If you think slashdot is bad, consider that you can at least post here... even it it does ultimately end up at -1.
Not quite. I was recently banned from posting to Slashdot for around a month (every time I attempted to post to a form, Slash would tell me I `wasn't allowed' to do that). I generally post intelligently, my karma sits at a perpetual excellent and has since it was fifty.
The only reason I can think of is because around the time I was banned I made a joke that the Slashdot editors didn't like. A joke that was
The mainstream success of Linux was inevitably going to be based on it being the best solution for a particular kind of job, and perhaps realizing that that quality comes as a result of Open Source licensing. To get the mainstream public to believe in the `ethics' of Free Software (that non-free software is immoral) was never realistically going to happen.
I always thought real techies used the best tool for the job. If Linux is that tool, and that's why the Linuxworld attendees are there, more power to them
Its a good thing that Linux now has more users than the developers. It means the developers were doing something right. Just like the Windows world, there will be seperate, smaller shows that will cater for developers - OSCon here we come. As a system admin and someone who often has to work out the best way to perform a given tak on Linux I like the fact that they're seperate - system admins have a different set of skills and desires than coders do.
Apart from their desktops...
on
Linuxworld Fun
·
· Score: 2
Where most are happily running Internet Explorer on Windows. Check the user agents that appear in the logs of anyone who's been Slashdotted.
You mean incompatiblities between lazy web designers and the web standards?.. Why should the web browser pretend to be something else and bend the standards and allow those designers to continue with the non-compliant code?
Because a web browser is a document viewer. If it can't view documents, it is a failure. If there's a reason for that, and you can get an end user to listen to you, they'll acknowledge what you have to say, then use IE to `fix' the fact that Mozilla doesn't seem to view many documents.
Don't believe me? Try running a profiler on Mozilla sometime and report back the hotspots.
You're pushing the app, the burden is on you. I can't be bothered. All I know is that what I see in front of me with Mozilla 1.0 and 1.1 If the benchmarks don't reflect that then there's something wrong with the benchmark.
But that would mean that they pulled XUL performance stats out of their asses.
Or they used Moz. Athlon 900, 640MB RAM, XP or Red Hat Linux 7.3, it feels slow. Same with many other PCs of a similar spec, even on other PCs where people tell me Moz feels fast (I guess I have have higher standards). I've not done a benchmark because I haven't needed too. This many people wouldn't be saying these things if the app had been fixed.
Does XUL intrinsically look exactly like native widgets? No.
Look is irrelevant. Look and feel. But that's true.
Does the classic theme look very much like native widgets. Absolutely.
Does look on is own matter as much as look and feel? Not quite. Does Moz look and feel like a Native app on Windows, Linux, or OSX? No.
Does the modern theme look like native widgets? No.
Agreed.
Was it planned to look "native"? No!
Yes, this is a bad default.
Modern theme looks the same no matter what platform you are on. If you want consistency of browser UI when using multiple operating systems (as I do), then use Modern.
That's great for all (both?) of you that web browse across multiple platforms regularly. Poor for the vast majority of users that just wanted a web browser on their platform.
I can hear it now. "But it's not as fast as compiled UIs." "It uses more memory." In a couple of years, advances in the rendering engine and the XUL processor (think 'compiler') will narrow the gap so far as to make the gap imperceptible.
Cool. So you admit it currently feels slow? When/if these advances happen, I might like Moz when I use it and even disregard my opinion that Mozilla was supposed to be like this before 1.0. But I can't afford to wait two years. On Linux, Konqueror, on Windows, IE or Opera.
However, what I will mention is software such as QuickTime player, RealOne, MusicMatch Jukebox, and literally anything written in Java. None of these use the MFC toolkit (not the widgets, anyway) nor do they follow the theme of the widgets in WinXP.
That's true, but a lot of people hate those applications too. I'm not batting an eyelid, I'm swtiching to something else, just like I do on Linux when a KDE (or sometimes GTK) frontend to some skinned monstrosity appears.
I'm a computer geek from Melbourne Australia currently spending some time interstate in Sydney in a vacation. Meeting up with a fellow freelance journalist and on our way to a club, our taxi passed a museum advertising its upcoming exhibit, Chinese Dinosaurs.
