Getting Our Beans In A Row To Fight Vista Circa Q4 2006

[Brazen]

What do I think linux needs?

To nab home users:

1. A "killer" game that will only run on linux

2. A commercial, much like IBM's from a few years ago, but focusing on how much better the UI is, how it doesn't crash (like Windows does) and how it resistant OOB to spyware and viruses (unlike Windows, which requires additional expensive software to be even marginally resistant).

For the corporate world:

1. A competitor to Active Desktop. I know SAMBA is working on this, and I hope SAMBA 4 blows me away. I'd like to see something that can have all the control over Windows desktops as Group Policy and can have the same controll over linux desktops. In particular: I have an automated installation of XP, then I add it to the domain and depending on what department it goes to and what software it should have, I just drop it into a certain OU and it gets all the software and settings it needs and I don't have to do anythings else. If I could do this in a SAMBA domain, I would switch.

2. A logon scripting tool as powerful and user-friendly as ScriptLogic. I would be happy to use ScriptLogic, but I'm pretty sure it won't work in a Samba domain.

...

And to the people saying "Why a fight?" I think you are critisizing semantics here. I wouldn't call it a "fight" either, but I know what the OP is meaning. I want linux to succeed because I think the free-as-in-freedom-of-speech model produces better software. I would say 70% of the time, opensource software is better than the proprietary counterpart. And to people who say "quit posting on a forum and do something about it" posting on a forum is not going to hurt anything; I can't code but who knows maybe a coder will be reading through this forum and get some ideas; and I do bug our software vendors about support for opensource servers. My biggest problems with it are:

1. I don't want to put Linux on people's desktops because they wouldn't be able to just pop into Wal-Mart or Best Buy, pick up the latest game or tax software or whatever and put it on their machine. (yes, this will work in a small number of cases due to wine and cedega, but cedega negates the advantage of not having to buy extra software such as antivirus for Windows, and the real solution has to be native ports from the software developers).

2. I don't want to give up Active Directory at work because of the power and ease it has over managing workstations.

3. One example: MySQL. It's better, faster, and easier to manage (IMO) than MS-SQL, but we have to have MS-SQL because we have enterprise software vendors who only support MS-SQL. They don't bother supporting MySQL because they don't need to, to be competitive. If Linux servers with opensource software became the norm, developers would have to support it to be competitive. It's kinda a catch22 in this respect.

 

[Nothinman]

The people who want to use Linux will, regardless of Vista. This is also good. I think you're making much ado about nothing.

Those people will also generally be forced to pay for a copy of Vista whether they plan on using it or not, you consider this good?

 

[Brazen]

Originally posted by: Nothinman

The people who want to use Linux will, regardless of Vista. This is also good. I think you're making much ado about nothing.

Those people will also generally be forced to pay for a copy of Vista whether they plan on using it or not, you consider this good?

Another very very good point. If I go out and buy a new computer (I build my own desktops, but I gotta buy a whole laptop) and I wipe the harddrive the first time I turn it on without ever even booting the Windows partition, and put linux on it, I'm still fueling the monopoly and Microsoft is getting $100 (or something like that) that should be going to the opensource developers.

 

[Nothinman]

2. A commercial, much like IBM's from a few years ago, but focusing on how much better the UI is, how it doesn't crash (like Windows does) and how it resistant OOB to spyware and viruses (unlike Windows, which requires additional expensive software to be even marginally resistant).

Only the last point has any merit, IMO. People are already used to the Windows UI, so switching to anything else will be problematic no matter how much better or worse it is. And in general Windows doesn't crash anymore unless you fill it with crappy drivers and software.

1. A competitor to Active Desktop. I know SAMBA is working on this, and I hope SAMBA 4 blows me away. I'd like to see something that can have all the control over Windows desktops as Group Policy and can have the same controll over linux desktops. In particular: I have an automated installation of XP, then I add it to the domain and depending on what department it goes to and what software it should have, I just drop it into a certain OU and it gets all the software and settings it needs and I don't have to do anythings else.

That's never been done because no one's done any large installations of Linux on their desktops. Most of it could be scripted pretty easily, but there's no packages out there to do it for you yet. Maybe one of those big euro contracts will fund something to handle this.

2. A logon scripting tool as powerful and user-friendly as ScriptLogic. I would be happy to use ScriptLogic, but I'm pretty sure it won't work in a Samba domain.

