[nycbug-talk] review of "state of demon address"

[G. Rosamond]

if i get some positive comments here, i will submit to the nycbug site. 
  There's nothing wrong with us putting up pieces like this in our 
library section, IMHO.

yes, i write compulsively and i did not even reread this.

*	*	*	*

I hold the frustration many others have with the recent 
osOpinion/osViews article that recently ran titled "The State of Demon 
(sic) Address" by Paul Webb.  What "demon" he refers to is a mystery: 
he must be referring to the "daemon" logo of the BSDs and more 
generally with the "daemon" of processes on a UNIX box.

Parenthetically, my reading of the article is based upon a posting 
supplied by a helpful Slashdot reader, as the osViews web site was not 
prepared to handle the traffic generated by article about a often 
considered peripheral operating system family.  I am restricted by only 
having access to this particular post.

  Paul Webb seems quite excited by the state of the BSD projects, which 
is rather surprising considering his review is full of dated and 
confused information.  I myself wonder if Paul has ever installed BSD, 
much less viewed any of the projects web sites in the past few months 
with non-enebriated eyes.

Paul breaks up his meager review by project: FreeBSD, NetBSD, OpenBSD 
then Darwin.  It's notable to mention that he makes no mention of 
DragonFlyBSD, which has caused excitement beyond the DragonFly 
community.  It's safe to guess that as it's only been around a year or 
so means that Paul hasn't quite caught up with it yet.

FreeBSD

Paul's approach to FreeBSD is way more confusing than the project 
itself having  4.x stable and 5.x "new technology" branches.  Sure, 
having two branches in simultaneous development may be confusing to the 
sysadmin figuring out which one to run for this or that server.  
However, it's far more confusing to read an excited interview that 
opens with the first project being in a "precarious state."  What is 
this, a Vincent Price movie?

He is delighted that FreeBSD has almost hit a critical mass in the 
corporate world, which is not provided any backing with any particular 
data.  Not that FreeBSD doesn't have critical mass, there's just no 
indication as to why Paul believes this.  FreeBSD has certainly made 
its name a long way before the recent Netcraft news that some 2.5 
million web hosts now run FreeBSD, to which Paul may or may not be 
referring.  But maybe Paul hasn't yet entered the world of seach 
engines so hasn't heard of Yahoo! or the extra specialized and covert 
existence of Pair and Verio, all of whom are dedicated FreeBSD users.  
What a wonder it will be when Paul finds that the internet is actually 
searchable with indexed searches.

Maybe with that new search engine capability, Paul could find the 
FreeBSD web site, mysteriously parked at freebsd.org and provide some 
clues to the the 4.x and 5.x branch issue to which he refers 
repeatedly.

Fortunately, as I had bookmarked "FreeBSD.org" from a previous search 
in my web browser, this author was able to quickly bring up the site.

In a large off-yellowish box in the right margin with red lettering, 
the site states:

	New Technology Release: 5.2.1

then a little lower down:

	Production Release: 4.10

That should make things a bit clearer, my friend.  And if you really 
want to get into the nuts and bolts of FreeBSD, you might even want to 
delve into the "Early Adopter's Guide" under that 5.2.1 section, which 
clearly states that:

	. . . the Release Engineering Team <re at FreeBSD.org> specifically 
	discourages users from updating from older FreeBSD releases to 
5.2.1-	RELEASE unless they are aware of (and prepared to deal with) 
possible 	regressions in the newer releases. Specifically, for more 
conservative users, we 	recommend running 4.X releases (such as 
4.9-RELEASE) for the near-term 	future. We feel that such users are 
probably best served by upgrading to 5.X 	only after a 5-STABLE 
development branch has been created; this may be 	around the time of 
5.3-RELEASE.

This seems pretty clear to me, but then again I am also familiar with 
the old refrain of RTFM, something as important with building a 
production server as it is with writing a software review.

NetBSD

If FreeBSD developers are angered by Paul's review of their project, I 
could only wonder how the NetBSD and OpenBSD developers feel.  Maybe 
they are marching to Paul's home as I write.

