This module does have some value. As most functional programmers will say, you rebuild the language to solve your problem. This module does solve an abstraction problem. There is always a cost, though. In this case, the module is somewhat large.
I do, however, like many others, have issues with the name of the module. Anyone who's attempted to find a useful module for Python or Ruby will know what I'm talking about. Namespace pollution is a serious issue for CPAN and a valid criticism here. On top of that, the name is ... well, it just sucks. The UNIX:: namespace should be reserved for modules that are UNIX-specific. This module does not appear to be. Also, the name is not at all descriptive of either its functionality or or other modules it may be related to. For example, the documentation mentions Pipe and Proc::SafePipe were investigated and didn't fit the purpose. So maybe Proc::SafePipe::Simple would be a better name? As far as the idiom being UNIX inspired, so if I had a module that was buggy and broke down, should I call it "Yugo". That's just ridiculous. So, if I had a module that I developed on Windows that listed a directory with a function called "dir", I could call it "Win32".
Finally, using the documentation to argue your point. Poor taste. Newsgroups, PerlMonks, etc. are great places for this, not your module.
Simple modules for very simple things are useful sometimes (I use
File::Slurp regularly for example). So I don't mind this module being
on CPAN as *long* as it is renamed. As other reviews said it has
*nothing* to do with Unix and if it had it still has no rights to
occupy top level namespace. As Chris Reinhardt suggested it should be
named like Proc::Pipe::Simple - quite clear descriptive name (Proc for
module category, Pipe for what it does, Simple for simplicity of the
interface). Lowest rating for bad naming.
Firstly, the name is simply horrendous. The functionality of the module has almost nothing to do with UNIX. Proc::Pipe::Simple would be a better name. Secondly, releasing a new version of a distribution without updating the change-log is a pet-peeve of mine. Third, nay-saying your critics in the documentation is at best a faux-pa. The reviews of this module have more than one good comment on how to improve it, they should be listened to, not rebutted.
Aside from being almost entirely pointless, this is supremely badly named. Not only does it not merit a top-level name space IT'S NOT EVEN UNIX-SPECIFIC. I cannot comprehend the thought processes behind this.
This module does not add much to standard perl. As the POD says, it is just syntactic sugar.
Unix.pm lets you write "$foo = UNIX $bar;" instead of "open $foo, '-|', $bar". It takes 800 kB (on my system) to get this functionality.
The POD is broken in the SYNOPSIS. The author should run the POD through pod2text before making dist.
The module includes a test script, but it is not portable and does not check for errors. (The test script is the same code as the SYNOPSIS.) The author should really read Test's documentation, and learn to use ok().