Linux format review: Fedora Core 4

[Rahul Sundaram]

Hi

Ever since reading http://distrowatch.com/weekly.php?issue=20050711#1 I 
have been curious to see this review and thanks to Andy Hudson for 
sending me a copy of this article.  This review which awards 4/10 points 
for the Fedora Core 4 release runs into a couple of pages and is a nice 
change from the usual ill informed MP3 rants and cursory looks. Time for 
feedback now!

The reviews goes into a general introduction that notes the change in 
the default GNOME desktop theme from Bluecurve to Clearlooks, switch to 
GCC 4.0 and stresses the movement of packages from core to extras as a 
controversy that has alarmed Fedora users and notes that PPC is now a 
supported architecture.  The controversy over extras has been compared 
to the refusal of Red Hat to include 2.6 kernel in RHEL3.  RHEL 3 was 
released on October 2003, a few months before the first release of 2.6 
version of the Linux kernel, providing such a major version bump during 
errata updates is something which no distribution has done in my 
knowledge. In both cases, the "annoyances" for any users are seemingly 
irrational to me. The review has several sections covering different 
aspects of the changes in this release.

Software shelved:
http://distrowatch.com/weekly.php?issue=20050711#1
It has been that mentioned many  packages like Abiword, Exim, XEmacs and 
XFCE has moved into extras as a effort to reduce bloat and this would be 
a problem for users who want to retrieve such packages over a dial up 
connection. I dont buy this argument. Lets assume for a moment that 
Fedora Core 4 included all of these packages in a 5 CD collection 
without moving anything to the extras repository, would users with a 
dial up connection be able to download it then?.  At any case, Fedora 
Core wouldnt ever serve as a set of all the software that any user could 
ever want. As Fedora Core gets trimmed to a more manageable collection 
of default applications, integration with Fedora Extras and potentially 
other third party repositories should be transparent enough that the 
users wouldnt have to care which repository their favorite application 
is in . As a first step, FC4 ships with the extras repository enabled by 
default as mentioned in the review.  Now further being work in Anaconda 
in using a yum backend would enable users to mix and match their 
applications from various repositories during installation time. If 
there is sufficient interests, anyone could spin off the other Fedora 
compatible repositories into ISO images for redistribution.

GUI Limbo:

GCC 4.0 has been described as a gamble here. Red Hat developers have 
significant stake in the development of GCC with in depth knowledge that 
enabled them to rebuild nearly all of Fedora Core with this compiler 
including fixes to packages and even to the compiler when necessary. A 
previous version has been shipped for compatibility reasons. While 
pushing new technology always has risks, it also has its benefits and 
this is what Fedora is meant to be.  The focus on whether GCC 4 would 
yield any performance benefits, in my opinion misses a important point. 
GCC 4 includes a significantly improved version of GCJ which has enabled 
the inclusion of Eclipse, Openoffice.org 2.0 milestone release including 
the Java parts, Apache Jakarta among several natively compiled Java 
components. This is a significant advancement of a completely Free Java 
stack which includes extensive work done over many years,  the 
importance of this and the relationship with the new compiler seemed to 
have gone unnoticed in the review.

The criticism of the lack of improvements in the system configuration 
tools (system-config*) especially system-config-package's lack of 
understanding of the yum repositories is indeed valid and significant 
work is being done during the FC5 timeframe to address this.

Comfortably yum:

This short section has a well deserved praise in this section for 
improvements in speed using XML headers and SQLite backend. Kudos to the 
yum developers on this. 

Core Proposition:

The review mentions that this release is just a standard update despite 
the inclusion of Xen and  trimming down a few packages in comparison 
with SUSE 9.3 Pro which includes Mono and Beagle. Setting aside the Mono 
factor*,  Fedora Core 4's inclusion of GFS cluster filesystem, Evince, 
Apache Jakarta along with the improvements in the SELinux policies have 
gone unnoticed in the review

http://fedora.redhat.com/docs/release-notes/fc4/#sn-new-in-fc

Taking into consideration, the only major criticism, "GUI limbo" as 
mentioned in the review, the overall score seems unfair to me even after 
reading the comments in the forum from the site admin, who mentions that 
the scores have been readjusted in such a way that 5/10 means a average 
one which many distributions would get.

http://www.linuxformat.co.uk/index.php?name=PNphpBB2&file=viewtopic&t=709

A amusing comment from the side admin caught my eye there "On the flip 
side, I would never, ever use it as a server distro either, simply 
because it doesn't provide a good enough security infrastructure for my 
requirements. ". FC4 includes Exec Shield, GCC 4 security improvements 
and 91 daemons covered under SELinux targeted policy by default along 
with the strict policy as an alternative. I would have thought that 
would provided enough of a security infrastructure in comparison to any 
other distribution or even operating system out there.

regards
Rahul

* http://fedora.redhat.com/docs/release-notes/fc4/errata/#sn-why-no-mp3

PS: It would perhaps be a good idea to look at how reviewers go about 
understanding new release highlights to help them see through such 
changes in an evident manner. It would be interesting to hear Andy 
Hudson's comments on his approach