"Chinese Dinosaurs?" "Chinese Dinosaur Ninjas" "Undead Chinese Dinosaur Ninjas" "John Romero's Undead Chinese Dinosaur Ninjas" (we both burst out laughing)
Sign, nobody's still answered the question. I assume it *is* possible, but nobody's been bothered. Obviously your comparison isn't very fair: you recreated the updatedb index from scratch, that would be unnecessary in the suggested system.
Thats for your post, but I still don't think that there any reason not to give a file a place in more than one location - a heirarchical tree and an alphabetic sort. I can't see it as being too expensive either - and you always exclude/var/tmp if you were worried about that kind of thing...
While the newly started two programs would make updates on this basis to improve the software to a level of Win98 and compatible with Office2000 and Word.
Your post: China: Running StarSuite under Linux
The article didn't say they were runnign StarSuite. I'd have thought the article was talkign about China makign something like Crossover Office. It doesn't matter what kernel, APIs, and windowing system it uses. If it runs MS Office, for many people, its a Windows clone (it obviously has many of the same APIs).
Slocate indexes the contents of the disk. So does the MFT, FAT and whatever the Linux equivalent is called (I'm surprised I don't know this, but I don't).
Would it be possible to organize that information in a manner so that it could be used to find / locat files in a very quick and efficent manner? I guess what I'm looking for is indexes which are updated on writes / unlinks. Would this be possible? Would there be drawbacks, and could they be counteracted? If someone with more skill than I implements this we'd all save a bit of time.
You can run both CMD and bash via OpenSSH on Windows with Cygwin. It works reliably, and there's quite a few useful command line utilities for the newer versions of windows (2000, XP), especially if you grab the resource kits. However, if you have the bandwidth (and hopefully you do) why not run terminal services?
This was inevitable as conflict between Turbolinux (who have recently released a product calleed PowerCockpit) and Caldera (whose former CEO was named Ransom Love) over who has the `sexiest' business threatened the UnitedLinux alliance.
Judging people by the clothes they wear is immature.
Most RPM based distributions have been able to download a package and all their dependencies a while now using a variety of different mechanisms. There's a few good uniques features of dpkg (just as there are rpm feaures) but its easier to implement these features on rpm than to convert most Linux systems to using another packaging format.
The LSB exists to provide the standards. No Linux distribution, not Debian, not Red Hat, not anyone else, has a current LSB complaint distribution. Download the test suites from linuxbase.org and see for yourself.
The `new' document isn't much of a replacement for the old. The new one is talking about embedded OSs, the formar one was talking about servers in general. I think someone got confused...MS still publish their Competing With Linux Partner Guide (or did last month, anyway) which has the same arguments re: TCO as their old comparative guides.
Not to mention, the new guide isn't very fair either:
SMB is integrated into Linux about as much as it is under Windows (the service is called smbd or server in each OS, turn it on, and go). We have multiple clients for network browsing and attaching to shares is handles as natively as NFS is. Server Appliance manufacturers simply don't have to do any Samba programming to make Samba function in an ordinary network - it works by default. Its also repeatedly benchmarked faster than the Windows implementation.
Linux can and is performing Active Directory in real world enterprise environments. Check out Quantum's Guardian 14000 NAS device, which runs a AD Enabled Samba to provide Linux native AD support for its 1.4TB of storage. Although the Samba code used contains beta code from Samba 3, but these aren't cheap boxes and the utmost of reliability is expected from them.
Scalability does not mean the ability to run on massively parallel x86 boxes. Windows 2000 runs on currently one platform. It does not scale to server class hardware beyond IA64. Linux does.
last time I saw a Specweb test, Tux on Linux trouced IIS on Win32, just like Samba on Linux trounced `Server' on Win32.
Again. Microsoft's definition of integrated is flawed. Its possible to build a modular OS where applications communicate with each other using standard protocols - you don't have to turn on everything by default.
Windows File Protection isn't necessary on Linux because Linux doesn't allow Joe User to save a trojan as C:\EXPLORER.EXE. Kudzu handles automatic configuration of hardware and requires less reboots to do so (test: configure a Linux / Windows XP dual boot on one system, pull out the hard disk, put it in another machine, time how long it takes for Kudzu/Windows Plug and Play to fix things)
NTFS is a semi journaling filesystem. Hence chkdisk takes a few minutes to recover the journal on NTFS 5.1, whereas Ext3 does it in a half second.
