I mean: am i wrong or those boxes are intended to be servers ? No video card pre-installed, small size, relatively high price, kinda "rackable": they're servers, aimed at competing with IBM servers (well... small servers...).
They're not competitors to Apple's desktop Macs. They're not aimed at desktop use... or am i completely missing the point ?
right ! but's that a global EU problem (what if u're Danish ?): in what languages should u make european-wide documents available ?
for governments, the use is now to have it translated at least in English, French, German, Spanish and Italian, and optionnaly in any other of the 11 languages of the EU... believe me, i've worked on this, and it's a nightmare: translation of legal texts seems to be a complex speciality, and managing all that stuff synchronously is pure hell !
That's probably why Mandrake only put French (they're French after all), and English versions.
ummm.. well.. say u're european, but live in france ?
and besides, there's an interesting sentence at the end of the disclaimer:
"if you are an institutional investor, you hereby guarantee that you are not US, Japanese, Australian or Canadian based."
so, say u're a german institutional investor, wouldn't u be happy to have the version in english (instead of french only) ?
actually AOL has two doors to the internet, one in the USA, and the other in Germany.
but that changes barely nothing to ur point, that is right: AOL is pure crap;-)
anyway, the situation is even worse. Germany and France share one very bad situation: the TLD is used for everything (schools, universities, companies, etc.). The french nic has promoted.asso.fr (=.org) and.gouv.fr (=.gov). but even the french government itself doesn't stick to those domains...
that leads me to my point: in France,.com domains are very popular, because "commercial" in french is said.. "commercial"... but.net is unused, as a contrary (network=reseau in french). So, why would we choose.co or.net or.gov or anything as pre-extensions for TLDs (".co".fr), instead of language-specific extensions (.res.fr instead of.net.fr) ? So, there i think we start the mess....
>Good Point - Forgot about the french subs - but the US and France are allied and france developed the capability against Russia - still wonder why they havent decomissioned them ?
in fact, the french nuclear weapons were re-deployed after the end of the cold war. those aimed specifically against the USSR were decomissioned (ground-ground ICBMs at Plateau d'Albion), and focus was put on nuclear subs, which are aimed at any target. but still, i would really find odd that they would be used against the US, although it's possible...
> France - Mainly short range but do have ICBM capability - You can bet US knows where they all are at any time
I'm not that sure the US knows where french nuclear submarines are at any time. they are smaller and far less noisy than Ohio class equivalents. but have nuclear warhead missiles too (M4-5 ICBMs).
what is ridiculous is that i really wonder why France will launch those missiles against the US. It's senseless !!!
> China - These guys are the only REAL nucleat power left other than the US - ICBM's and well mainitained - huge conventional forces HOWEVER China is the least likely to use the weapons as it doesnt fit with their national Psyche or their view of the world - they are by tradition a defensive nation and thus these weapons are seen as defensive only.
i'm not that sure too about their "defensive" views... especially considering Taiwan. what i am sure is that they are smart enough not to start a war in which they'll loose lotta people. they are far from able to avoid any of the US ICBMs that would be launched against them, nor to destroy them before they are launched.
so i agree ! Nuclear Weapons are the least of the world worries !!
yeah.. very strange.. Led Zep is the only band ever to have put 3 albums in the top 100 hits at the same time in the US (it was in the 70's), and Led Zeppelin IV is still in the top 5 of the most sold albums ever. so i bet Jimmy Page didn't do it for money !!!
btw what's shocking to me there is that there's a huge gap between that vacuum that is Puff Daddy and Led Zep, that basically invented heavy metal rock (some say it's Hendrix, but i'm not that convinced about that..)
actually, 2.4.3 kernel from redhat is available for a while now, as an update (RPMS). That's why i wonder why the GamePC guy didn't use this kernel, which has a lot of improvments. Moreover, i had several problems with my Duron and that 2.4.2, which, while investigating a bit, made me think the AMD architecture wasn't very well supported using pre-compiled RH 2.4.2 kernel. I simply recompiled it de-activating MTRR stuff and so on, and everything worked very well afterwards.
I also think choosing a RH to test AMD processors is far from fair, because they usually focuses on Intel platforms when pre-compiling their ugly kernel.
Lastely, 2.4 kernel series haven't been a heaven till 2.4.6 (remember 2.4.5 ReiserFS bug ?), so why bothering testing on 2.4.2 ? I don't know much about SMP support in 2.2.x, is it that bad ?
