Yeah, sure (yawn). And red hat is going bankrupt... for how many years now? So when it is goint to be completely screwed over, hmm? Next year? Two years from now? Five? RH just proves the opposite of what you were implying - it is possible to make money selling a GPLd product, by offering what free competitors will not: professional grade support. Now we can argue whether or not paying a geek to do all the support work for a CentOS farm is would be cheaper or not - probably it will. But PHBs don't care or understand that: they care about reputation, marketing, words like "enterprise grade support"... something (reputation for instance) RH or Novell has. RH want be screwed over because of CentOS. If it would be screwed oever at all, it will be by the competition offering better services/support options (better marketing) - and that would be another company selling their Enterprise line of products (Novell, Mandrake, you name it).
Well, that was one of the most surprising posts I have ever seen on slashdot... thanks:) Open mindedness was not something I expected from the average slashdotter, and you proved me wrong:)
If you want to give FreeBSD a spin, wait a few days for 5.4-RELEASE. As to pf, I have written this a few months ago - a quicky about how to set up pf. If you don't need anything complex (like applying different queue algorythms for traffic shaping from different hosts behind the firewall) you just kldload pf, enable it in rc.conf and write a simple rule file like the one a described there:) Also, you don't need natd to do nat like with ipfw/ipfw2, it is as simple as:
# macros, lists ext_if="vr0" # replace with actual external interface name i.e., dc0 int_if="ed0" internal_net="192.168.0.1/16" external_addr="172.17.141.160"
# options set block-policy drop set loginterface ed0
# Scrub[1] options [2] scrub on ed0 all reassemble tcp
#nat nat on $ext_if from $internal_net to any -> ($ext_if)
pass quick on lo0 all
#default policy block all
# allow outgoing and return traffic pass out quick on $ext_if proto tcp all modulate state pass out quick on $ext_if proto { udp, icmp } all keep state
# filter rule to allow traffic through nat pass in quick on $int_if from $internal_net to any keep state pass out quick on $int_if from any to $internal_net keep state
#example for a service - samba (the most complex, a webserver is just a one-liner pass in on $ext_if proto udp to port { 137, 138 } keep state pass out on $ext_if proto udp to port { 137, 138 } keep state pass in on $ext_if proto tcp to port 139 modulate state
[1]"Scrubbing" is the normalization of packets so there are no ambiguities in interpretation by the ultimate destination of the packet. The scrub directive also reassembles fragmented packets, protecting some operating systems from some forms of attack, and drops TCP packets that have invalid flag combinations.
[2]reassemble tcp
Statefully normalizes TCP connections. When using scrub reassemble tcp, a direction (in/out) may not be specified. The following normalizations are performed:
Neither side of the connection is allowed to reduce their IP TTL. This is done to protect against an attacker sending a packet such that it reaches the firewall, affects the held state information for the connection, and expires before reaching the destination host. The TTL of all packets is raised to the highest value seen for the connection.
Modulate RFC1323 timestamps in TCP packet headers with a random number. This can prevent an observer from deducing the uptime of the host or from guessing how many hosts are behind a NAT gateway. (from pf's faq I linked to on openbsd.org)
That's a damn lie if there is any - FreeBSD supports both IPFW and OpenBSD's PF - if fact 50% of FreeBSD uses PF on FreeBSD. Now don't tell me PF is not up to the task of iptables - it is, and it eats it for breakfest. MANGLE, IP_POSTROUTING, string-matching. It does all that and more - does iptables have anything like CARP. See what's missing from PF here. Debian might have the largest cesspool of tools. Can I install xorg 6.8.2 and KDE 3.4 on it while still having official security updates to my OS? Because I can in FreeBSD. Do I have timely openoffice.org 2.0 binary packages for debian?
Not only that, but name one distro (perhaps gentoo) that has binaries for all localized versions as well. Can I apt get the latest hungarian openoffice2 built for debian?
FreeBSD might be the second to debian that has the largest package repository/ports system there is after debian (and how many of the 16000+ packages are actually useful?). Oh, I guess you didn't like what I wrote about apt a few posts above? If that is the case, instead of spreading FUD, it would be more mature to debate whatever claims I made.
Then I was wrong in my previous post (above yours). Thanks for the info. Screenshot of jerkiness of bottom part of screen - playing the 1920x816 batman trailer (using -vo gl2 to be able to capture the screen).
hmmmm... mplayer plays it, it is only that the bottom part of the screen has some artifects (mplayer: latest version, on an athlon xp 2200+ with 512MB ram, FreeBSD) - and of course it is slow! As to the big news part: it is not big news, except that it is APPLE. Some info mplayer displays:
MOV track #1: 858 chunks, 6765 samples Audio bits: 16 chans: 2 rate: 48000 Audio extra header: len=91 fcc=0x77617665 MOV: Found MPEG4 audio Elementary Stream Descriptor atom (51)! Fourcc: mp4a
and
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family Selected video codec: [ffh264] vfm:ffmpeg (FFmpeg H.264)
So the news part is that Quicktime7 is here, and that's it: there is nothing progressive in making high resolution movies or using h.264 (ffmpeg had it for some time).
Valószínüleg a RAM mennyiség a kulcs a Firefox-hoz. Hozzáteszem, nem borzasztóan lassú egy jobb gépen minimum 256 mega rammal. Viszont látványosan lassabb akár az operánál, akár a konqueror-nál (föleg nagyobb táblázatok renderelésénél - tudom, mert van pont egy ilyen oldalam", és mivel egy ideig kizárólag firefoxot használtam, azt hittem hogy a szerver lassú. Szóval mint örült elkezdtem mindenféle apache trükköt bekonfigurálni (mint az accept filte, míg rá nem jöttem hogy csak a firefox küzd vele annyira. No mindegy, azért még persze használható, csak azért kritizálom (most már többször is) mert oda kéne erre figyelniük. A két gépen amire utaltam korábban win98 futott, és a felhasználók csak arra emlékeznek, hogy IE-re nem kellett fél percet várni míg elindul. (A fél perc nem túlzás, és két különbözö gépröl van szó.) Opera viszont nem probléma (kivéve hogy több a hibásan megjelenített oldal, bár még nem frissítettem a 8-as verzióra).