Red Hat does only have 2 official `certified' RAID for 7.3 drivers but according to the same HCL will support over thirt drives - whose vendors have not used Red Hat for testing, and thus are not `certified'. Likewise, there were only two certified apps for Windows 2000 professional when it was released (Omnipage and another app, IIRC).
Integration is why distributions exist. That is their function. Standards are handled by the Linux Standard Base.
The Web User Interface in Windows 2000 SAK is limited in out of the box functionality, requiring users to log directly into the device (rendering the device prone to significant danger through user tampering) to perform basic functions advertised in Win2000 SAK devices, like send alerts to an email address. As soon as users go out of your web based GUI (and they are required to) your server appliance is no longer an appliance - its a 2GB default install of Windows 2000 which users can and will modify (because the system encourages them to), increasing your support costs.
If you want to program an app for Linux, and don't wish to Open Source your application, simply write your own code. You don't even get this choice with Windows 200 SAK - you MUST write your own code.
That Microsoft would assume that the number of bug announcements for a product is indicative of that products security status illustrates a non existent understanding of basic security principles. All vulnerabilities are not reported, and bugs differ in severity and mitigation. According to ZDNet, Microsoft take a average of more than four times longer than Red Hat to patch a known flaw in their products, leaving MS customers exposed for longer periods of time. Furthermore Microsoft has a habit of not patching vulnerabilities at all - anyone who purchased Exchange 5.0 years ago (which had a known vulnerability allowing spammers to steal bandwidth from companies deploying the product) will know this - Microsoft never fixed the problem - customers had to PAY to upgrade to Exchange 5.5 to do this. As far as I know, the problem persists to this day (I'm not as involved in the Microsoft world as I used to be, having focused on Linux for the last three years).
Last time I saw, Kerberos 5 was supported in Red Hat Linux as a stanard option. Run setup, turn on Kerberos, enter the server details, done. Modern Linux does not authenticate in clear text, this is a falsehood. The MD5 algorithm used in Linux's shadow password file is stronger than the MD4 used in NTLM2 authentication, which has known flaws and is no longer recommended by those who originally created it. NTLM2 is necessary to authenticate Windows clients pre Windows 2000, such as NT4 and Windows 98, the most popular versions of Windows in existence.
Soon after Microsoft's alleged `security refocus' third parties found major vulnerabilities in Microsofts web server and browser platforms that were missed during the one month long `audit'
The GNU Public licensing model also does not contain licensing provisions that require an OEM, and potentially its licensees, to disclose the source code for its intellectual property in a widespread fashion to open source participants. To suggest otherwise is a fabrication. It does offer the option to use Open Source application code within an OEMs products. If this option is exercised, and the resultant application distributed, the company has obligations to also distribute source code of that application. This does not apply to Windows systems because Windows OEMs do not have this option. Linux OEMs may choose as they wish.
We already have a standardized packaging system and I don't think autopackage will change that. Why not create a good apt-get frontend? Or, if you need to actually change somethign within RPM, submit your changes to the RPM maintainer. I wouldn't install autopackage on my system the same way I wouldn't install any other unpackaged app: the effectiveness of *any* management system is linked with its ubiquity.
Why stop there? Most of the files transmitted through p2p can just as easily be sent through the mail on a disk. Why not ban mail?
Because the makers of all the email clients don't promote their systems as a warez network - Napster and Kazaa have in the past. These days Kazaa certainly goes out of its way to deliberately not promote the issue of piracy, but at the same time they refuse to implement even the most simplest of filters over a network *they* control (Fasttrack still has centralized directory services).
Stop being so damned agressive. Why assume the other fellow doesn't know about middle click (as it turns out he does) and is an idiot?
I'd sure like to know what you're using and how you're trying to cut and paste, because (at least in X), 99% of everything responds to the standard select-copy and middle-button-paste
He might be using OpenOffice, the major contender for being `the' Linux office suite.
The quality of hinting and AA in Mozilla makes gecko look worse, not better, because apparently Mozilla doesn't use Freetype the same way QT and GTK2 do - check Bugzilla for more info. If you have a screenshot to prove otherwise, post it.
If we fix a lot of the "problems" with Linux -- for example, radically restructuring the security and filesystem models to be more Windows-like.
In terms of security, that's already done, with minimal impact on users and much benefit for system administrators. Solaris and Trusted BSD have real permission (ACL) support, and so do current Mandrake and future Red Hat releases. This had no effect on yoru ability to work with the system, but rather gives Unix the ganular permissions its always needed.