>Lets face it, is MSN messenger really an essential service?
Well, i kinda agree and disagree with you:
The fact i really do not understand is that if IM is such a critical issue for you (say, a company), why don't you run your own IM server (said jabber ? or others: i even developped one small IM system for windows in the past -IFY-...) ???
That's the part i really don't understand. It seems odd to me to be confident in external public services (esp. microsoft's one) in core business applications, or for handling confidential data. I think for some part the "magic" of the web, where everything seems easy as compared to old time computing or services make people forget that nothing is magic, and that it doesn't prevent from thinking.
That's also mainly the reason why i think that HailStorm stuff is hugely dangerous.
Right, but it's not very easy to use, buried in the conf menu...
In kde 2.2, u'll have a quick and simple User Agent changer (i didn't see it, but i assume it's more or less like Opera's one).
well.. actually, changing the UserAgent is a feature for Konqueror in KDE 2.2, already available in Beta 1 (if this is the problem of course... but it usually is).
ok, that's only for perl, but that's great, when it works;-)
anyway i don't really see in what port system, apt get, rpm or whatever have to do with.NET. I thought.NET was dealing more with services (passport stuff, and some C# bullshit progs around -maybe other languages too-), than with apps, that are most of the time simply useable on any UNIX-like system (./configure,make,make install works on most recent progs out here, on ever GNU/Linux distro, most *BSD, and probably MacOS X too, and thus usually binaries packages exist as well) so what's the point of this article ? As far as apps are concerned, i don't see why we would use *BSD ports, that seems very odd to me, and instead i think it would be far better if things were cleared up for everyone about where to install all the stuff from a prog (/usr ?/usr/share ?/usr/local ?/var ???) i've seen in the past years that much funny things about that. see for apache, in the past default content directories were under/home/httpd, now it is in/var/www (using RedHat), and it's probably different on another distro. That makes no sense to me. it's silly. I think LSB 1.0 is far better news than that port thing we don't care.... (well, FHS seems to be still working on, but it's about ending with a stable version, despite i don't find clear enough the difference between/usr and/opt)
but maybe i'm only completely missing the point about that non-event...
Actually, i think things aren't exactly going that way.
It's even worse. I've seen (too) many techies fond of MS stuff. I've even seen security experts fond of MS stuff !!! Because MS fog around the exact possibilities of their software, clean and nice windows make those people THINK the product itself is as clean as the UI, and as new and innovative as MS claims it to be. We all know it isn't.
Moreover i've heard some executives techies saying "well, seen what happened to IBM ? thereofore i'll stuck with MS as i'd used to stuck with IBM until someone else clearly becomes the 'industry standard'".
Finally, i would just add one thing: what made MS software succeed at home ? nice UI ? comfort of use ? certainly not, Apple was there before for that. What made MS succeed at home use is that they succeeded at work use: people don't wanna make the effort to "learn" two different "computers", and they wanna be able to bring back home progs and docs from work.
Well, on RH and Mandrake, i use autopdate.
one main advantage is that u can put it in your cron.daily... it's pure commande line, and works well. I think that as far as updates as concerned, this should always work this way.
Re:Konqueror's HTML rendering: too good to be trut
on
Galeon At A Glance
·
· Score: 2
actually they did reuse code from mozilla:
konqueror has two rendering engine from which u can choose:
-kmozilla well, it's gecko
-khtml, which is QT/KDE development, from scratch, and which is really great indeed:-))
in fact only one thing remain really weak in konqueror: javascript. but i've seen that KDE 2.2 should fix that.
anyway, a reason why konqueror (well u're talking about khtml in fact) is such a great thing compared to mozilla, is that it's only a browser, and it doesn't try to reinvent the wheel for each task. it relies deeply on KDE for many things such as rendering images and so on. and they didn't try to make an abstraction bloatware based on GTK/MFC/whatsoever, in order to have it run under every platform on earth...
i don't understand really.
why is AOL shitting that much with linux clients for linux ? they've only got a kinda pre-release for linux, dated from december, which is completely unusable (well.. it doesn't have a correct accentuated characters handling, making it unusable outside the us).. say it doesn't exist.
first they tried to shit OSS people making good work for them (i'm not sure, but isn't it their interest that more and more people use AIM ?) by disallowing OSCAR connections (leaving only that shitty TOC, that only KIT uses;-)), and now that legal stuff... really, all that it's pointless.