A FreeBSD-röl: nekem problémamentes. Elsösorban azért ez, mert nagyságrendekkel könnyebb volt megtanulni mint a linuxot. Kezdve a tüzfal konfigurációtól (ami angolul van - tényleg, megírni még egy bonyolultabb filtert is olyan, mintha angolul beszélnél) az oprencer konfigurációig. Aztán meg ott van a csomagkezelöje, mely ötvözi az APT és a ports funkcionalitását. pkg_add -r openoffice-2.0-devel az egyenlö az apt-get install openoffice-blah paranccsal. Szóval nem muszály ports-t használni, lehet teljesen a bináris csomagkezelöre hagyatkozni, amely funkcionalitását tekintve minden szempontból lefedi a debian csomagkezelöjét. A bináris csomakog relatíve frissek (kábé 1 hónap csúszás van a ports-hoz képest, de ezen még akarnak a jövöben javítani): xorg 6.8.2, kde 3.4, openoffice-2 (még a magyar verziója is!!!) stb.. Aztán ott van a webes infrastuktúrájuk. Lehetöség van feltölteni az installált portjaidat a a freshports-ra, és emailben értesítenek ha valamelyik csomagot frissítették.
Aztán: nem kell skipfirst meg ilyesmi. Mikor felteszel több portot, ha az egyik nem sikerül, egyszerüen továbblép, tehát nagyobb lelki nyugalommal hagyom ott hétvégére a gépet fordítani, mert mire megérkezem, ott lesz egy lista hogy miket installált, mik nem sikerültek, és hogy miért nem sikerültek. Aztán: a már felinstallált (mindegy hogy forrásból vagy binárisból) csomagokból egy paranccsal tudsz binárist csinálni: pkg_create -b pkg_name. Ez pedig pont olyan lesz mint egy.deb csomag: tartalmazza a csomag telepítéséhez szükséges dependenciák információit, tehát amikor egy másik gépre fel akarod tenni, automatikusan felteszi a dependenciát vagy az általad megadott útvonalról, vagy a netröl. Így telepítettem a két ratyi gépet is. A jobbikon telepítettem a csomagokat, -p kapcsolóval (mondjuk portinstall -p blackbox) - ami felteszi a csomagot és csinál egy bináris csomagot is egyben. Majd átmásoltam ftp-én keresztül a binárisokat a gyeng
Well, all depends on your motivation - do you want to learn how a unix-like OS works or not? Some want just a quick replacement for windows. You did not learn linux because of gentoo, you learned it because you had the motivation and curiousity to learn it:)
There are pros and cons against both portage and apt. Portage last time I tried it had no proper reverse dependency lookup, but used something like a worldfile as a workaround. As a result, sometimes it was really hard to completely remove a package (all its files and installed libs). APT doesn't have that problem, however, dependencies are fixed. Portage handles dependencies more flexibly, that's a +. However, it does it in an unnecessarily complex way: useflags. Without configuration portage is brain dead (installing xfree as a dependency of mc?). However, to configure it properly, you have to know the interdependencies of 300+ useflags. I think a much better approach is the original ports, where you have to remember one thing: if you don't want to accept the defaults (which are more sane than with portage), you can check the makefile for additional options. Not only that, but before a port is built, you can choose additional dependencies from a menu. Your choice will be saved, and next time you upgrade, it will be remembered (ie. no user interaction is needed) See this for example.
What I don't really understand is why they didn't clone the damn thing (I mean ports) instead of inventing their hodge-podge of a ports system. They would have the best of both worlds (source based and debians apt). I mean the package management of freebsd doesn't care about the origin of the package. In fact, you can create binary packages from an installed port with one simple command: pkg_create -b pkgname. And that's not a simple binary - it has all the functionality of a.deb package: it knows of the dependencies, and if you move it to another machine, it would fetch its dependencies from either the place you specify or from the net. pkg_add -r mplayer has exactly the same functionality that apt-get install mplayer has. Same thing with deinstallation.
The linux distribution of my dreams would be slackware with ports (the original one, without modification)! I would love that!
Hmmm... You should send that to funroll-loops.org. Althought it is not as impressive as this guys setup [grin]:
omg optimized"There are so many steps involved in tweaking the last uumph out of your linux system- and it really is a work of art to pull it off- I have used many different kernels and all sorts of optimization combinations-yesterday I finally used -noatime and -notail for my reiserfs file system: The single biggest performance boost I have yet to see-now I can have gnome2.2 running using gnome-terminal to compile the latest j2sdk from source (nice -n 19)while browsing with mozilla while running e17 in a seperate login with two eterms and run Unreal Tournament at full speed (this with an apache webserver running for my dyndns pseudo-domain and a mysql for my answering-machine software for my isdn card-which keeps track of all incomming phonecalls and manages my telephone book app and ntfsd/sshd/dhcp server/squid)..."
Let me give you an advice: try out opera. On a 333Mhz system Opera runs faster (yes RUNS - not only starts up, it renders pages faster) even if you have a kernel building in the background than Firefox runs on the same machine while it is idle.
And because it is notorius for its userbase. Well, some of it of course, but unfortunately they are very vocal.
I remember this guy (one of the optimization freaks) that tried to convince folks that you need to rebuild the toolchain three times to get it properly optimized - and there was no way to convince him that what he sais is nonsense.