Re: the filesystem, I don't see to many users complaining once they have the benefits of the FHS explained to them - buy a new hard disk, move the files, and have instant new storage without having to reinstall any new apps, have a single partition (/) with everything you need to recover the base system, be able to mount/home and/var rw while / is mounted ro (great for servers), etc. Every person I've ever spoken to about storage under Linux thinks its great and doesn't want to change it. The main problem is Unix people who don't know the FHS, and therefore can't expalin it to the Windows user who asks. I do agree that the Linux filesystem (which is NOT the Unix filesystem) rocks and the system would suffer if changed.
migrating to non-PostScript-centric applications
Every app I know can generate Postscript, from Tex to OpenOffice to Staroffice to KOffice. I don't see anybody advocating this change.
changing X to be more Windows-like (i.e. no virtual desktop
Every windows user I've spoken to likes the virtual desktops. PS, Windows has these too - you're looking at a different VT when the trusted path is displayed on screen, and you can unlock the others with various user apps.
color depth switch on-the-fly
I don't think anyone wants this anymore. Cards have been able to do 32 bit color all the time in every practical res for quite some time now. However if it was implemented I can't see how this would negatively impact your work.
no X stream but direct drawing instead
Generally DRI is used for things which wouldn't be practical under X. Every app that one would want to run under X (desktop / server apps) still is, and nobody AFAICT want to remove the ability do do this under X.
then Linux won't actually be useful to me anymore.
I hope I've illustrated that for all of the above that: a) That change has already been done and hasn't impacted you. Eg, ACLs. b) That the change wouldn't be bad and won't affect you when it happens. Eg, DRI. c) People don't want many of the changes you've suggested they do. Eg, no Postscript apps, no Virtual Desktops.
Weird HW detection...sometimes after a reboot i have to rmmod sb/sbawe/soundcore/etc by hand and restart them.
That is odd.
To watch divx5 movies, it is not enough to download a codec like with WMP, but you have to recompile your media player, upgrade your ALSA, upgrade your kernel... in fact, this is the reason i ditched linux and returned to 98. I prefer reboots to downloading endless MBs and recompiling for hours and not being sure it will work.
This is what I do on my Red Hat 7.3 box: Find an app which plays DivX: apt-cache search divx
This pops up a list of apps available, such as xine, kxine, or mplayer. To download
apt-get install xine
And a link will appear in my KDE menu under multimedia.
It is slower. End of story. No matter what you say, no matter what benchmarks or other stuff you come up with, qt/gtk widgets are STILL slower than win32 widgets
Agreed. The next-egeneration are compiling their default kernels with various low-latency drivers built in. This has a noticable effect on interactive response time in X/KDE/GNOME. The beta for what will likely be RedHat 8 already does this, and other distros will do so too - nothing needed on this one but to wait. That said, the current performance issues aren't that noticable on my home system.
watching dvd with XINE takes 40% of my CPU while under windows it takes 5%(five)
I haven't noticed, but it sounds reasonable that this is the case. My XP install certainly chugs down a lot (Tbird 900, 640MB) too though. Either way, I watch film fullscreen otherwise I can't concentrate. I do agree the Linux players still have a little catching up to do. Region-freeness makes it just a little bit more preferable to watch movies under Red Hat than XP in my own case.
process spawning is slower (under windows if i run iexplore.exe repeatedly, it pops up new windows at a rate about 5 windows/second. Under linux, the best i could do is 0.5 new windows/sec.
I can replicate that here too, and I don't think its a bad test at all. Again, if you wait till the next gen distros come out, you'll get a noticable speed increase in X for nothing but a download. Z
Lyx owns, blah blah blah, but under windows,
Hehehe. I'm `smart' enough to understand Tex and I hate it too:)
to do word processing/type setting, it is 10 clicks away to write in my native, non-english, language.
There's a lot of work being put into internationalization, but I'm not sure about Openoffice. I agree its an important point.
I don't even want to think what is necessary to actually print.
Kmenu - System - Printer Configuration on my box. Linux does somethings right:).
*PLEASE* tell me what to do to correct them! i am NOT bashing linux! i WANT to use linux! i WANT it to get better!
Slashdot Mirror
Not quite. I was recently banned from posting to Slashdot for around a month (every time I attempted to post to a form, Slash would tell me I `wasn't allowed' to do that). I generally post intelligently, my karma sits at a perpetual excellent and has since it was fifty.
