Date: Sun, 7 Mar 1999 16:33:27 -0600 (CST) From: Tony Kimball <alk@pobox.com> To: robert@kudra.com Cc: freebsd-hackers@freebsd.org Subject: Re: base64 Message-ID: <14049.29637.311448.247778@avalon.east> References: <14048.10089.598598.919239@avalon.east> <36E07AEC.101F3467@newsguy.com> <14048.48864.918087.631128@avalon.east> <19990306095927.B53145@kudra.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Quoth Robert Sexton on Sat, 6 March: : One point I'd add is that Linux without add ons is pretty much : unusable. Since its a kernel, you need add ons to even log in. : FreeBSD is a coherent OS because its a complete package out of the : box. We have a usable base system. And there's no dispute of which : things are add ons and which aren't. Attempts to expand the base : system have met with what I'd consider justified resistance. I object to the current modularization, precisely because it is either bloated or inadequate, depending on your application. The circulation would only benefit if the system was easier to tailor to the application. This would result from layering the functional areas -- a development package, and admin tool package, a networking package, an X server package, etc. The base system has expanded. DHCP is one recent example. In the modern world, a system without DHCP is much less generally useful than one with DHCP. MIME should be another functional adaptation to the modern environment. Uu-coding can go. (Well, not yet, but before I die.) : Thats why we have the ports tree. If we switched to : elm for instance, that would be a wasted megabye of disk for me, and : many other people. Can you say 'Creeping Featureism?' Like DHCP. Or C++. Creeping Featurism? I don't think so. But it would be better to be able to trivially configure a system without DHCP, or C++, or MIME, I agree. : If I may be so presumptious, I'd say that the coherent vision is that : we have the smallest base distribution which makes up a complete : system, and that the rest can be convieniently flavored to taste. 'Complete system' is not a fungible term. It's a semantic game. One person's complete system in another person's bloated monstrosity or underpowered midget. : > MHO: If the included mail reader is not 'decent', throw it away. Oh, : > replacing UCB mail is socially difficult, but such compunctions seem : > to be damaging the future of the product, its coherent vision, and its : > ability to keep up with the times. : : This is the path that has made linux such a mess. There is _no_way_ : that FreeBSD could even include a mail reader that would make most : people happy. Nor should it try! I'm not arguing for making people happy. I'm arguing for going the minimum lawful speed on the freeway. That speed has gone up in the past few years, on this particular freeway. If you can't drive at the minimum speed, stay off the freeway. Linux qua apple, not orange. You should compare the kernel to Linux, And the distribution to SUSE/RedHat/Debian/Caldera -- all of which are much better candidates for a general end-user *or* server OS than is FreeBSD on the evidence of circulation alone. (FreeBSD will never be given a fair shot unless its circulation rises sufficiently to be a known contender.) FreeBSD 2.1.x is perfect for some people. FreeBSD 2.2.x is perfect for some people. FreeBSD 3.x is perfect for some people. -current is perfect for some people. That's great, but -current is becoming something, something new. That new thing will not be an improvement for those who find 2.2.x to be perfect. I am of the impression that -current exists for a number of distinct reasons, significant among them being 1) a desire to sell CDs on the part of Walnut Creek 2) a desire for control/self-actualization/money on the part of core/committers 3) a desire for a certain product by the end-users with the last being as diverse as the user base. While the 4.x user base will overlap with the 3.x user base, it is (1) hopeful that the subscription base will increase by virtue of the product being more competetive in order to provide a hope of a business future, and (2) hopeful that it will fulfill the needs of its contributors in order to provide a hope of a technical future. : You've touched upon what can only be called a religeous issue. : Based on my take of of the user base, nobody here wants FreeBSD to be : a general purpose platform 'out of the box'. The consensus seems to : be that the minimal system is quite popular. Well, Sam, it certainly is not 'minimal'. PicoBSD is more like minimal. There is a minimal Linux distribution too, the name of which I forget. : I for one don't want to schlep around a lot of mime code that my news : machine will never run. Likewise for uucp :-) Its not just that we've : got to support the old machine crowd, but that bloat is in general : bad. Ugh. Then stick to 2.1.X -- certainly don't run 3.x! But I agree that bloat is bad (that's part of the meaning of the word 'bloat'). The current system is much more bloated than past systems, in the sense that if you add up the number of unused bits in each installation, it is much, much larger. : > I still think FreeBSD should be packaged as a layered system, a group : > of functional complexes serving different areas. And sooner rather : > than later, for as it grows the components of what should be distinct : > modules become increasingly interdependent. : : This the solution to the problem that could make everybody happy. : So hows it coming along, Jordan? :-) Is there actual motion in this direction? To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?14049.29637.311448.247778>