I have no idea what ScriptLogic is, but Linux has no shortage of scripting languages so I really doubt there needs to be another one introduced.

1. I don't want to put Linux on people's desktops because they wouldn't be able to just pop into Wal-Mart or Best Buy, pick up the latest game or tax software or whatever and put it on their machine.

IMO that's a good thing, most of the stuff you get at BestBuy is crap and causes more problems than it fixes.

3. One example: MySQL. It's better, faster, and easier to manage (IMO) than MS-SQL, but we have to have MS-SQL because we have enterprise software vendors who only support MS-SQL. They don't bother supporting MySQL because they don't need to, to be competitive. If Linux servers with opensource software became the norm, developers would have to support it to be competitive. It's kinda a catch22 in this respect.

Those vendors would be better off supporting standard SQL via ODBC, then you could use whatever backend you wanted. And I wouldn't call MySQL better than MSSQL, PostgresSQL might be close, but not MySQL at least not until 5.0 is released.

 

[Brazen]

Originally posted by: Nothinman
Only the last point has any merit, IMO. People are already used to the Windows UI,....

Not, sure if you're arguing against me or just expanding on my points but I'll give my thoughts on your response...

People are fickle and don't care about facts. If they see commercials telling them it's better they will believe. People were comfortable with Win98 but they upgraded to Win2000 without batting an eye even though it's about (bear with me here) as different as linux is. They did it because it was hyped and Microsoft told them it was better.

That's never been done because no one's done any large installations of Linux on their desktops.

True. But you took my argument the wrong way, and after reading my post I see that's my fault. I was meaning in the corporate world that the easiest thing to do would be to nab the servers first, and the desktops will/may gradually follow, possibly into some sort of mixed win32/*nix environment. I meant to emphasize that the Active Directory competitor on the server side and the logon scripting tool would have control of Windows clients, not Linux clients, not to begin with anyway.

But once that gets established, as long as the same program can have the same (or more) control over linux clients, it will greatly aid the conversion from Windows workstations to Linux workstations. The server side has to happen first though, because Microsoft will never add Linux support to Active Directory. (also I'll put a note below describing ScriptLogic**)

IMO that's a good thing, most of the stuff you get at BestBuy is crap and causes more problems than it fixes

very true, but people still want to buy games, and they still want to buy that personal finance/educational/productivity software to take home and install it but never use it.

Those vendors would be better off supporting standard SQL via ODBC,

Again, you are correct, BUT they don't. And there's nothing I can do about since none of the competitors in this very specialized field are any different and I'm just one person out of their 100 or so clients who complains but buys the software anyway because I have no choice to do what my company wants to do.

-----
** ScriptLogic is BASICALLY just a logon script. It is invoked as a typical logon script, assigned as slogic.bat in the user's profile and the slogic.bat is stored in the scripts folder under the NETLOGON share on the Domain controller. So far, no big deal right? I assume you can assign logon scripts in a SAMBA domain. The problem is, that slogic.bat calls an installation of ScriptLogic components onto the workstation if not already there (pulled from the share I think, still no big deal, but), then that client program interfaces with a ScriptLogic component that must be installed on the server and running as a service. Hence Linux/SAMBA domain controller = no install of server-side service. I assume this server side service is what gives ScriptLogic it's magical powers.

 

[Nothinman]

People are fickle and don't care about facts. If they see commercials telling them it's better they will believe. People were comfortable with Win98 but they upgraded to Win2000 without batting an eye even though it's about (bear with me here) as different as linux is. They did it because it was hyped and Microsoft told them it was better.

People are also lazy, stupid and dislike change. Most people only moved to Win2K and XP when they bought a new machine or when some other requirement forced them. Switching OSes when they already have one that came with the machine and is familiar with them seems unthinkable, if they've even heard of Linux.

I was meaning in the corporate world that the easiest thing to do would be to nab the servers first, and the desktops will/may gradually follow, possibly into some sort of mixed win32/*nix environment. I meant to emphasize that the Active Directory competitor on the server side and the logon scripting tool would have control of Windows clients, not Linux clients, not to begin with anyway.

Eventually that will happen. We have dozens of Linux servers now, mostly in-house custom stuff ported from commercial unix and oracle. And because of that a few (less than 5 IIRC) people did ask to run Linux on one of their machines, but usually only a second machine in their cube because there's still a lot of things that they can't do on Linux because of crap in-house custom tools that require Windows.