Then I was reading freebsd-question mailing list, and there was this fellow who wanted to install freebsd the gentoo way. People were genuinly bewildered - not because they had anything against gentoo, but because they didn't know how it works, so they had him explaining it. He basically said that he wants everything built from source optimized and all from the ground up. They told him to install the OS and then rebuild everything. He said no no no, he wants to build the kernel before installing, he wants to build the toolchain before installing and so on. The idea was so... well, how shall I put it... strange, that poor folks at first thought that they don't understand the bloke properly. And if you come to think of it: the guy instead of wanting to have a sytem with basic functionality (xfree, a wm, a browser, networking, email, whatnot) up and running in ten minutes, and then rebuilding everything from source in the background (while posting gentoo roxorz messages on./) he wanted to have an unusable system for hours if not half a day - for what reason?
I was always stuck by the sheer senselessness of the install procedure (even starting from stage3) - it seems to me that gentoo has it backwards: instead of installing the damn thing and have the ability to read your emails in 10 minutes while having the choice to optimize when it suits you, you are (or you were, I'm not aware of the current status) forced to fumble with the system for hours just to get it installed.
Don't tell me this helps newbies understand linux better. It doesn't. I had this friend who came from a brief mandrake background to gentoo (I recommended it, for he wanted linux, and I thought gentoo must be cool because of portage - since then I changed my opinion seeing the deficiencies of portage like skipfirst kinda hacks, worldfile instead of proper reverse dependency lookup, useflags and its interdependency hell, etc.). He spent a day installing gentoo (I know, I know, it takes you only a few hours, but for a complete noob, it was a day) than almost a week configuring it (and rebuilding kde when it turned out that nspluginviewer is missing because he forgot a useflag). At the end of it, he was no closer to understanding how linux (or generally a unix-like operating system) works. He just followed the documentation, and succeeded, b/c the doc. is good. I had to explain the most basic things (like file permissions) after he had a more or less (it must have been another useflag that prevented kpdf from opening pdf files) working system up and running!
The problem with this is that I think this senseless install procedure might give some users the (false) sense that, yes, I did it, I built a linux manually - I must be geek. These users, as they build and rebuild their system from day to day, trying out more and more compilation options will grow into those arrogant fanboys who can't resist posting a "yeah baby, I will emerge whatever today, gentoo is so cool" kinda messages in all news regardless of how related or unrelated that is to the topic at hand. Oh yeah, and they will they you to rebuild your tool chain at least 3 times to make sure its properly optimized.
Yes, but that's not focused enough, with no formalized test criteria. I'm thinking of a distro that is not very useful beyond testing the kernel, attracting a userbase that is more capable of sending useful bugreports than your average ubuntu user. For instance, you have to enable debugging support, you must be able to break into the debugger - if all else fails via a serial line, provide stack traces, etc... If a small but detacated (payed?) userbase would send in high quality bug reports, cooperating with each other, having a single goal in their mind, than it would be less strain on the kernel maintainers. For instance, a kernel release that goes through such a testing phase might have a few bugs, lets say 13. With such a focused team with its own mailing list there would be exactly 13 bug reports all with everything the developer needs to quickly solve the problem. Otherwise, there might be hundreds of bug reports from users, some of which might be along the lines of "KDE stopped working, wtf?", and just wading through them to find a useful one can be time consuming.
A formalized testing structure, using most of the available testing tools (ranging from mysql benchmarks to compiling and running KDE) with clear descriptions would be more useful I think. Participating in Ubuntu's QA/Release process is not the same, for the kernel is just one thing among many that it tests - so it is not focused primarily on one task.
Aha, I see now. So it is up to distro makers to give extensive testing for each kernel release. Hmmm... I wonder how a middleman distro would work out. I mean a ditro used by enthusiasts just to build the latest and greatest linux kernel, providing clean (vanilla packages) userland and system layout. Slackware comes to mind - but with more simplicity - fewer packages available - and supporting more archs. Each major distro maker can collaborate in using that middleman distribution and testing the kernel there before inclusion in their own distributions. So kernels could be released first into that distro and retain RC status for a short testing period there before released to the public as stable.
I know this sound more complicated, but it might have its own advantages. For instance, for those who run customized distributions on their servers this might mean fewer gotchas - as it stands now, one cannot be absolutely sure which particular release will be actually stable and fully ready for production use. 2 weeks in that middleman distro, with active stress-testing done by major distributions (or anyone who volunteers) is not much, but it would help I think. Especially since users of such a distro would provide useful bugreports (stack traces, good debug info in general). Then it would be possible to draw up a charter for pronouncing the kernel stable. For instance, it might have to pass various test conditions (make -j12 kernel while running a webserver serving content that would attract huge number of slashdotters - half-naked sexy programmer girls comes to mind) before achieving the stable tag.
Hmmm, I think I agree with you, but I still have an impression that we debate something;) - I just don't know what:) (maybe because english is not my native language).
But again, this does not mean that FreeBSD is bad or poorly organized or useless. It's a fine OS that I recommend to people all the time. It's just that there's a different audience.
Yes, I know, I mean I didn't take any of your comments to have something agains fbsd. I'm just curious about the linux release process. I know bsd's, that's why I mentioned it:) - but I don't know if linux has something similar or not, or how does it look like. Or is it a good or a bad thing? (I mean my perceived absence of transparent rules for releasing a new version, ABI/API stability/change policies, etc.). Whether FreeBSD is good or not is not important in this context:)))
About scalability of support for different archs - I can't argue with that (not that I can argue with whatever you said:)))
FreeBSD does what it can with the resources it has, and that's a good thing. Let's not try to compare them to Linux. Linux is Linux and BSD is BSD. They are excellent tools for different jobs.
