long as each program is substantially different from the previous ones.  Turn in a report with printouts of your source code, compiled code, and sample inputs & outputs or a trace of an interactive session.   Your report should also say what is the order of growth of space, time, and spacetime costs in each of your algorithms.
3. [20 points] Medium programming project.  Extend the R compiler to implement some new language features.  You could add additional types of control-flow constructs, like the ones in Janus.  Or, you could add language support for more sophisticated data types such as structures and classes.  Or, you could add support for a new floating-point data type (with an associated arithmetic library).
4. [100 points] Large programming project.  Design & implement your own high-level reversible programming language.  It should be at least as expressive as the existing reversible languages Janus, PsiLisp, and R.  Implement an interpreter or compiler for the language.  Turn in the source code and a report documenting the language and tool in detail.

Lecture 33 homework (reversible algorithms)

1. [3 points ea.] Algorithm design & analysis.  Invent and write out in reversible pseudo-code a reversible algorithm for some problem(s) of interest.  This may include any of the reversible algorithms mentioned in lec. 33 where the algorithm was not already written out for you in detail - sorting, searching, division, factoring, matrix multiplication or inversion, all-pairs shortest-path, etc.  You can also make up other problems to solve - anything of interest to you.  Say what is the order of growth of space, time, and spacetime costs in each of your algorithms, as a function of input size.  If the algorithm is only partially reversible, also specify the order of growth of entropy produced (# of irreversible erasures) as a function of input size.  Be very careful to make sure that you do not accidentally use a step that is irreversible without realizing it!  To help you make sure of this, you may want to stick with the language constructs that are available in the existing reversible languages such as Janus, Psi-Lisp, and R.