This is a version of Linux based on Slackware that has been constructed using the "PGCC" Pentium-optimized version of GCC, so as to (at least hopefully) provide the fastest possible performance on the IA-32 architecture. It is also reputed to have been "tuned" for security, and might thus be more secure than some of the other distributions...
Seeking to build a generic distribution usable for any architecture
It was once based on Slackware, but uses a somewhat makefile-like dependency scheme in the rc subsystem to start up system services, eschews the use of GLIBC , instead using libc version 4, is compatible with the a.out binary format, and uses a fair bit of BSD "user world," whether from FreeBSD or OpenBSD.
Originally produced by people at LinuxCare, this provides a Linux distribution stored on business-card-sized compact disks.
This is a minimalist Linux distribution; it is consciously missing lots of stuff that is common in bigger distributions.
One interesting thing is that it includes, as compilers, GCC and gforth, and is thus intended to be programmed using Forth.
cLIeNUX is configured for client-only use; it uses a somewhat peculiar Dotted Standard Filename Hierarchy that is decidedly different from the UNIX tradition of having notable directories such as /usr , /sbin, /etc , /bin.
GoboLinux is an alternative Linux distribution which redefines the entire filesystem hierarchy. In GoboLinux paths such as /Programs/XFree86/4.3/ and /System/Settings/BootScripts/Reboot are used. The main idea of the alternative hierarchy is to store all files belonging to an application in its own separate subtree; therefore we have directories such as /Programs/GCC/2.95.3/lib. There are then symbolic links to mimic a traditional Unix hierarchy, with /usr, /sbin and such.
A Linux distribution using a BSD Ports -like package management system called Portage
Recently discussed in an IBM DeveloperWorks article, Gentoo Portage .
Some of the aficionados of Gentoo get a little overenthusiastic. Many seem to be expatriate Mandrake fans that get way too excited about the ability to compile code specifically for their rice cooker, um, particular Pentium variation.
The grand error is in imagining that it is actually useful to locally compile code in order to improve efficiency. Reality is that if the CPU load on your system is sufficiently high that the optimizations would do any good, any good that would be done would be spectacularly outweighed by the cost of compiling everything.
Claims about the value of optimization are vastly overrated.
It takes a lot of work to do a meaningful benchmark, controlling for all the possible extraneous effects, and you never see any of that.
All you actually get are vague anecdotes.
There are plenty of claims about superiority of Gentoo's package management over other systems.
That superiority only exists if you ignore the capabilities of the other systems package management systems.
Some include additional "value-added" commercial software packages and feature such things as customized versions of the Red Hat distribution or extended documentation.