I absolutely agree - I'm not comparing the tools, but the way they organize resources. The freebsd project might be something like a good example here nothing more. This news gave me the impression that there is something 'wrong' in linux kernel release/testing process, but frankly, I'm not even sure that there is. Even though it might seem to be a long time since 2.6 was declared stable, I guess it is too early to tell where current procedure leads. It might even be something positive. Just musing over a few things:))
Well, that would really be a problem, but despite Theo's personality (which I think might have its own charms) doesn't necessarily get in the way of development. Just think of the huge contributions OpenBSD made. Common Address Redundancy Protocol (CARP) for instance. Or their excellent firewall, pf (now present in all BSDs). Not to mention OpenSSH. And beside these standalone or highly portable applications, they released a secure and stable OS. Not 'just' a kernel. They write their own libc. They maintain a lot of software in their base system. Apache 1.3.x can be almost considered a fork, with their security/stability related patchset. Which comes down to my main point: The problem is not lack of resources, monetary or otherwise
Currently there are ~100 developers payed fulltime just to work on the kernel (at various organizations). There are none in FreeBSD. There are perhaps a dozen devs whose employers let them work on FreeBSD part-time, or there are various works that are sponsored by companies (pair network comes to mind) from time-to-time. But all in all, FreeBSD, that writes its own kernel, its own C library, and generally speaking maintains an OS (userland apps like their package management and ports system for instance, burncd - the native cd burning app of freebsd, etc.) does that with 1/50 of the resource Linux & co has just to develop the kernel.
This is not about linux vs. freebsd btw. I chose to use the latter, you chose the former, I really don't care, and I'm not willing to engage in yet another linux vs. bsd flamefest. You can argue endlessly about why linux is better, and I can do the same about FreeBSD, but I think we can agree on one point: either way, neither is that much better (lets cut down that figure to 10x - you can't possibly claim that linux is 10x better or something). In other words, my point is that it is not about (monetary) resources. It is a problem of organization imho. Less frequent releases, more API/ABI stability, a controlled release engeneering process might be a solution. Perhaps a branch split like it was done during 2.5.x (current 2.6) development. Pronounce the current 2.6.x branch STABLE, meaning introducing a POLA (policy of least astonishment in freebsd) and forbid API/ABI changes, then continue development in a new, 2.7 branch at the current pace.
I don't mean to imply that there is no release engeneering in linux kernel development whatsoever. But somehow FreeBSD's (and I assume the other BSD's as well) release engeneering seems to me a lot more transparent. Click the first few links at the top of this page to see what I mean by "controlled release engeneering process."
Force it on the entire world? Last time I checked, it was still possible to make Nautilus use "Windows File Browser" mode, and the gnome developers hadn't rendered the dozens of other Windows-esque file managers available for X inoperable.
A few days ago I read a review on a news portal (index.hu - it's hungarian) about suse linux. This is not a technology portal, it is more like cnn or bbc - politics, culture etc. The title of the review caught my attention, it was something like a SuSe Linux review - it is working!. And it did, mostly (reviewer had problems on a laptop, but desktop puter was fine). He mentioned one gripe though - spatial browsing. He didn't name it like that, he has no idea about the gnome lingo, he just simply didn't like the fact that each folder he opened took up desktop space, and he criticized the file manager for cluttering his desktop. Yes, I know that there is a key combo that closes previous window. He didn't know that, he just didn't understand why it works like that.
So the problem with spatial nautilus is that it was made the default, which imho flies in the face of their precious HIG. Whenever it comes to debates like KDE vs. GNOME, it is always about hig this and hig that, but when it comes to implementing (or not implementing) features, often it is in violation of the HIG. Not the GNOME hig, which I didn't read (except for the first chapter). A generic HIG that has something like a POLA (policy of least astonishment) - in it (which it should). Spatial browsing as the default flies in the face of such policy. So does the reverse button order (and don't start me on its justification, I think someone is always having a good laugh when it is mentioned). The way I see it, lately gnome devs blindly copy features of OS X, because we all know that OS X is a powerful yet user friendly desktop. However, each time I sit down (not many times unfortunately, I worked on Macs extensively years ago - 7.5.x times) I have to readjust to the button order. I guess most PC users will have the same problem when trying out either GNOME or OS X. However, in the case of Macs, it was always like that, it is a tradition. In the case of GNOME - well, its what I said: blind copying for no good reason. GNOME devs cannot expect a great number of newcomers from the Mac - if one can afford a mac (and now with mac mini more and more ppl could) why would he or she want to switch? However, they can expect newcomers from Windows - and these newcomers are forced to get accustomed to this order, and during that period they will make accidental mistakes (I did a lot when trying out GNOME - which I do each year out of curiousity). This is again an example of violating HIG.
God forbid those of us who think the Windows browser model is a horrible User Interface design should have an actual, viable option to choose.
Exactly! I cannot agree more. But making that choice the default for everyone is a mistake imho - and it somehow undermines the prestige of their own HIG. I very much doubt that some of the choices they made is based on viable research - spatial browsing is just one example for that. When the GNOME desktop is presented to a newbie, usually the one thing that stands out as something uncool is spatial browsing. Note that I'm talking about newbies here, not users who are accustomed to using various desktop environments. Having a good interface design is only part of the picture - another part which is hugely underestimated is familiarity. This goes hand in hand with drawing up a picture of The Generic User which would usually be your computer-agnostic grandma. This is wrong. Users learn new things more readily than gnome devs would readily admit. Documentation and good tutorials are the key here, and windows leads in that area. It is also a wrong concept that users prize simplicity above all. Yes, you heard it right. This simplicity fetish is simply wrong - being an admin of a small computer lab, the extent users go to make their desktop unusable (in MY opinion) is unbelievable.
That's what is great about the new OpenOffice.org format. (trying to build it right now, fingers crossed). According to oo.o, it is not only supported by the community, but also the European Commission as well:
Beginning with version 2.0 OpenOffice.org uses the open standard OASIS OpenDocument XML format as the default file format. The OASIS OpenDocument format is a vendor and implementation independent file format, and thus guarantees freedom and independence.