That's the biggest hurdle, custom in-house stuff that was developed without any form of portability in mind.

Again, you are correct, BUT they don't. And there's nothing I can do about since none of the competitors in this very specialized field are any different and I'm just one person out of their 100 or so clients who complains but buys the software anyway because I have no choice to do what my company wants to do.

But if you keep complaining eventually they might fix it, especially if others start complaining because they don't want to pay the licensing fees for Oracle or MSSQL.

 

[mechBgon]

and Microsoft is getting $100 (or something like that) that should be going to the opensource developers

Microsoft is not getting $100 for the Windows installation on a Dell/HP/eMachine. I think it's about 10% of that. You don't seriously think Dell is selling a $400 Dementia 3000 and then handing 25% of the sale to Microsoft, do you?

(unlike Windows, which requires additional expensive software to be even marginally resistant).

The word "requires" is pretty strong. I run antivirus software as backup both at home and at the office, but my main line of defense is simply to keep my Windows rigs patched, and use a Limited or Restricted-User account except when I need Admin powers. This should be no news to the *nix guys, it's a page straight from your own book. Available in the Windows world since WinNT.

Empirical evidence shows that it's a pretty solid technique, too. SpywareInfo put out a typical article about how OMG they went to a wrestling site with WinXP and IE6 (and obviously an Admin account) and clicked an ActiveX prompt and then Bad Things happened. So I inquired privately as to what wrestling site, then waded in (yes even the Babes pages) and tempted Fate with my WinXP/IE6 combo under my Limited account. No dice. It did lower my IQ about 10 points to be exposed to that drivel but my computer was unscathed. Anyone really surprised? And as you probably have read, Vista will run the browser at reduced privilege level. Back to the drawing board with Point #2 there... At least, Microsoft can play the same spin game by pointing out the security improvements in Vista, if it's just the PR aspect you're discussing.

 

[Brazen]

Originally posted by: mechBgon

and Microsoft is getting $100 (or something like that) that should be going to the opensource developers

Microsoft is not getting $100 for the Windows installation on a Dell/HP/eMachine. I think it's about 10% of that. You don't seriously think Dell is selling a $400 Dementia 3000 and then handing 25% of the sale to Microsoft, do you?

No, in fact I know they are not for the home edition, which is why I picked a nice round number and pointed out "or something like that." I would guess that they are for the Pro edition, which does not come on the $400 Dimension 3000. But that does not matter, you are already light-years away from getting the point.

The word "requires" is pretty strong.

*sigh* yes it is. You have quoted me horribly out of context and backed up your argument with how things SHOULD be done and not how home users actually do things. Again, this is really just drivel and not relevant to my previous posts because you have once again HORRIBLY missed the point.

 

[Brazen]

People are also lazy, stupid and dislike change. Most people only moved to Win2K and XP when they bought a new machine or when some other requirement forced them. Switching OSes when they already have one that came with the machine and is familiar with them seems unthinkable, if they've even heard of Linux.

Again, I'm really assuming you are not trying to negate what I'm saying. At least, I agree with everything you are saying, but that doesn't go against my "suggestions." To the above quote I say:

You are right, the key is not necessarily to get people to switch, but to get Linux on new computers. And how do you do that? "Hype" and advertising, as I said.

That's the biggest hurdle, custom in-house stuff that was developed without any form of portability in mind.

Ok. So how do you suggest getting past it? If that's YOUR biggest problem with getting rid of Windows, then you'll have to address it. Such as, are you sure there is not an open source alternative to your in house software? Maybe you should get your company to start one. Highly unlikely I know. In house stuff isn't a problem for us because, well, we don't have any. For me, as a sysadmin for a small/medium business (~250 workstations), the Active Directory functionality is my biggest hurdle.

But if you keep complaining eventually they might fix it, especially if others start complaining because they don't want to pay the licensing fees for Oracle or MSSQL.

Maybe they will. I'm trying. And hoping. And maybe if I keep posting on forums and telling colleagues about the merits of opensource software, it will help fuel that change.

 

[Rilex]

because Microsoft will never add Linux support to Active Directory.

I'm not sure what exactly you're shooting for here, but Microsoft is already doing this and is soon to release it with Windows 2003 R2 -- Active Directory Federation Services (ADFS). It will have OOTB support for NFS & NIS.

I'm not sure why you'd run Active Directory via Samba. It doesn't make much sense unless you're migrating an entire corporation away from Windows to OD or another form of LDAP.