i've given up using AIM because of all that fuss they did... and i've switched to yahoo (gtkyahoo and kyahoo are indeed good clients, u just have to choose depending on ur favorite desktop), and i find it great for now... just hope that yahoo won't start to shit 'em as well...
moreover, i think that's a strange way for AOL to thank OSS people for the good job they do for their Mozilla/Netscape project (hey, Netscape6 comes with AIM !), and i wonder what could happen to people using Mozilla name or software.. will they be sued as well ?;;-))
another question: where did AOL register AIM trademark ? did they do it in all european countries as well ? what happens if we start distributing it from here ?
well.. u're missing the chat function (irc/chatzilla).. but right, it was not in Moz 0.7 which is in RH 7.1.
That's what i wonder about; Considering mozilla is now at 0.9.1, with far more features in it, and that there's a chance it keeps on growing, by what time will mozilla be larger than kernel, making it the largest internet-centered OS-weight program we've all dreamt for ?;;-)))
videolan project started in the old days beeing aimed at broadcasting mpeg2 video on the students' campus' LAN in Chatenay.
I remember they got huge funding to buy a complete 100MB switched ethernet network, mostly sponsored by companies such as Cisco (or maybe else..). I saw the very first demos in 1999 or so... they were using a mpeg2 decoder card, and focused on developing the streaming over the LAN, which i remember as beeing quite complex.
I didn't even know they GPL'ed the whole thing and started developing the decoders themselves. nice thing !:-))
Still, i don't really understand how DVD format fit in the initial picture. ability to brodcast DVDs over the LAN ? Isn't that a big copyright and broadcast issue ?
that's not exactly what this guy is saying.
as a summary he says tux can handle dynamic requests by forwarding them to apache (or any other web server).
so, who is right or wrong ?
KDE-ishly speaking, that means they separated the config tool for the equivalent of KCOP from the config tool for the equivalent of kdelibs, am i right ? so, they failed where KDE guys succeeded ? is there any obvious reason to that ? has Bonobo far more features than KCOP ?
sorry to go on with the parallel with KDE, but i'd like to understand what is wrong with GNOME development.. that looks very unclear to me: ok people fight, that happens all the time. but i think also that GNOME development is very slow, and i really wonder why.
actually i asked the same question below (#27), but except one nice guy, people here seem more interested in flamebait than in answering me.. too bad:-(
(i know i'll get modded down for that, but anyway, truth is still a thing one have to fight for;-))
Well at first i have to say i am a happy KDE user. I think the KDE team did a great work on KDE2, and for me it's fast enough and perfectly useable.
Anyway, i was wondering really what all that stuff about GNOME-2.0 means, and i'd wish to have some additional explanation.
As i understood there was two goals in GNOME-2.0:
1- use of GTK+-2.0
2- changes in the architecture (between gnome core and bonobo)
So, what exactly are the roles of bonobo, lignome and libgnome(ui) in gnome, and what exactly was at stake there ?
Because of the way NT is designed, if USER crash, regardless of where USER lives, NT goes down as well. Why? Because without USER, all the information that Win32 has goes south and dies.
This has adverse affects on any program that uses Win32 (99.999999999% of them does).
ok, therefore we agree: not enough separation between UI (User Interface, which meant to me User Space if i get it, again, pardon me if i don't have the right english terms) and the "kernel".
and i agree with all your part about win9x as well
Forth, X does makes direct hardware calls, and as such, capable of making the machine hang.
right as well, but when an app crashes X for some reason, it is therefore less likely to hang the whole machine, as NT does (regarding ur comment about Win32 use, the apps, Explorer.exe and Krnl ARE closer than an app,X and the kernel on Linux, no ?). X is an app, the wm is another app and so on... i just mean there are better chances that the faulty app crashes or locks only the "layer" (which i don't mean as NT's HAL.DLL counterpart.. actually, for me, KDE or GNOME are layers, since they provide more or less their own set of libraries to access hardware, communication between processes and so on...) closest to it than the whole machine, that can happen only every time with NT's design (not to talk about 9x), and therefore there is a chance on linux that you can use another UI (such as text console) to kill and restart X (no need to restart the whole thing).. well if X didn't lock the whole machine, which can still happen, but less often. or am i missing something again ?