In addition to OpenOffice.org itself, the open source office suite KOffice as well as OpenOffice.org derivatives like the StarOffice software support the OASIS OpenDocument file format. The OASIS OpenDocument file format is also one of the file formats recommended by the European Commision.
oo.o-2.0 feature-guide
Fileextensions:
OpenDocument Text [.odt]
OpenDocument Text [.odt]
OpenDocument Text Template [.ott]
OpenDocument Master Document [.odm]
OpenDocument Spreadsheet [.ods]
OpenDocument Spreadsheet Template [.ots]
OpenDocument Drawing [.odg]
OpenDocument Presentation [.odp]
OpenDocument Chart [.odc]
OpenDocument Database [.odb]
I think that this standarization might help in persuading governments to choose this new format. Although not an office suite strictly speaking, I wonder about abiword's default file-format... Does/will it use this new standard as the default as well (seems to be a good idea).
It is because of your comment. Now that you have drawn attention to the fact that this is the least commented story, readers are tentative (myself included) to post new comments for fear of ruining this new record;)
"This version of Apache is principally a bug fix release.
This release is compatible with modules compiled for 2.0.42 and later versions.
We consider this release to be the best version of Apache available and encourage users of all prior versions to upgrade."
Just finished upgrading:)
I agree - with one minor correction. In the case of Firefox, being free software and all, I can only demand one thing: correctness in information. Even though I still have the get firefox logo on each of the sites I maintain, I have a failure story to report. Today, the 4th user came to me demanding back their IE icon (I have disabled access in windows xp to IE, which amounts to disabling access to its icon) - why? Because they could not access one site or another in Firefox, that worked in IE.
Also, don't get me started on performance. 3 machines in the lab range between 300Mhz celerons with 96MB ram and an IBM Personal Computer 300 (600Mhz celeron with 96M ram). Firefox on those is a no-no. Not only b/c painfully slow startup times, but also, painfully slow rendering of pages. Opera renders pages faster while running a kernel compile in the background than Firefox does on an idle computer. What's there in gecko that makes it so much slower than Opera or khtml? (Yes, you heard it right, starting Konqi from a foreign - Blackbox - wm is actually much faster both in startup and rendering of pages than firefox).
These slow machines function as simple 'terminals' btw - they have opera, gaim, xmms, rox - that can be choosen from a simplified menu.
This must be said at the risk of loosing karma (I have plenty, so go ahead) - there is something wrong with Firefox and its rendering engine, not only in compatibility or correct rendering of pages, but in performance as well. And this is not a minor issue, the performance difference b/w say opera or khtml and gecko is significant. So I have only one demand: inform the potential users correctly, don't give them the false impression that Firefox is better in every way than IE. It is not, and such misinformation will only create a backlash. 2 of those users are now actively looking for more and more justfications to have IE back as the standard browser. They are not interested in philosophy or open source ideals. They are interested in accessing the sites they want.
What part of their specs are they lying about? Wondering what to look for.
Every part:) Response time for instance: if you see 12ms, that doesn't apply to the whole color spectrum. It might be true for grey to grey, but red to blue for instance might be 16 or 20. You never know how they calculate the given response time (average? fastest? greytogrey?). Or think of the 178degree viewing angle - lol - why not 180? Oh, you would only see the edge of the monitor... just imagine how much you can see of the picture from 89 degree. Basically every spec:) That's why it is important to know the specs given by the panel manufacturer... not that they are not lying or something... they are, but at least you can see how various vendors using exactly the same panel like to make some hilarious claims.
True. Question is: does it really matter? You won't see the difference anyway:) Otherwise, usually the very fast (sub 16ms) tn+film LCD screens are 6 bit. To the best of my knowledge, all MVA/PVA/S-IPS panels are 8bit.
You agree? Well, for how many years are you planning to use your monitor? If I assume at least 3, than you can calculate how much the energy consumed by your CRT will cost you. My vx912 LCD: ~35W A 20" CRT (the equivalent of 19" LCD): ~120W. Do the math.:)
Slashdot Mirror
Are you sure it is the OS to blame? This might come as a surprise then. 6 win2k servers in the longest uptimes top 100 - according to netcraft.
Yeah, sure (yawn). And red hat is going bankrupt ... for how many years now? So when it is goint to be completely screwed over, hmm? Next year? Two years from now? Five? RH just proves the opposite of what you were implying - it is possible to make money selling a GPLd product, by offering what free competitors will not: professional grade support. Now we can argue whether or not paying a geek to do all the support work for a CentOS farm is would be cheaper or not - probably it will. But PHBs don't care or understand that: they care about reputation, marketing, words like "enterprise grade support" ... something (reputation for instance) RH or Novell has. RH want be screwed over because of CentOS. If it would be screwed oever at all, it will be by the competition offering better services/support options (better marketing) - and that would be another company selling their Enterprise line of products (Novell, Mandrake, you name it).
Can someone give us a link to an explanation of why it was blacklisted? What specific problems they had with it?
If I looked at your profile, I would have spared the lecture on pf.conf :)))
If you want to give FreeBSD a spin, wait a few days for 5.4-RELEASE. As to pf, I have written this a few months ago - a quicky about how to set up pf. If you don't need anything complex (like applying different queue algorythms for traffic shaping from different hosts behind the firewall) you just kldload pf, enable it in rc.conf and write a simple rule file like the one a described there :) Also, you don't need natd to do nat like with ipfw/ipfw2, it is as simple as:
[1]"Scrubbing" is the normalization of packets so there are no ambiguities in interpretation by the ultimate destination of the packet. The scrub directive also reassembles fragmented packets, protecting some operating systems from some forms of attack, and drops TCP packets that have invalid flag combinations.[2]reassemble tcp
Statefully normalizes TCP connections. When using scrub reassemble tcp, a direction (in/out) may not be specified. The following normalizations are performed:
- Neither side of the connection is allowed to reduce their IP TTL. This is done to protect against an attacker sending a packet such that it reaches the firewall, affects the held state information for the connection, and expires before reaching the destination host. The TTL of all packets is raised to the highest value seen for the connection.