I don't ever see the Linux NTFS driver going stable. NTFS is a closed file system and writing to NTFS outside of Windows is dangerous because of the journal log. As Microsoft moves to transaction-based NTFS (and registry, too!), again the driver will need another update.

I also don't see Linux and Windows competing. We've seen enough "Linux on the Desktop" threads (well, at least at Ars) to know that it simply isn't ready for the desktop, not just yet -- and we've heard this for 2+ years. Linux and Windows don't compete in the server space. Both are replacing aging (and sometimes not so aging) UN*X Iron. "Offical" statics support this, with Windows having the largest server marketshare.

Internet Serving stuff; where do Apache and Mysql/Postgresql/Firebird need to be to compete with MSsql?

I really don't see Apache competing with MS-SQL But regardless, Microsoft has shown huge improvements with IIS6 when it comes to security. In fact, compared to Apache 1.x & 2.x, it looks just rosey.

I don't think MySQL has much going for it if you can afford MS-SQL, especially SQL 2005.

IMO, YMMV

By the way, why use ScriptLogic over say straight vbscript?

 

[R3MF]

Originally posted by: Nothinman

i don't need to guess about the modularity of X.org, it has been released with Mandriva this week. what matters is why they modularised it............ because they thought it would allow rapid performance and feature gains in future.

I know why the modularized it, the old monolithic build system was the cause of much pain for many developers and package maintainers. But that means absolutely nothing to end users.

yeah so what, the fact is games arn't developed for linux with a few exceptions, i want to see that more prevalent and see a better desktop as a way forward.

Well then start calling game developers and asking for Linux versions, posting a threat on AT won't get the attention of any of them. Nearly all of the technology is there and ready to be used, they're not targeting Linux for other reasons.

thus i'm not terribly excited by the imminent releaes of 6.9/7.0, i'm far more excited by the prospect of what they can achieve with 7.2 given that they have this new modular environment that should result in faster feature adoption and performance.

better than that, i buy linux compatible games. but this is not the point.

"That's the biggest hurdle, custom in-house stuff that was developed without any form of portability in mind."

what needs to change? linux needs greater desktop penetration.
hos is this achieved? by providing advanced technologies to distro's who then package it up and serve it to users as a compelling alternative to windows.

 

[R3MF]

Originally posted by: kamper

Originally posted by: R3MF
I just want a better linux distro

From what I can tell, all you want is for linux to become exactly like windows.

thus we come back to: ...

Thus you come back to. Nobody here seems too concerned about autum 2006 or the "Big 4."

how did you get this impression?

well then i don't imagine you will add further of your incisive input to this discussion.................?

 

[Nothinman]

thus i'm not terribly excited by the imminent releaes of 6.9/7.0, i'm far more excited by the prospect of what they can achieve with 7.2 given that they have this new modular environment that should result in faster feature adoption and performance.

So you're just blindly hoping that something good will happen by the time 7.2 is released?

hos is this achieved? by providing advanced technologies to distro's who then package it up and serve it to users as a compelling alternative to windows.

Not always. For instance, the timesheet program where I work is web based and there's nothing inherently Windows-specific about it other than the fact that the developers only tested on IE and because of that some of the JS stuff doesn't work in FF even on Windows.

I don't ever see the Linux NTFS driver going stable. NTFS is a closed file system and writing to NTFS outside of Windows is dangerous because of the journal log. As Microsoft moves to transaction-based NTFS (and registry, too!), again the driver will need another update.

So? If MS isn't willing to release the NTFS docs under a usable license, of course the driver will take forever to write. And I personally don't care, the only people MS is hurting is their own customers.

I also don't see Linux and Windows competing. We've seen enough "Linux on the Desktop" threads (well, at least at Ars) to know that it simply isn't ready for the desktop, not just yet -- and we've heard this for 2+ years

IMO Windows isn't ready for the desktop either. It's got some really huge warts that people have just grown accustomed to. The only real problem is lack of games, just about every other program on Windows has an alternative on Linux that's just as good or better.

 

[Brazen]

Rilex

I'm not sure what exactly you're shooting for here, but Microsoft is already doing this and is soon to release it with Windows 2003 R2 -- Active Directory Federation Services (ADFS). It will have OOTB support for NFS & NIS.

*ugh. Microsoft has supported NFS and NIS for some time. Neither of those are related to ADFS, and none of this address administering Linux workstations through Active Directory.