Ok, i am wrong, X is the Linux kernel, you are right. I'm gonna tell Linus we've found a new feature of his kernel.
My statement IS NOT an oversimplification, since there are BOTH a LINUX KERNEL MODULE AND AN X DRIVER, which means "X and the Kernel are clearly different". BTW the kernel module is usually generic, whereas the X driver is usually specific. This distinction DOES NOT EXIST on Windows.
And i can go further: with X-systems, the GUI server, window manager, etc. are clearly independant piece of software. That's what "layer programming" (sorry i must not know the exact english term) means, and that's one of the reasons why the UI is heavier but more stable on Linux than on Windows... (the main other beeing the open-source coding)
Of course, this doesn't mean it's impossible to lock the machine with an UI app, but it's far more difficult... The main problem beeing that if you don't have a network access to your machine, the OS can actually be still running and functionnal, but the keyboard beeing locked by X (well... the faulty X app), you can't log on a text console and kill X. I barely never had a case when i had to reboot my linux box because of a faulty app (and believe me, i use crappy apps !).. I think it happens twice a year... on windows it used to be twice a day... Last time i had serious total locks-up problems with Linux, was simply because of faulty hardware...
I didn't meant drivers didn't run in kernel space (i hope they do ! that's the point of a kernel !!), i meant the UI run too close from the kernel on Windows.
In other words, i mean that if you don't have clearly different layers in an OS, this is a design flaw... and don't tell me that kernel32.exe and explorer.exe are two different layers...
BTW 9x isn't single task at all. That is even the problem, because the "DOS-like" design worked, as long as only one task was launched at a time.
And neither is the single user design of 9x a problem, as for end-users. The problem is that it makes no difference between the end user, the apps it launches, and the kernel. NT does, but the UI remains, as i said too close from the kernel, and too "integrated": explorer.exe is a window manager, a file manager, a GUI server (as X), etc.
As for linux, X and the Kernel are clearly different, and it happens to me often having X crashed, whereas my linux box was still working.
Apparently that's not happening when using NVidia drivers (X drivers AND Kernel modules), which makes my point about closed-source drivers.
Slashdot Mirror
I mean: am i wrong or those boxes are intended to be servers ? No video card pre-installed, small size, relatively high price, kinda "rackable": they're servers, aimed at competing with IBM servers (well... small servers...).
They're not competitors to Apple's desktop Macs. They're not aimed at desktop use... or am i completely missing the point ?
right ! but's that a global EU problem (what if u're Danish ?): in what languages should u make european-wide documents available ?
for governments, the use is now to have it translated at least in English, French, German, Spanish and Italian, and optionnaly in any other of the 11 languages of the EU... believe me, i've worked on this, and it's a nightmare: translation of legal texts seems to be a complex speciality, and managing all that stuff synchronously is pure hell !
That's probably why Mandrake only put French (they're French after all), and English versions.
ummm.. well.. say u're european, but live in france ?
and besides, there's an interesting sentence at the end of the disclaimer:
"if you are an institutional investor, you hereby guarantee that you are not US, Japanese, Australian or Canadian based."
so, say u're a german institutional investor, wouldn't u be happy to have the version in english (instead of french only) ?
actually AOL has two doors to the internet, one in the USA, and the other in Germany. ;-)
.asso.fr (=.org) and .gouv.fr (=.gov). but even the french government itself doesn't stick to those domains...
.com domains are very popular, because "commercial" in french is said .. "commercial"... but .net is unused, as a contrary (network=reseau in french). So, why would we choose .co or .net or .gov or anything as pre-extensions for TLDs (".co".fr), instead of language-specific extensions (.res.fr instead of .net.fr) ? So, there i think we start the mess....
but that changes barely nothing to ur point, that is right: AOL is pure crap
anyway, the situation is even worse. Germany and France share one very bad situation: the TLD is used for everything (schools, universities, companies, etc.). The french nic has promoted
that leads me to my point: in France,
>Good Point - Forgot about the french subs - but the US and France are allied and france developed the capability against Russia - still wonder why they havent decomissioned them ?
in fact, the french nuclear weapons were re-deployed after the end of the cold war. those aimed specifically against the USSR were decomissioned (ground-ground ICBMs at Plateau d'Albion), and focus was put on nuclear subs, which are aimed at any target. but still, i would really find odd that they would be used against the US, although it's possible...
