Date: Mon, 12 Mar 2001 12:55:34 +0100 From: Jeroen Ruigrok/Asmodai <asmodai@wxs.nl> To: Luigi Rizzo <luigi@info.iet.unipi.it> Cc: Josef Karthauser <joe@tao.org.uk>, Ian Dowse <iedowse@FreeBSD.ORG>, cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Subject: The Project and onward [was: Re: cvs commit: src/sys/netinet ip_output.c] Message-ID: <20010312125534.B46529@daemon.ninth-circle.org> In-Reply-To: <200103121021.LAA87908@info.iet.unipi.it>; from luigi@info.iet.unipi.it on Mon, Mar 12, 2001 at 11:21:49AM %2B0100 References: <20010312104538.B45619@daemon.ninth-circle.org> <200103121021.LAA87908@info.iet.unipi.it>
next in thread | previous in thread | raw e-mail | index | archive | help
-On [20010312 12:08], Luigi Rizzo (luigi@info.iet.unipi.it) wrote: [gestation period in current] >i said "for some things". I should have added "now". Yes ok. I think that symantics on determining when it exactly started to fail. In other words, falls outside of the scope of the problem at hand. :) >You mention another example where broken code (with insufficient >protection for critical sections) can happily live untested in >-current for ages. As you point out, having two very different versions >of the system (one of which is at the moment hard to use) >is the main reason for the above. Once things will converge, >then the testing on -current will work perfectly >again. At the moment, it just does not in many cases, and >I think it is better that committers realize this, in the interest >of code quality. I agree totally to the above statement/conclusion. I know for a fact that a bunch of committers and developers are actively moving back to STABLE to do development on because CURRENT wasted/wastes their time too often by having them find and solve the problems in there before they could continue their own work. What is done is done, SMPng was necessary, in how much it will affect UP systems in a negative and/or positive way still remains to be seen and will probably be 20/20 hindsight. The question however, which you sort of raised in the above, is that we have to see how we can ease transition of things like this in the future. We have one problem, although John Baldwin et al. have worked very hard on providing a lot of documentation for the current system under which FreeBSD works I see a distinct number of kernel hackers not being able to keep up with the recent changes. Newbus and everything associated is also still under-documented/explained [I should know, since I am but slowly moving on and on with that piece of documentation =( ] and now we have SMPng. I am not sure in how much mbuf [typical BSD/Stevens stuff] has changed symantics or operation in such a way that you can disregard anything mentioned in Stevens' books. And the list continues on and on. I might be wrong about this, but I have a lot of questions still inside my head, which people might assume I should answers to being committer and all that. - We definately lack architectural guidance, things keep getting added and added, but were do we draw the line? - The above probably is a step which needs to be taken after we consider what to target as FreeBSD. I mean, an Unix OS targetted at x86 and Alpha just doesn't fit anymore. I mean, on out frontpage we pride ourselves of being THE choice for Internet related applications/serving, but in practice we're still behind on certain things. To name one small, but very important thing: clustering [typical usage meaning webclustering and the likes in my vocabulary]. So, what do we target? Servers? Workstations? Embedded systems? I know it runs perfect on all, but I cannot seem to get a definate answer somewhere, nor do I doubt we could even agree on one. So mayhap there is a chance on the things we might want to see get done in the future with regard to things implemented. My personal goals still lie in a BSDL compiler suite and many, many cool network extensions, plus a full regression testsuite for a lot of system critical regions so that we can assure more QA. - Who will be part of an architecural team if it is to be formed? - How can we best inform our community about what cool new things FreeBSD is doing? Sure -announce is one good way [and should be used more IMHO]. Furthermore we need more conspectia [I forgot the correct plural again *sigh*] for the appropriate lists [I need to get my stuff automated for the -net list], we need more active FAQ maintainers [hello, that means YOU!], we need experienced webmasters to make our site usable [yes, sorry to say I think it sucks to navigate around], and I am sure people could think of more things to do. I am sure a lot more questions to be formed and made, and I hope they at least seek to try and address issues at the `helicopter view' level, since going into minutae at this point might not work. Reading back this might seem as a rant, but I think it is time we at least got some project management to come into view, aside from core's tasks, since otherwise we might be digging our own hole in certain areas. I welcome peoples ideas, suggestions and what not, and prefer to keep it clean, to the point and flamefree. I have no idea where to ask any follow-ups to be posted to though... -- Jeroen Ruigrok van der Werven/Asmodai .oUo. asmodai@[wxs.nl|freebsd.org] Documentation nutter/C-rated Coder BSD: Technical excellence at its best D78D D0AD 244D 1D12 C9CA 7152 035C 1138 546A B867 Might makes right... To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010312125534.B46529>