I'm not sure why you'd run Active Directory via Samba

Well, people do; SAMBA currently supports it to a minor extend. SAMBA4 is supposed to improve their Active Directory implementation by leaps and bounds.

I don't ever see the Linux NTFS driver going stable. NTFS is a closed file system and writing to NTFS outside of Windows is dangerous because of the journal log. As Microsoft moves to transaction-based NTFS (and registry, too!), again the driver will need another update.

I also don't see Linux and Windows competing. We've seen enough "Linux on the Desktop" threads (well, at least at Ars) to know that it simply isn't ready for the desktop, not just yet -- and we've heard this for 2+ years. Linux and Windows don't compete in the server space. Both are replacing aging (and sometimes not so aging) UN*X Iron. "Offical" statics support this, with Windows having the largest server marketshare.


quote:

--------------------------------------------------------------------------------
Internet Serving stuff; where do Apache and Mysql/Postgresql/Firebird need to be to compete with MSsql?
--------------------------------------------------------------------------------



I really don't see Apache competing with MS-SQL But regardless, Microsoft has shown huge improvements with IIS6 when it comes to security. In fact, compared to Apache 1.x & 2.x, it looks just rosey.

I don't think MySQL has much going for it if you can afford MS-SQL, especially SQL 2005.

IMO, YMMV

By the way, why use ScriptLogic over say straight vbscript?

You are so far off on so many things. It seems that you do not understand a single thing you have talked about. No offense, there's nothing wrong with that. You just need to maybe take it a little slower.

 

[drag]

thus i'm not terribly excited by the imminent releaes of 6.9/7.0, i'm far more excited by the prospect of what they can achieve with 7.2 given that they have this new modular environment that should result in faster feature adoption and performance.

Modular X is going to be great. More rapid updates with Debian and smaller updates and bugfixes is nice. It's a pretty big deal.

After they get 7.0 out the door I hope that developement of XGL picks up more. Get rid of this multiple driver crap and move it to a OpenGL api. That'll help move it off of the hardware, get the drivers in the OS itself rather then in the X server, and make it so that it can run completely in a usermode and close a big security hole.

That (and combined with modularity) should allow it to more easily adapt to multiple operating systems, increase security, increase performance, allow it to more easily use things like the excellent Linux hotplug (and related userspace apps/systems) support for things like mice, keyboards, joysticks, tablets (which it has absolutely no support for now) and support actual multiple users on a single computer with no weird hacks.

If you think about it a X Server should be a rather a insignificant program (in terms of complexity and size) in a larger OS rather then practically be a entire OS itself like it currently is.

 

[Brazen]

Originally posted by: Nothinman

IMO Windows isn't ready for the desktop either. It's got some really huge warts that people have just grown accustomed to. The only real problem is lack of games, just about every other program on Windows has an alternative on Linux that's just as good or better.

:thumbsup:QFT. As it is, I agree it is a superior product. The problem is with 3rd party developers. At this point, microsoft is retaining it's OS domination because.... it IS the dominant OS. It's a vicious catch22 for linux. It is only going to break that cycle by being so far better than Windows that people use it regardless of the poor 3rd party support, which it IS doing, very slowly.

 

[bersl2]

Originally posted by: drag

thus i'm not terribly excited by the imminent releaes of 6.9/7.0, i'm far more excited by the prospect of what they can achieve with 7.2 given that they have this new modular environment that should result in faster feature adoption and performance.

Modular X is going to be great. More rapid updates with Debian and smaller updates and bugfixes is nice. It's a pretty big deal.

After they get 7.0 out the door I hope that developement of XGL picks up more. Get rid of this multiple driver crap and move it to a OpenGL api. That'll help move it off of the hardware, get the drivers in the OS itself rather then in the X server, and make it so that it can run completely in a usermode and close a big security hole.

That (and combined with modularity) should allow it to more easily adapt to multiple operating systems, increase security, increase performance, allow it to more easily use things like the excellent Linux hotplug (and related userspace apps/systems) support for things like mice, keyboards, joysticks, tablets (which it has absolutely no support for now) and support actual multiple users on a single computer with no weird hacks.

If you think about it a X Server should be a rather a insignificant program (in terms of complexity and size) in a larger OS rather then practically be a entire OS itself like it currently is.