> France - Mainly short range but do have ICBM capability - You can bet US knows where they all are at any time
I'm not that sure the US knows where french nuclear submarines are at any time. they are smaller and far less noisy than Ohio class equivalents. but have nuclear warhead missiles too (M4-5 ICBMs).
what is ridiculous is that i really wonder why France will launch those missiles against the US. It's senseless !!!
> China - These guys are the only REAL nucleat power left other than the US - ICBM's and well mainitained - huge conventional forces HOWEVER China is the least likely to use the weapons as it doesnt fit with their national Psyche or their view of the world - they are by tradition a defensive nation and thus these weapons are seen as defensive only.
i'm not that sure too about their "defensive" views... especially considering Taiwan. what i am sure is that they are smart enough not to start a war in which they'll loose lotta people. they are far from able to avoid any of the US ICBMs that would be launched against them, nor to destroy them before they are launched.
so i agree ! Nuclear Weapons are the least of the world worries !!
yeah.. very strange.. Led Zep is the only band ever to have put 3 albums in the top 100 hits at the same time in the US (it was in the 70's), and Led Zeppelin IV is still in the top 5 of the most sold albums ever. so i bet Jimmy Page didn't do it for money !!!
btw what's shocking to me there is that there's a huge gap between that vacuum that is Puff Daddy and Led Zep, that basically invented heavy metal rock (some say it's Hendrix, but i'm not that convinced about that..)
actually, 2.4.3 kernel from redhat is available for a while now, as an update (RPMS). That's why i wonder why the GamePC guy didn't use this kernel, which has a lot of improvments. Moreover, i had several problems with my Duron and that 2.4.2, which, while investigating a bit, made me think the AMD architecture wasn't very well supported using pre-compiled RH 2.4.2 kernel. I simply recompiled it de-activating MTRR stuff and so on, and everything worked very well afterwards.
I also think choosing a RH to test AMD processors is far from fair, because they usually focuses on Intel platforms when pre-compiling their ugly kernel.
Lastely, 2.4 kernel series haven't been a heaven till 2.4.6 (remember 2.4.5 ReiserFS bug ?), so why bothering testing on 2.4.2 ? I don't know much about SMP support in 2.2.x, is it that bad ?
>Lets face it, is MSN messenger really an essential service?
Well, i kinda agree and disagree with you:
The fact i really do not understand is that if IM is such a critical issue for you (say, a company), why don't you run your own IM server (said jabber ? or others: i even developped one small IM system for windows in the past -IFY-...) ???
That's the part i really don't understand. It seems odd to me to be confident in external public services (esp. microsoft's one) in core business applications, or for handling confidential data. I think for some part the "magic" of the web, where everything seems easy as compared to old time computing or services make people forget that nothing is magic, and that it doesn't prevent from thinking.
That's also mainly the reason why i think that HailStorm stuff is hugely dangerous.
Right, but it's not very easy to use, buried in the conf menu...
In kde 2.2, u'll have a quick and simple User Agent changer (i didn't see it, but i assume it's more or less like Opera's one).
well.. actually, changing the UserAgent is a feature for Konqueror in KDE 2.2, already available in Beta 1 (if this is the problem of course... but it usually is).
and for perl:
;-)
.NET. I thought .NET was dealing more with services (passport stuff, and some C# bullshit progs around -maybe other languages too-), than with apps, that are most of the time simply useable on any UNIX-like system (./configure,make,make install works on most recent progs out here, on ever GNU/Linux distro, most *BSD, and probably MacOS X too, and thus usually binaries packages exist as well) so what's the point of this article ? As far as apps are concerned, i don't see why we would use *BSD ports, that seems very odd to me, and instead i think it would be far better if things were cleared up for everyone about where to install all the stuff from a prog (/usr ? /usr/share ? /usr/local ? /var ???) i've seen in the past years that much funny things about that. see for apache, in the past default content directories were under /home/httpd, now it is in /var/www (using RedHat), and it's probably different on another distro. That makes no sense to me. it's silly. I think LSB 1.0 is far better news than that port thing we don't care.... (well, FHS seems to be still working on, but it's about ending with a stable version, despite i don't find clear enough the difference between /usr and /opt)
perl -MCPAN -e install :
suck source accross the net, resolve dependencies, buils, installs...
ok, that's only for perl, but that's great, when it works
anyway i don't really see in what port system, apt get, rpm or whatever have to do with
but maybe i'm only completely missing the point about that non-event...