- Modulate RFC1323 timestamps in TCP packet headers with a random number. This can prevent an observer from deducing the uptime of the host or from guessing how many hosts are behind a NAT gateway. (from pf's faq I linked to on openbsd.org)
Good Luck!I didn't make that clear, sorry. I downloaded the batman trailer (the 1920x816 big one).
Then I was wrong in my previous post (above yours). Thanks for the info. Screenshot of jerkiness of bottom part of screen - playing the 1920x816 batman trailer (using -vo gl2 to be able to capture the screen).
A FreeBSD-röl: nekem problémamentes. Elsösorban azért ez, mert nagyságrendekkel könnyebb volt megtanulni mint a linuxot. Kezdve a tüzfal konfigurációtól (ami angolul van - tényleg, megírni még egy bonyolultabb filtert is olyan, mintha angolul beszélnél) az oprencer konfigurációig. Aztán meg ott van a csomagkezelöje, mely ötvözi az APT és a ports funkcionalitását. pkg_add -r openoffice-2.0-devel az egyenlö az apt-get install openoffice-blah paranccsal. Szóval nem muszály ports-t használni, lehet teljesen a bináris csomagkezelöre hagyatkozni, amely funkcionalitását tekintve minden szempontból lefedi a debian csomagkezelöjét. A bináris csomakog relatíve frissek (kábé 1 hónap csúszás van a ports-hoz képest, de ezen még akarnak a jövöben javítani): xorg 6.8.2, kde 3.4, openoffice-2 (még a magyar verziója is!!!) stb.. Aztán ott van a webes infrastuktúrájuk. Lehetöség van feltölteni az installált portjaidat a a freshports-ra, és emailben értesítenek ha valamelyik csomagot frissítették.
Aztán: nem kell skipfirst meg ilyesmi. Mikor felteszel több portot, ha az egyik nem sikerül, egyszerüen továbblép, tehát nagyobb lelki nyugalommal hagyom ott hétvégére a gépet fordítani, mert mire megérkezem, ott lesz egy lista hogy miket installált, mik nem sikerültek, és hogy miért nem sikerültek. Aztán: a már felinstallált (mindegy hogy forrásból vagy binárisból) csomagokból egy paranccsal tudsz binárist csinálni: pkg_create -b pkg_name. Ez pedig pont olyan lesz mint egy .deb csomag: tartalmazza a csomag telepítéséhez szükséges dependenciák információit, tehát amikor egy másik gépre fel akarod tenni, automatikusan felteszi a dependenciát vagy az általad megadott útvonalról, vagy a netröl. Így telepítettem a két ratyi gépet is. A jobbikon telepítettem a csomagokat, -p kapcsolóval (mondjuk portinstall -p blackbox) - ami felteszi a csomagot és csinál egy bináris csomagot is egyben. Majd átmásoltam ftp-én keresztül a binárisokat a gyeng
There are pros and cons against both portage and apt. Portage last time I tried it had no proper reverse dependency lookup, but used something like a worldfile as a workaround. As a result, sometimes it was really hard to completely remove a package (all its files and installed libs). APT doesn't have that problem, however, dependencies are fixed. Portage handles dependencies more flexibly, that's a +. However, it does it in an unnecessarily complex way: useflags. Without configuration portage is brain dead (installing xfree as a dependency of mc?). However, to configure it properly, you have to know the interdependencies of 300+ useflags. I think a much better approach is the original ports, where you have to remember one thing: if you don't want to accept the defaults (which are more sane than with portage), you can check the makefile for additional options. Not only that, but before a port is built, you can choose additional dependencies from a menu. Your choice will be saved, and next time you upgrade, it will be remembered (ie. no user interaction is needed) See this for example.
What I don't really understand is why they didn't clone the damn thing (I mean ports) instead of inventing their hodge-podge of a ports system. They would have the best of both worlds (source based and debians apt). I mean the package management of freebsd doesn't care about the origin of the package. In fact, you can create binary packages from an installed port with one simple command: pkg_create -b pkgname. And that's not a simple binary - it has all the functionality of a .deb package: it knows of the dependencies, and if you move it to another machine, it would fetch its dependencies from either the place you specify or from the net. pkg_add -r mplayer has exactly the same functionality that apt-get install mplayer has. Same thing with deinstallation.
The linux distribution of my dreams would be slackware with ports (the original one, without modification)! I would love that!
Let me give you an advice: try out opera. On a 333Mhz system Opera runs faster (yes RUNS - not only starts up, it renders pages faster) even if you have a kernel building in the background than Firefox runs on the same machine while it is idle.
I remember this guy (one of the optimization freaks) that tried to convince folks that you need to rebuild the toolchain three times to get it properly optimized - and there was no way to convince him that what he sais is nonsense.
Then I was reading freebsd-question mailing list, and there was this fellow who wanted to install freebsd the gentoo way. People were genuinly bewildered - not because they had anything against gentoo, but because they didn't know how it works, so they had him explaining it. He basically said that he wants everything built from source optimized and all from the ground up. They told him to install the OS and then rebuild everything. He said no no no, he wants to build the kernel before installing, he wants to build the toolchain before installing and so on. The idea was so ... well, how shall I put it ... strange, that poor folks at first thought that they don't understand the bloke properly. And if you come to think of it: the guy instead of wanting to have a sytem with basic functionality (xfree, a wm, a browser, networking, email, whatnot) up and running in ten minutes, and then rebuilding everything from source in the background (while posting gentoo roxorz messages on ./) he wanted to have an unusable system for hours if not half a day - for what reason?