Well, true X via GL is a long way off. EXA has the advantage of being ready now and having financial backing (from Trolltech), so not everyone is convinced that Xgl needs to be a high priority at the moment. This (and also because he wasn't getting paid for it ) is why one of the two developers, Jon Smirl, stopped his work on Xgl. He subsequently wrote up a piece on his view of the future, and an (at times) very ugly thread ensued. A lot of it is over whether further EXA development duplicates functionality that OpenGL already provides.

One thing that you don't know is holding back development is memory management by the Direct Rendering Infrastructure/Direct Rendering Manager. Apparently, it's in an abysmal state right now and needs to be rewritten once X is modular.

If you are interested in the future of the Linux desktop, reading the mailing lists hosted on freedesktop.org every so often is one thing you'll want to do. HAL, D-BUS, Cairo, and X discussion all happen there.

 

[mechBgon]

*sigh* yes it is. You have quoted me horribly out of context and backed up your argument with how things SHOULD be done and not how home users actually do things. Again, this is really just drivel and not relevant to my previous posts because you have once again HORRIBLY missed the point.

If your point is that Linux enjoys "security by obscurity," then you surely realize that would not last forever. Imagine a world where Linux is the de facto standard OS and the home users treat it like they now treat Windows. Logged in as root all the time, never updating because they heard that their co-worker's best friend's husband messed up the computer by trying that update stuff, letting the kids download and install anything they want to, opening that dramatic-sounding email titled *DETECTED* ONLINE VIOLATION that their ISP seems to have sent... yeah. Gee, the cable-modem sure blinks a lot lately...

Anyway, good luck putting a dent in Microsoft's machine. I may get around to trying some Linux someday, thanks to positive examples by people like drag and Nothinman and n0c, but frankly it doesn't bug me to be shelling out $250 for a product that will be supported for ten years, either.

 

[Rilex]

So? If MS isn't willing to release the NTFS docs under a usable license, of course the driver will take forever to write. And I personally don't care, the only people MS is hurting is their own customers.

How does this hurt MS' own customers? Microsoft's customers == Windows users. Windows users == native NTFS support.

There is no logic in what you just said.

IMO Windows isn't ready for the desktop either. It's got some really huge warts that people have just grown accustomed to. The only real problem is lack of games, just about every other program on Windows has an alternative on Linux that's just as good or better.

The "application game" is not something I'm going to play because it is pointless.

Microsoft has supported NFS and NIS for some time.

Thank you, Cpt. Obvious. I know this but R2 has them OOTB (did you read a little too quickly?).

Neither of those are related to ADFS

ADFS relates to UN*X interoperability.

address administering Linux workstations through Active Directory.

Hence my "I'm not sure what exactly you're shooting for here".

Well, people do.

Which is just asinine. Why bother?

You are so far off on so many things.

Start naming them, I'm waiting.

It seems that you do not understand a single thing you have talked about

I'd have to say "likewise".

 

[Brazen]

Originally posted by: mechBgon

*sigh* yes it is. You have quoted me horribly out of context and backed up your argument with how things SHOULD be done and not how home users actually do things. Again, this is really just drivel and not relevant to my previous posts because you have once again HORRIBLY missed the point.

If your point is that Linux enjoys "security by obscurity," then you surely realize that would not last forever.

I never mentioned security _by obscurity_. "Security by obscurity" is an unproven theory anyway. Apache is the most popular webserver, how come it has always had fewer exploits than IIS?

Anyway, good luck putting a dent in Microsoft's machine

Every year, Linux's market share goes up and Microsoft's market share goes down. I would say the "dent" has already been "put."

I may get around to trying some Linux someday, thanks to positive examples by people like drag and Nothinman and n0c,

I think that's all the linux advocates really want. Just try it before you knock it. Nothinman and n0c have impressed me too with their technical knowledge of linux (drag, though I do recognize the name, has not made an impression that I remember, no offense drag :beer: ). The Fedora and Ubuntu forums also great places, even just to browse and find out things you can do on linux that you never thought of. Good luck.

 

[stash]

ADFS relates to UN*X interoperability.

Not necessarily. ADFS is basically a method of securely sharing information about users across organizational and security boundaries. Additionally, ADFS in R2 will only provide SSO to allow a user to authenticate to multiple web applications that run on IIS. Third parties like Centrify and Vintela will extend this to non-windows web applications (J2EE, etc), but ADFS is not fundementally about interop with Unix systems.