The only reason I can think of is because around the time I was banned I made a joke that the Slashdot editors didn't like. A joke that was
Judge for yourself - and decide how free of censorship Slashdot truly is
The mainstream success of Linux was inevitably going to be based on it being the best solution for a particular kind of job, and perhaps realizing that that quality comes as a result of Open Source licensing. To get the mainstream public to believe in the `ethics' of Free Software (that non-free software is immoral) was never realistically going to happen.
I always thought real techies used the best tool for the job. If Linux is that tool, and that's why the Linuxworld attendees are there, more power to them
Its a good thing that Linux now has more users than the developers. It means the developers were doing something right. Just like the Windows world, there will be seperate, smaller shows that will cater for developers - OSCon here we come. As a system admin and someone who often has to work out the best way to perform a given tak on Linux I like the fact that they're seperate - system admins have a different set of skills and desires than coders do.
Where most are happily running Internet Explorer on Windows. Check the user agents that appear in the logs of anyone who's been Slashdotted.
You mean incompatiblities between lazy web designers and the web standards? .. Why should the web browser pretend
to be something else and bend the standards and allow those designers to continue with the non-compliant code?
Because a web browser is a document viewer. If it can't view documents, it is a failure. If there's a reason for that, and you can get an end user to listen to you, they'll acknowledge what you have to say, then use IE to `fix' the fact that Mozilla doesn't seem to view many documents.
Don't believe me? Try running a profiler on Mozilla sometime and report back the hotspots.
You're pushing the app, the burden is on you. I can't be bothered. All I know is that what I see in front of me with Mozilla 1.0 and 1.1 If the benchmarks don't reflect that then there's something wrong with the benchmark.
But that would mean that they pulled XUL performance stats out of their asses.
Or they used Moz. Athlon 900, 640MB RAM, XP or Red Hat Linux 7.3, it feels slow. Same with many other PCs of a similar spec, even on other PCs where people tell me Moz feels fast (I guess I have have higher standards). I've not done a benchmark because I haven't needed too. This many people wouldn't be saying these things if the app had been fixed.
Does XUL intrinsically look exactly like native widgets? No.
Look is irrelevant. Look and feel. But that's true.
Does the classic theme look very much like native widgets. Absolutely.
Does look on is own matter as much as look and feel? Not quite. Does Moz look and feel like a Native app on Windows, Linux, or OSX? No.
Does the modern theme look like native widgets? No.
Agreed.
Was it planned to look "native"? No!
Yes, this is a bad default.
Modern theme looks the same no matter what platform you are on. If you want consistency of browser UI when using multiple operating systems (as I do), then use Modern.
That's great for all (both?) of you that web browse across multiple platforms regularly. Poor for the vast majority of users that just wanted a web browser on their platform.
I can hear it now. "But it's not as fast as compiled UIs." "It uses more memory." In a couple of years, advances in the rendering engine and the XUL processor (think 'compiler') will narrow the gap so far as to make the gap imperceptible.
Cool. So you admit it currently feels slow? When/if these advances happen, I might like Moz when I use it and even disregard my opinion that Mozilla was supposed to be like this before 1.0. But I can't afford to wait two years. On Linux, Konqueror, on Windows, IE or Opera.
However, what I will mention is software such as QuickTime player, RealOne, MusicMatch Jukebox, and literally anything written in Java. None of these use the MFC toolkit (not the widgets, anyway) nor do they follow the theme of the widgets in WinXP.
That's true, but a lot of people hate those applications too. I'm not batting an eyelid, I'm swtiching to something else, just like I do on Linux when a KDE (or sometimes GTK) frontend to some skinned monstrosity appears.
First, are there application or user experience standards for KDE, Gnome, X, or command line apps?
n /standards/k de/style/basics/
u p/hig/draft_ hig/
t y-Apps-HOWTO/
Why yes, yes there are (strokes beard):
KDE User Interface Guidelines
http://developer.kde.org/documentatio
GNOME 2.0 Human Interface Guidelines
http://developer.gnome.org/projects/g
Designing Integrated High Quality Linux Applications
http://www.tldp.org/HOWTO/HighQuali
This is ludacris
hell no nigga it ain't. This is Ludacris, foo'. Check yo spelling, y`all sucka MCs meant ludicrous.
Ahem...