I was always stuck by the sheer senselessness of the install procedure (even starting from stage3) - it seems to me that gentoo has it backwards: instead of installing the damn thing and have the ability to read your emails in 10 minutes while having the choice to optimize when it suits you, you are (or you were, I'm not aware of the current status) forced to fumble with the system for hours just to get it installed.
Don't tell me this helps newbies understand linux better. It doesn't. I had this friend who came from a brief mandrake background to gentoo (I recommended it, for he wanted linux, and I thought gentoo must be cool because of portage - since then I changed my opinion seeing the deficiencies of portage like skipfirst kinda hacks, worldfile instead of proper reverse dependency lookup, useflags and its interdependency hell, etc.). He spent a day installing gentoo (I know, I know, it takes you only a few hours, but for a complete noob, it was a day) than almost a week configuring it (and rebuilding kde when it turned out that nspluginviewer is missing because he forgot a useflag). At the end of it, he was no closer to understanding how linux (or generally a unix-like operating system) works. He just followed the documentation, and succeeded, b/c the doc. is good. I had to explain the most basic things (like file permissions) after he had a more or less (it must have been another useflag that prevented kpdf from opening pdf files) working system up and running!
The problem with this is that I think this senseless install procedure might give some users the (false) sense that, yes, I did it, I built a linux manually - I must be geek. These users, as they build and rebuild their system from day to day, trying out more and more compilation options will grow into those arrogant fanboys who can't resist posting a "yeah baby, I will emerge whatever today, gentoo is so cool" kinda messages in all news regardless of how related or unrelated that is to the topic at hand. Oh yeah, and they will they you to rebuild your tool chain at least 3 times to make sure its properly optimized.
A formalized testing structure, using most of the available testing tools (ranging from mysql benchmarks to compiling and running KDE) with clear descriptions would be more useful I think. Participating in Ubuntu's QA/Release process is not the same, for the kernel is just one thing among many that it tests - so it is not focused primarily on one task.
I know this sound more complicated, but it might have its own advantages. For instance, for those who run customized distributions on their servers this might mean fewer gotchas - as it stands now, one cannot be absolutely sure which particular release will be actually stable and fully ready for production use. 2 weeks in that middleman distro, with active stress-testing done by major distributions (or anyone who volunteers) is not much, but it would help I think. Especially since users of such a distro would provide useful bugreports (stack traces, good debug info in general). Then it would be possible to draw up a charter for pronouncing the kernel stable. For instance, it might have to pass various test conditions (make -j12 kernel while running a webserver serving content that would attract huge number of slashdotters - half-naked sexy programmer girls comes to mind) before achieving the stable tag.
But again, this does not mean that FreeBSD is bad or poorly organized or useless. It's a fine OS that I recommend to people all the time. It's just that there's a different audience.
Yes, I know, I mean I didn't take any of your comments to have something agains fbsd. I'm just curious about the linux release process. I know bsd's, that's why I mentioned it :) - but I don't know if linux has something similar or not, or how does it look like. Or is it a good or a bad thing? (I mean my perceived absence of transparent rules for releasing a new version, ABI/API stability/change policies, etc.). Whether FreeBSD is good or not is not important in this context :)))
About scalability of support for different archs - I can't argue with that (not that I can argue with whatever you said :)))
I absolutely agree - I'm not comparing the tools, but the way they organize resources. The freebsd project might be something like a good example here nothing more. This news gave me the impression that there is something 'wrong' in linux kernel release/testing process, but frankly, I'm not even sure that there is. Even though it might seem to be a long time since 2.6 was declared stable, I guess it is too early to tell where current procedure leads. It might even be something positive. Just musing over a few things :))
Well, that would really be a problem, but despite Theo's personality (which I think might have its own charms) doesn't necessarily get in the way of development. Just think of the huge contributions OpenBSD made. Common Address Redundancy Protocol (CARP) for instance. Or their excellent firewall, pf (now present in all BSDs). Not to mention OpenSSH. And beside these standalone or highly portable applications, they released a secure and stable OS. Not 'just' a kernel. They write their own libc. They maintain a lot of software in their base system. Apache 1.3.x can be almost considered a fork, with their security/stability related patchset. Which comes down to my main point: The problem is not lack of resources, monetary or otherwise
Currently there are ~100 developers payed fulltime just to work on the kernel (at various organizations). There are none in FreeBSD. There are perhaps a dozen devs whose employers let them work on FreeBSD part-time, or there are various works that are sponsored by companies (pair network comes to mind) from time-to-time. But all in all, FreeBSD, that writes its own kernel, its own C library, and generally speaking maintains an OS (userland apps like their package management and ports system for instance, burncd - the native cd burning app of freebsd, etc.) does that with 1/50 of the resource Linux & co has just to develop the kernel.
This is not about linux vs. freebsd btw. I chose to use the latter, you chose the former, I really don't care, and I'm not willing to engage in yet another linux vs. bsd flamefest. You can argue endlessly about why linux is better, and I can do the same about FreeBSD, but I think we can agree on one point: either way, neither is that much better (lets cut down that figure to 10x - you can't possibly claim that linux is 10x better or something). In other words, my point is that it is not about (monetary) resources. It is a problem of organization imho. Less frequent releases, more API/ABI stability, a controlled release engeneering process might be a solution. Perhaps a branch split like it was done during 2.5.x (current 2.6) development. Pronounce the current 2.6.x branch STABLE, meaning introducing a POLA (policy of least astonishment in freebsd) and forbid API/ABI changes, then continue development in a new, 2.7 branch at the current pace.
I don't mean to imply that there is no release engeneering in linux kernel development whatsoever. But somehow FreeBSD's (and I assume the other BSD's as well) release engeneering seems to me a lot more transparent. Click the first few links at the top of this page to see what I mean by "controlled release engeneering process."