Actually, i think things aren't exactly going that way.
It's even worse. I've seen (too) many techies fond of MS stuff. I've even seen security experts fond of MS stuff !!! Because MS fog around the exact possibilities of their software, clean and nice windows make those people THINK the product itself is as clean as the UI, and as new and innovative as MS claims it to be. We all know it isn't.
Moreover i've heard some executives techies saying "well, seen what happened to IBM ? thereofore i'll stuck with MS as i'd used to stuck with IBM until someone else clearly becomes the 'industry standard'".
Finally, i would just add one thing: what made MS software succeed at home ? nice UI ? comfort of use ? certainly not, Apple was there before for that. What made MS succeed at home use is that they succeeded at work use: people don't wanna make the effort to "learn" two different "computers", and they wanna be able to bring back home progs and docs from work.
Well, on RH and Mandrake, i use autopdate.
one main advantage is that u can put it in your cron.daily... it's pure commande line, and works well. I think that as far as updates as concerned, this should always work this way.
actually they did reuse code from mozilla: :-))
konqueror has two rendering engine from which u can choose:
-kmozilla well, it's gecko
-khtml, which is QT/KDE development, from scratch, and which is really great indeed
in fact only one thing remain really weak in konqueror: javascript. but i've seen that KDE 2.2 should fix that.
anyway, a reason why konqueror (well u're talking about khtml in fact) is such a great thing compared to mozilla, is that it's only a browser, and it doesn't try to reinvent the wheel for each task. it relies deeply on KDE for many things such as rendering images and so on. and they didn't try to make an abstraction bloatware based on GTK/MFC/whatsoever, in order to have it run under every platform on earth...
i don't understand really. ;-)), and now that legal stuff... really, all that it's pointless.
;;-))
why is AOL shitting that much with linux clients for linux ? they've only got a kinda pre-release for linux, dated from december, which is completely unusable (well.. it doesn't have a correct accentuated characters handling, making it unusable outside the us).. say it doesn't exist.
first they tried to shit OSS people making good work for them (i'm not sure, but isn't it their interest that more and more people use AIM ?) by disallowing OSCAR connections (leaving only that shitty TOC, that only KIT uses
i've given up using AIM because of all that fuss they did... and i've switched to yahoo (gtkyahoo and kyahoo are indeed good clients, u just have to choose depending on ur favorite desktop), and i find it great for now... just hope that yahoo won't start to shit 'em as well...
moreover, i think that's a strange way for AOL to thank OSS people for the good job they do for their Mozilla/Netscape project (hey, Netscape6 comes with AIM !), and i wonder what could happen to people using Mozilla name or software.. will they be sued as well ?
another question: where did AOL register AIM trademark ? did they do it in all european countries as well ? what happens if we start distributing it from here ?
well.. u're missing the chat function (irc/chatzilla).. but right, it was not in Moz 0.7 which is in RH 7.1. That's what i wonder about; Considering mozilla is now at 0.9.1, with far more features in it, and that there's a chance it keeps on growing, by what time will mozilla be larger than kernel, making it the largest internet-centered OS-weight program we've all dreamt for ? ;;-)))
.. it ended up as a DVD-player ?
:-))
videolan project started in the old days beeing aimed at broadcasting mpeg2 video on the students' campus' LAN in Chatenay.
I remember they got huge funding to buy a complete 100MB switched ethernet network, mostly sponsored by companies such as Cisco (or maybe else..). I saw the very first demos in 1999 or so... they were using a mpeg2 decoder card, and focused on developing the streaming over the LAN, which i remember as beeing quite complex.
I didn't even know they GPL'ed the whole thing and started developing the decoders themselves. nice thing !
Still, i don't really understand how DVD format fit in the initial picture. ability to brodcast DVDs over the LAN ? Isn't that a big copyright and broadcast issue ?
that's not exactly what this guy is saying.
as a summary he says tux can handle dynamic requests by forwarding them to apache (or any other web server).
so, who is right or wrong ?