I'm a computer geek from Melbourne Australia currently spending some time interstate in Sydney in a vacation. Meeting up with a fellow freelance journalist and on our way to a club, our taxi passed a museum advertising its upcoming exhibit, Chinese Dinosaurs.
"Chinese Dinosaurs?"
"Chinese Dinosaur Ninjas"
"Undead Chinese Dinosaur Ninjas"
"John Romero's Undead Chinese Dinosaur Ninjas"
(we both burst out laughing)
Fucking campers :)
Sign, nobody's still answered the question. I assume it *is* possible, but nobody's been bothered. Obviously your comparison isn't very fair: you recreated the updatedb index from scratch, that would be unnecessary in the suggested system.
Thats for your post, but I still don't think that there any reason not to give a file a place in more than one location - a heirarchical tree and an alphabetic sort. I can't see it as being too expensive either - and you always exclude /var/tmp if you were worried about that kind of thing...
While the newly started two programs would make updates on this basis to improve the software to a level of Win98 and compatible with Office2000 and Word.
Your post: China: Running StarSuite under Linux
The article didn't say they were runnign StarSuite. I'd have thought the article was talkign about China makign something like Crossover Office. It doesn't matter what kernel, APIs, and windowing system it uses. If it runs MS Office, for many people, its a Windows clone (it obviously has many of the same APIs).
So yes. Something to see here.
Slocate indexes the contents of the disk. So does the MFT, FAT and whatever the Linux equivalent is called (I'm surprised I don't know this, but I don't).
Would it be possible to organize that information in a manner so that it could be used to find / locat files in a very quick and efficent manner? I guess what I'm looking for is indexes which are updated on writes / unlinks. Would this be possible? Would there be drawbacks, and could they be counteracted? If someone with more skill than I implements this we'd all save a bit of time.
You can run both CMD and bash via OpenSSH on Windows with Cygwin. It works reliably, and there's quite a few useful command line utilities for the newer versions of windows (2000, XP), especially if you grab the resource kits. However, if you have the bandwidth (and hopefully you do) why not run terminal services?
This was inevitable as conflict between Turbolinux (who have recently released a product calleed PowerCockpit) and Caldera (whose former CEO was named Ransom Love) over who has the `sexiest' business threatened the UnitedLinux alliance.
- Being non commercial doesn't prevent lies.
- Judging people by the clothes they wear is immature.
- Most RPM based distributions have been able to download a package and all their dependencies a while now using a variety of different mechanisms. There's a few good uniques features of dpkg (just as there are rpm feaures) but its easier to implement these features on rpm than to convert most Linux systems to using another packaging format.
The LSB exists to provide the standards. No Linux distribution, not Debian, not Red Hat, not anyone else, has a current LSB complaint distribution. Download the test suites from linuxbase.org and see for yourself.F**k advocacy. Use the best tool for the job.
Not to mention, the new guide isn't very fair either:
We already have a standardized packaging system and I don't think autopackage will change that. Why not create a good apt-get frontend? Or, if you need to actually change somethign within RPM, submit your changes to the RPM maintainer. I wouldn't install autopackage on my system the same way I wouldn't install any other unpackaged app: the effectiveness of *any* management system is linked with its ubiquity.
Why stop there? Most of the files transmitted through p2p can just as easily be sent through the mail on a disk. Why not ban mail?
Because the makers of all the email clients don't promote their systems as a warez network - Napster and Kazaa have in the past. These days Kazaa certainly goes out of its way to deliberately not promote the issue of piracy, but at the same time they refuse to implement even the most simplest of filters over a network *they* control (Fasttrack still has centralized directory services).
Stop being so damned agressive. Why assume the other fellow doesn't know about middle click (as it turns out he does) and is an idiot?
I'd sure like to know what you're using and how you're trying to cut and paste, because (at least in X), 99% of everything responds to the standard select-copy and middle-button-paste
He might be using OpenOffice, the major contender for being `the' Linux office suite.
When exactly did ignorance become OK?
Never has. Arrogance either.
The quality of hinting and AA in Mozilla makes gecko look worse, not better, because apparently Mozilla doesn't use Freetype the same way QT and GTK2 do - check Bugzilla for more info. If you have a screenshot to prove otherwise, post it.
If we fix a lot of the "problems" with Linux -- for example, radically restructuring the security and filesystem models to be more Windows-like.
