*BSD News Article 32983


Return to BSD News archive

Xref: sserve comp.os.386bsd.misc:2723 comp.os.linux.misc:19781
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msuinfo!agate!howland.reston.ans.net!vixen.cso.uiuc.edu!uxa.cso.uiuc.edu!rkb55989
From: rkb55989@uxa.cso.uiuc.edu (Rafal Boni)
Newsgroups: comp.os.386bsd.misc,comp.os.linux.misc
Subject: Re: I hope this won't ignite a major flame war, but I've got to know!
Date: 19 Jul 1994 17:12:16 GMT
Organization: University of Illinois at Urbana
Lines: 83
Message-ID: <30h1hg$1i4@vixen.cso.uiuc.edu>
References: <1994Jul18.093302.19670@wmichgw> <30e4dr$jno@bigblue.oit.unc.edu> <1994Jul18.212715.18684@dde.dk> <1994Jul19.025153.10080@cs.cornell.edu>
NNTP-Posting-Host: uxa.cso.uiuc.edu

mdw@cs.cornell.edu (Matt Welsh) writes:

>In article <1994Jul18.212715.18684@dde.dk> kim@dde.dk (Kim Andersen) writes:
>>>9) The Linux Documentation Project (LDP) is doing a fantastic job. Anyone
>>>   trying to decide between BSD and Linux would find the Linux's Installation
>>>   and Getting Started Guide a good incentive to follow Linux. I can't
>>>   give Matt, Michael, et all enough kudos for their good work there.
>>The *BSD's have had documentation from the very start, you can even get
>>a book about the os.

>Yet another reason why Linux is better in many people's opinions---there's
>a lot more to be done. FreeBSD is far from a closed-book issue, but the
>nature of development isn't as swiftly-paced nor participatory. From the
>hacker's point of view, Linux has many more frontiers to explore. From
>what I understand, the *BSD projects don't consider experimental development
>vital---hence, the semi-closed development hierarchy.

	I think this observation is rather un-founded.  Having followed 
	NetBSD's development from the first release I see things in a
	different light.  The *BSD camps try to focus on a set of well
	defined features, and keep those features conforming to well-
	known conventions [whether they be formatting conventiions such 
	as the KNF or traditional BSD system-call interfaces, or...].
	Thus, many pieces that may be in the Linux kernel that offer 
	alternative interfaces are looked down upon.

	Someone interested in a experimental project for BSD will usually get
	in touch with the core members who will coordinate the integration of
	that person's work into the tree.  This is how things like SysV 
	compatibility code started out, among other things.... It is sort-of
	expected that the person taking on the project will take some thought
	in comforming to said conventions... Minor style nits, etc. are 
	then fixed up in the integration.

>These are just observations---make of them what you will. The strongest
>categorization of the two projects is that *BSD moves at a somewhat slower,
>and perhaps more careful, pace. On the other hand, Linux is more hackish,
>less organized, more dynamic. Some consider *BSD to be more stable, easier 
>to install, less confusing, and so on. Nevertheless, I have heard the same 
>said in reference to Linux. 

	I think maybe the "chaotic" [not in a bad sense, necessarily] vs.
	"directed" is more an apt description of development.  I won't
	comment on stability or installation, since I'm obviously biased.

	NetBSD [and I'm sure FreeBSD] is more directed in that a core
	group have acking/nak'ing power over changes and hence directs
	the development in some sense.  Linux is more chaotic because
	everyone hacks and hacks and Linus does integration/releasing
	more than direction [note: I use BSD, but I do read the Linux
	development newsgroup and I use Linux on friend's systems].

	Linux is good in that it has an active doc project going on [LDP],
	but BSD has a score of literature concerning it [the deamon book,
	many papers by the developers].  In that sense, I think BSD has more
	in-depth docs, since many of the interfaces presented in the books,
	papers etc. about BSD are being preserved by the core-team oriented
	development style.  LDP is a nice idea, but last a friend of mine 
	attempted to write a device driver for Linux, we used kernel source 
	for SysV to answer the many q's we had about various sematics not
	covered by the kernel hacker's guide.  [Believe it or not, the 
	SysV sources were much clearer in docs in certain places than the
	Linux kernel sources, and filled in many blanks about usage semantics
	the Linux sources left open to guessing].

	When I went hacking around the BSD device driver area, I had real
	published docs on many of the interfaces, and didn't have to resort 
	to peeking any extraneous source...

>It's all been said before. 

	Maybe, but more often it gets said in the "we're better than you"
	context.

>M. Welsh

						--rafal

/--------------------------------------------------------------------------\
|"Blessed are the meek, for they shall inherit   |              Rafal Boni |
| 15% of the earth, which is 100% more than they |         r-boni@uiuc.edu |
| have now..." -Cartoon caption in New Yorker    | My opinions, not UIUC's |
\--------------------------------------------------------------------------/