First, Paul seems to judge operating system by development by 
self-assigned version numbers, which is pretty funny in itself.  He 
states that after NetBSD forked in 1993, it has "continued on its own 
in parallel to FreeBSD. . . . albeit at a slower pace.  Does that mean 
Microsoft Office 2004 is a thousand units better than 2003?  Or that 
while NetBSD is currently at 1.6.2, not 2.6.2 as Paul states, it is 
behind OpenBSD 3.6 by 2.0.something?  How developers label their 
software branches and how they mark their development can not be 
compared by the version numbers of other projects.  Certainly the 
enormous portability of NetBSD to dozens of architectures and the 
brilliance of pkgsrc accounts for some significant pace of development.

Now comes the real gems of the review: "It is not secure" and "you can 
safely ignore NetBSD unless you have old or obscure hardware or are 
looking to port your operating system project to new hardware."  Woah.  
Where do we start?

To say NetBSD is "not secure" is pretty scary.  Paul, turn that web 
browser window to that new site Yahoo.com, and look for this cool list 
called BugTraq.  For your information, it lists vulnerabilities 
released to the security community, and they are usually mentioned 
there first.  And like the rest of the BSD community, NetBSD developers 
don't hide their vulnerabilities, they publicize them and fix them as 
soon as possible.  They also tend to be the ones who publicize them 
first.

And sure, NetBSD runs great on obscure hardware, but to "safely ignore 
NetBSD" would be a huge mistake.  Although it may not be the biggest 
mistake Paul makes.

There is real excitement about pkgsrc from a number of sysadmins I 
know.  A cross platform package system with over 4000 ports and 
counting is hardly something to ignore.  Maybe Paul should browse these 
before arguing that "(i)ts desktop and production applications are so 
limited to be nonexistennt" And then of course you can't forget to 
mention embedded systems, an area in which NetBSD is a serious force.

OpenBSD

If NetBSD developers are marching on Paul's home right now, they might 
not be able to find it, OpenBSD developers already got there and have 
burned it down.

You can look at Paul's basic factual errors first, such as "the site 
claims, there hasn't been a hole in the default install in over seven 
years."  Actually, if Paul read the technical news circuit online, he 
would know that OpenBSD celebrated its eighth birthday recently, and 
the site was updated to account for that.

Or his foolish mistake of stating that "OpenBSD runs on very few 
platforms", when it actually runs on twelve, and the new OpenBSD 3.6 
does run SMP.  Or that "OpenBSD is updated every three or four months" 
when the OpenBSD developers have religiously stuck to six month 
intervals in their release schedule since the beginning.  Most people 
who run OpenBSD know that, without major apocalyptic happenings in the 
universe, a new OpenBSD is released every May 1st and November 1st.

But what is really shocking about Paul's review is his personal attack 
on Theo de Raadt, someone who "is known to be highly unstable and even 
destructive at times. . . " and who has "infamous tantrums" and "many 
users have been flamed off the Internet due to his bad moods and 
compulsive control issues."  Finally, Paul warns us that "developers 
may wish to remain wary of this platform and its creator."

Woah.  Pretty strong words there tiger.  Paul?  Are you still there?

Software reviews are not meant to be personality profiles.  Does Theo 
have a reputation as a compulsively serious developer?  Absolutely.  Is 
this based on being "highly unstable"?  Unless Paul has known Theo for 
any length of time, and was his shrink, or at least has witnessed this 
behavior in person, Paul really should keep his personal inferences out 
of his software reviews.

Theo is not known for being an easy-going person.  He is very much 
dedicated to OpenBSD, and intends to keep focus on maintaining an 
operating system that doesn't cower to foolish assumptions about 
security or "revised" licensing schemes.

And the oft-repeated comments about "if you can't code it in C, please 
don't bother us" is the basis of the project, not pleading for new 
users without any technical background looking for a cool pc in their 
dorm room to impress their friends with.

Paul, if you are going to do a software review, stick to focusing on 
the technology of the operating system, not on the personalities behind 
that OS.

Darwin

I must honestly state that I no longer have the patience to read Paul's 
Darwin section.  It's nice that he does mention that Darwin is part of 
the BSD family, but I'm beginning to think that Paul leaving out 
DragonFlyBSD from his BSD review could only make Matt Dillon smile.

And Then Paul's Final Thoughts. . .

It's fair to believe that there was a lag of months between Paul's 
review of each BSD and his "Final Thoughts" wrap-up.  He espouses how 
great each is as a "workstation solution, network security. . ."  Wait, 
maybe the Slashdot poster who put up the text was referring to another 
article.