Peter Bindels is a Software Engineer who graduated in both Technical Informatics and Higher Informatics in February 2006 and has since started working at www.topic.nl, being placed at www.grassvalley.com . He develops AtlantisOS and moderates the Mega-Tokyo OS-Dev forum in his free time, and occasionally plans ahead & does some work on his view of a new computing environment.
You can 
drop him a IM message through MegaTokyo message board ...
I can be contacted at dascandy@gmail.com. I might occasionally accidentally send out an html mail, forgive me, haven't had time to write an email client.
If you came here for a view on who I am, stop reading. The next is about my view on computing, which is just my view.
My view of the new computing environment consists most of all about kicking out all the stuff that currently exists. New compilers, new assemblers, new ideas, and most of all, all of the old experience and concepts but none of the code. The code is buggy (bugtraq shows it), and converting code with bugs to new code without is harder than making new code without (my personal experience). So, I'm setting up a platform (like GNU) to create good software, not complete (FSF/GNU etc) or commercial (Microsoft/all the other companies). Complete results in lots of incompatible features at war with eachother (quote from PL/1 language comment, PL/1 was also designed to be complete) and commercial results in abusing methods such as adding features to keep people buying, in essence, the same problem as complete has.
Instead, think about a good thing for instance. It does what it should do, and does it well (GNU philosophy). It does that with one specific set of input files, a limited set of output methods, and does not grow to embrace&extend the rest of the world (which is what GNU does wrong).
Project list of all the things I've thought about that are needed:
Lowest level, for the programmers & designers, without graphical environment
Assembler (casm, started design) C Compiler (ccc, small prototype done) C++ Compiler (ccxx, not started) Operating System (atlantisos, started & working on it) Lexer generator (clex, started, not receiving the work it should get, but it is nearly functional enough for a 1.0 release. Just need to get that damn NFA->DFA converter started - still) Parser generator (cpg or something, didn't think of something good) Shell (ncsh, shell part is done, history still depends on readline)
Medium level, for the lowlevel programmers to design and the normal users and highlevel programmers to base their scripts and basic programs on:
C library (implement only, started) C Graphics Library C++ stl (implement only, started) C++ OO Graphics Library (partially started, just one file as of yet) Program Module System (thought going in at a constant rate, starters on SVN server, idea done at least once, it appears to work quite well)
Highlevel, the things users use as programs:
Word processor Spreadsheet Presentation program Programming environment Some small games (Set, Breakout & Tetris are candidates) Database system Web browser Email client (made for Windows first, then recoded) Complete Graphical User Interface
And in the end, there'll be a system that's stable, good, bugfree (at least, as far as possible), interoperable and easily programmable.
At least, I hope :)