A few days ago I read a review on a news portal (index.hu - it's hungarian) about suse linux. This is not a technology portal, it is more like cnn or bbc - politics, culture etc. The title of the review caught my attention, it was something like a SuSe Linux review - it is working!. And it did, mostly (reviewer had problems on a laptop, but desktop puter was fine). He mentioned one gripe though - spatial browsing. He didn't name it like that, he has no idea about the gnome lingo, he just simply didn't like the fact that each folder he opened took up desktop space, and he criticized the file manager for cluttering his desktop. Yes, I know that there is a key combo that closes previous window. He didn't know that, he just didn't understand why it works like that.
So the problem with spatial nautilus is that it was made the default, which imho flies in the face of their precious HIG. Whenever it comes to debates like KDE vs. GNOME, it is always about hig this and hig that, but when it comes to implementing (or not implementing) features, often it is in violation of the HIG. Not the GNOME hig, which I didn't read (except for the first chapter). A generic HIG that has something like a POLA (policy of least astonishment) - in it (which it should). Spatial browsing as the default flies in the face of such policy. So does the reverse button order (and don't start me on its justification, I think someone is always having a good laugh when it is mentioned). The way I see it, lately gnome devs blindly copy features of OS X, because we all know that OS X is a powerful yet user friendly desktop. However, each time I sit down (not many times unfortunately, I worked on Macs extensively years ago - 7.5.x times) I have to readjust to the button order. I guess most PC users will have the same problem when trying out either GNOME or OS X. However, in the case of Macs, it was always like that, it is a tradition. In the case of GNOME - well, its what I said: blind copying for no good reason. GNOME devs cannot expect a great number of newcomers from the Mac - if one can afford a mac (and now with mac mini more and more ppl could) why would he or she want to switch? However, they can expect newcomers from Windows - and these newcomers are forced to get accustomed to this order, and during that period they will make accidental mistakes (I did a lot when trying out GNOME - which I do each year out of curiousity). This is again an example of violating HIG.
God forbid those of us who think the Windows browser model is a horrible User Interface design should have an actual, viable option to choose.
Exactly! I cannot agree more. But making that choice the default for everyone is a mistake imho - and it somehow undermines the prestige of their own HIG. I very much doubt that some of the choices they made is based on viable research - spatial browsing is just one example for that. When the GNOME desktop is presented to a newbie, usually the one thing that stands out as something uncool is spatial browsing. Note that I'm talking about newbies here, not users who are accustomed to using various desktop environments. Having a good interface design is only part of the picture - another part which is hugely underestimated is familiarity. This goes hand in hand with drawing up a picture of The Generic User which would usually be your computer-agnostic grandma. This is wrong. Users learn new things more readily than gnome devs would readily admit. Documentation and good tutorials are the key here, and windows leads in that area. It is also a wrong concept that users prize simplicity above all. Yes, you heard it right. This simplicity fetish is simply wrong - being an admin of a small computer lab, the extent users go to make their desktop unusable (in MY opinion) is unbelievable.
Fileextensions:
I think that this standarization might help in persuading governments to choose this new format. Although not an office suite strictly speaking, I wonder about abiword's default file-format... Does/will it use this new standard as the default as well (seems to be a good idea).
"This version of Apache is principally a bug fix release.
:)
This release is compatible with modules compiled for 2.0.42 and later versions.
We consider this release to be the best version of Apache available and encourage users of all prior versions to upgrade." Just finished upgrading
Also, don't get me started on performance. 3 machines in the lab range between 300Mhz celerons with 96MB ram and an IBM Personal Computer 300 (600Mhz celeron with 96M ram). Firefox on those is a no-no. Not only b/c painfully slow startup times, but also, painfully slow rendering of pages. Opera renders pages faster while running a kernel compile in the background than Firefox does on an idle computer. What's there in gecko that makes it so much slower than Opera or khtml? (Yes, you heard it right, starting Konqi from a foreign - Blackbox - wm is actually much faster both in startup and rendering of pages than firefox).
These slow machines function as simple 'terminals' btw - they have opera, gaim, xmms, rox - that can be choosen from a simplified menu.
This must be said at the risk of loosing karma (I have plenty, so go ahead) - there is something wrong with Firefox and its rendering engine, not only in compatibility or correct rendering of pages, but in performance as well. And this is not a minor issue, the performance difference b/w say opera or khtml and gecko is significant. So I have only one demand: inform the potential users correctly, don't give them the false impression that Firefox is better in every way than IE. It is not, and such misinformation will only create a backlash. 2 of those users are now actively looking for more and more justfications to have IE back as the standard browser. They are not interested in philosophy or open source ideals. They are interested in accessing the sites they want.
Every part :) Response time for instance: if you see 12ms, that doesn't apply to the whole color spectrum. It might be true for grey to grey, but red to blue for instance might be 16 or 20. You never know how they calculate the given response time (average? fastest? greytogrey?). Or think of the 178degree viewing angle - lol - why not 180? Oh, you would only see the edge of the monitor... just imagine how much you can see of the picture from 89 degree. Basically every spec :) That's why it is important to know the specs given by the panel manufacturer ... not that they are not lying or something... they are, but at least you can see how various vendors using exactly the same panel like to make some hilarious claims.
True. Question is: does it really matter? You won't see the difference anyway :) Otherwise, usually the very fast (sub 16ms) tn+film LCD screens are 6 bit. To the best of my knowledge, all MVA/PVA/S-IPS panels are 8bit.
You agree? Well, for how many years are you planning to use your monitor? If I assume at least 3, than you can calculate how much the energy consumed by your CRT will cost you. My vx912 LCD: ~35W A 20" CRT (the equivalent of 19" LCD): ~120W. Do the math. :)