KDE-ishly speaking, that means they separated the config tool for the equivalent of KCOP from the config tool for the equivalent of kdelibs, am i right ? so, they failed where KDE guys succeeded ? is there any obvious reason to that ? has Bonobo far more features than KCOP ?
sorry to go on with the parallel with KDE, but i'd like to understand what is wrong with GNOME development.. that looks very unclear to me: ok people fight, that happens all the time. but i think also that GNOME development is very slow, and i really wonder why.
actually i asked the same question below (#27), but except one nice guy, people here seem more interested in flamebait than in answering me.. too bad :-(
;-))
(i know i'll get modded down for that, but anyway, truth is still a thing one have to fight for
Well at first i have to say i am a happy KDE user. I think the KDE team did a great work on KDE2, and for me it's fast enough and perfectly useable.
/. crowd ;-)
Anyway, i was wondering really what all that stuff about GNOME-2.0 means, and i'd wish to have some additional explanation.
As i understood there was two goals in GNOME-2.0:
1- use of GTK+-2.0
2- changes in the architecture (between gnome core and bonobo)
So, what exactly are the roles of bonobo, lignome and libgnome(ui) in gnome, and what exactly was at stake there ?
Thank you all,
Because of the way NT is designed, if USER crash, regardless of where USER lives, NT goes down as well.
Why? Because without USER, all the information that Win32 has goes south and dies.
This has adverse affects on any program that uses Win32 (99.999999999% of them does).
ok, therefore we agree: not enough separation between UI (User Interface, which meant to me User Space if i get it, again, pardon me if i don't have the right english terms) and the "kernel".
and i agree with all your part about win9x as well
Forth, X does makes direct hardware calls, and as such, capable of making the machine hang.
right as well, but when an app crashes X for some reason, it is therefore less likely to hang the whole machine, as NT does (regarding ur comment about Win32 use, the apps, Explorer.exe and Krnl ARE closer than an app,X and the kernel on Linux, no ?).
X is an app, the wm is another app and so on... i just mean there are better chances that the faulty app crashes or locks only the "layer" (which i don't mean as NT's HAL.DLL counterpart.. actually, for me, KDE or GNOME are layers, since they provide more or less their own set of libraries to access hardware, communication between processes and so on...) closest to it than the whole machine, that can happen only every time with NT's design (not to talk about 9x), and therefore there is a chance on linux that you can use another UI (such as text console) to kill and restart X (no need to restart the whole thing).. well if X didn't lock the whole machine, which can still happen, but less often. or am i missing something again ?
Ok, i am wrong, X is the Linux kernel, you are right. I'm gonna tell Linus we've found a new feature of his kernel.
My statement IS NOT an oversimplification, since there are BOTH a LINUX KERNEL MODULE AND AN X DRIVER, which means "X and the Kernel are clearly different". BTW the kernel module is usually generic, whereas the X driver is usually specific. This distinction DOES NOT EXIST on Windows.
And i can go further: with X-systems, the GUI server, window manager, etc. are clearly independant piece of software. That's what "layer programming" (sorry i must not know the exact english term) means, and that's one of the reasons why the UI is heavier but more stable on Linux than on Windows... (the main other beeing the open-source coding)
Of course, this doesn't mean it's impossible to lock the machine with an UI app, but it's far more difficult... The main problem beeing that if you don't have a network access to your machine, the OS can actually be still running and functionnal, but the keyboard beeing locked by X (well... the faulty X app), you can't log on a text console and kill X. I barely never had a case when i had to reboot my linux box because of a faulty app (and believe me, i use crappy apps !).. I think it happens twice a year... on windows it used to be twice a day... Last time i had serious total locks-up problems with Linux, was simply because of faulty hardware...
I didn't meant drivers didn't run in kernel space (i hope they do ! that's the point of a kernel !!), i meant the UI run too close from the kernel on Windows.
In other words, i mean that if you don't have clearly different layers in an OS, this is a design flaw... and don't tell me that kernel32.exe and explorer.exe are two different layers...
BTW 9x isn't single task at all. That is even the problem, because the "DOS-like" design worked, as long as only one task was launched at a time.
And neither is the single user design of 9x a problem, as for end-users. The problem is that it makes no difference between the end user, the apps it launches, and the kernel. NT does, but the UI remains, as i said too close from the kernel, and too "integrated": explorer.exe is a window manager, a file manager, a GUI server (as X), etc.
As for linux, X and the Kernel are clearly different, and it happens to me often having X crashed, whereas my linux box was still working.
Apparently that's not happening when using NVidia drivers (X drivers AND Kernel modules), which makes my point about closed-source drivers.