/home and /var rw while / is mounted ro (great for servers), etc. Every person I've ever spoken to about storage under Linux thinks its great and doesn't want to change it. The main problem is Unix people who don't know the FHS, and therefore can't expalin it to the Windows user who asks. I do agree that the Linux filesystem (which is NOT the Unix filesystem) rocks and the system would suffer if changed.
In terms of security, that's already done, with minimal impact on users and much benefit for system administrators. Solaris and Trusted BSD have real permission (ACL) support, and so do current Mandrake and future Red Hat releases. This had no effect on yoru ability to work with the system, but rather gives Unix the ganular permissions its always needed.
Re: the filesystem, I don't see to many users complaining once they have the benefits of the FHS explained to them - buy a new hard disk, move the files, and have instant new storage without having to reinstall any new apps, have a single partition (/) with everything you need to recover the base system, be able to mount
migrating to non-PostScript-centric applications
Every app I know can generate Postscript, from Tex to OpenOffice to Staroffice to KOffice. I don't see anybody advocating this change.
changing X to be more Windows-like (i.e. no virtual desktop
Every windows user I've spoken to likes the virtual desktops. PS, Windows has these too - you're looking at a different VT when the trusted path is displayed on screen, and you can unlock the others with various user apps.
color depth switch on-the-fly
I don't think anyone wants this anymore. Cards have been able to do 32 bit color all the time in every practical res for quite some time now. However if it was implemented I can't see how this would negatively impact your work.
no X stream but direct drawing instead
Generally DRI is used for things which wouldn't be practical under X. Every app that one would want to run under X (desktop / server apps) still is, and nobody AFAICT want to remove the ability do do this under X.
then Linux won't actually be useful to me anymore.
I hope I've illustrated that for all of the above that:
a) That change has already been done and hasn't impacted you. Eg, ACLs.
b) That the change wouldn't be bad and won't affect you when it happens. Eg, DRI.
c) People don't want many of the changes you've suggested they do. Eg, no Postscript apps, no Virtual Desktops.
Weird HW detection...sometimes after a reboot i have to rmmod sb/sbawe/soundcore/etc by hand and restart them.
:)
:).
That is odd.
To watch divx5 movies, it is not enough to download a codec like with WMP, but you have to recompile your media player, upgrade your ALSA, upgrade your kernel... in fact, this is the reason i ditched linux and returned to 98. I prefer reboots to downloading endless MBs and recompiling for hours and not being sure it will work.
This is what I do on my Red Hat 7.3 box:
Find an app which plays DivX:
apt-cache search divx
This pops up a list of apps available, such as xine, kxine, or mplayer. To download
apt-get install xine
And a link will appear in my KDE menu under multimedia.
It is slower. End of story. No matter what you say, no matter what benchmarks or other stuff you come up with, qt/gtk widgets are STILL slower than win32 widgets
Agreed. The next-egeneration are compiling their default kernels with various low-latency drivers built in. This has a noticable effect on interactive response time in X/KDE/GNOME. The beta for what will likely be RedHat 8 already does this, and other distros will do so too - nothing needed on this one but to wait. That said, the current performance issues aren't that noticable on my home system.
watching dvd with XINE takes 40% of my CPU while under windows it takes 5%(five)
I haven't noticed, but it sounds reasonable that this is the case. My XP install certainly chugs down a lot (Tbird 900, 640MB) too though. Either way, I watch film fullscreen otherwise I can't concentrate. I do agree the Linux players still have a little catching up to do. Region-freeness makes it just a little bit more preferable to watch movies under Red Hat than XP in my own case.
process spawning is slower (under windows if i run iexplore.exe repeatedly, it pops up new windows at a rate about 5 windows/second. Under linux, the best i could do is 0.5 new windows/sec.
I can replicate that here too, and I don't think its a bad test at all. Again, if you wait till the next gen distros come out, you'll get a noticable speed increase in X for nothing but a download. Z
Lyx owns, blah blah blah, but under windows,
Hehehe. I'm `smart' enough to understand Tex and I hate it too
to do word processing/type setting, it is 10 clicks away to write in my native, non-english, language.
There's a lot of work being put into internationalization, but I'm not sure about Openoffice. I agree its an important point.
I don't even want to think what is necessary to actually print.
Kmenu - System - Printer Configuration on my box. Linux does somethings right
*PLEASE* tell me what to do to correct them! i am NOT bashing linux! i WANT to use linux! i WANT it to get better!
Same here. Hope that helped.