PCSU stands for POSIX Compliant Shell Utilities; it is an effort to create a set of 100% POSIX compliant shell utilities according to the [POSIX.1-2008 standard] (http://pubs.opengroup.org/onlinepubs/9699919799/). As far as I have been able to glean, there is no mention of implemention details for the utilities, only what their required behavior needs to be. Thus, the initial effort is to create them in Py3K. Implentations in other languages, such as Racket, may be added during future efforts.
Taking a page from the reference implementations of the PNG and DEFLATE standards, PCSU is liberally licensed under the zlib license. Thus, it can used almost anywhere for any purpose with the one main caveat being the the license notice must accompany source distributions. See the LICENSE file for details.
Certain utilities mention in the standard are currently beyond the scope of this project (and some may never be within the scope of this project). A good example would be shell built-in commands (such as cd); these will not be implemented unless/until a POSIX compliant shell is also implemented as part this project. A good example of a utility that is completely beyond the scope of this project would be creating a C compiler, thus the c99 shell utility will likely never be implemented as part of this effort. Besides, there are number of mature and robust C compilers that are liberally licensed and freely available (e.g. Clang, GCC, etc.).
This project is mainly for my own learning and benefit. If it is useful to someone else, then all the better. Also, it would be nice to have a reference implementation of strictly POSIX compliant utilities. It also makes it possible for people testing shell scripts to be sure their scripts will be completely portable. Implementations such as the GNU Coreutils often extend or sometimes even change the default behavior of these core utilities; this can give a false impression of functionality that may not exist in all implementations. Having utilities that are strictly compliant allows portability testing that might not otherwise be possible.
Compared to most other languages, Python is easy to use and fast to prototype in. Personal computers have become fast enough that decreases in speed due to an interpreted language will be relatively small, at least from a user's perception. Also, most bottlenecks in programming are from IO, which is independent of implementation language anyway.
Although Python 3 is not yet as widespread as its predecessor, Python 2, it will never get to that point of saturation until more people start using it. Since this is a new project, I see no reason not to use the latest version of the language since the previous version is rapidly moving into EOL status. Currently, the utilities require at least version 3.2 since they utilitize the argparse module (since argparse is also included in Python 2.7, it may theoretically be possible to use that version of Python as well, although no official testing has been done to verify that).
Lastly, future implementations in other languages, such as Racket, are both possible and probable.
I am currently only one person working on this in my free time, so my resources are limited. However, if you want a certain utility and have some skill in programing, you are free to contribute to the project. Your knowledge and support would be greatly appreciated.