I think maybe you are confusing ADFS with the three Unix interop components that will ship with R2: Microsoft Services for NFS, Subsystem for Unix-based apps and Identity Management for Unix.

 

[xtknight]

Going off my (albeit somewhat limited) experience of Linux distributions (collectively referred to as "Linux" from here on), the ease of use of applications needs a lot of work. This is assuming the goal of Linux is to be a competitor to Windows on the desktop. Keep in mind I never said the goal of Linux was indeed to do this or not, but that's what the OP wanted to talk about, so let me begin.

Linux needs standards. GTK? Qt? Tcl/Tk? Gnome? KDE? Fvm95? Windowmaker? Enlightenment? XFree86? X.Org? There is just too much confusion out there. Choice is good, but they need to be consolidated somehow, unified, or "nexused". Windows has one shell by default that supports a vast amount of themes and visual styles, and I think Linux should do the same. This is of course just one example.

Everything needs to be more user friendly. Actions need to be more consistent. Windows (at least the apps it ships with) does a stellar job of keeping functionality consistent across the board.

There is an overload of options. Gaim or Kopete? Firebird or Evolution? I don't know.

Linux also needs a "DirectX" equivalent. DirectShow is a decent implementation of media playback on Windows. On Linux, what do we have? Nothing, or have I just not heard of it? You need to download ffmpeg... but it's not all that bad. One thing I can give to Linux: ffmpeg supports damn near every media file out there. DirectShow by default? Let's just say...Cinepak-compressed AVIs, uncompressed AVIs, Microsoft MPEG-1, and the later versions Windows Media and MP3? That's diddly-squat. Still, DirectShow is a much better architecture because of it's chain-linked system. Filters are not few and far between.

The start menu on most distributions I've seen is an unorganized mess.

What to do when Xwindows fails to initiate? Who knows...

Plus it REALLY needs an easier way of mounting drives. Why in the damned hell is my SATA drive named as some SCSI/RAID/or whatever device (/dev/sda I think it was)? How was I supposed to know my Windows drive was /dev/sda? Where's my disk manager? I mean a graphical one, not fdisk print.

The other thing that is in Linux's favor now is driver configuration, IMO. Linux sets up a lot of your devices automatically. My printer worked, sound worked (on my P4 machine), video worked at full resolution (though unaccelerated). None of that on Windows.

 

[Brazen]

Originally posted by: Rilex

You are so far off on so many things.

Start naming them, I'm waiting.

Well, it's really a matter of going through your entire post and picking out every single computing and IT related term you threw out there. You can search through the forums or google on the internet to learn more about them. It's really beyond the scope of a single thread, let alone a single post. Take your time, just take it one term or concept at a time; don't get over your head.

If you are not sure what all terms you used are computing and IT related, maybe this isn't the place for you. Maybe sign up for an introductory computer course at your high school or college, whatever the case may be. Other than this, I'm not going to pollute this thread on your account any more.

 

[bersl2]

Originally posted by: xtknight
Going off my (albeit somewhat limited) experience of Linux distributions (collectively referred to as "Linux" from here on), the ease of use of applications needs a lot of work. This is assuming the goal of Linux is to be a competitor to Windows on the desktop. Keep in mind I never said the goal of Linux was indeed to do this or not, but that's what the OP wanted to talk about, so let me begin.

Linux needs standards. GTK? Qt? Tcl/Tk? Gnome? KDE? Fvm95? Windowmaker? Enlightenment? XFree86? X.Org? There is just too much confusion out there. Choice is good, but they need to be consolidated somehow, unified, or "nexused". Windows has one shell by default that supports a vast amount of themes and visual styles, and I think Linux should do the same. This is of course just one example.

And a boatload of developers disagree that projects need to be merged. What needs to be developed are more common protocols and APIs, ICCCM-style.

Everything needs to be more user friendly. Actions need to be more consistent. Windows (at least the apps it ships with) does a stellar job of keeping functionality consistent across the board.

This statement is nebulous.

There is an overload of options. Gaim or Kopete? Firebird or Evolution? I don't know.

Why does it matter then? Choose one; if you don't like it, try the other.

Linux also needs a "DirectX" equivalent. DirectShow is a decent implementation of media playback on Windows. On Linux, what do we have? Nothing, or have I just not heard of it? You need to download ffmpeg... but it's not all that bad. One thing I can give to Linux: ffmpeg supports damn near every media file out there. DirectShow by default? Let's just say...Cinepak-compressed AVIs, uncompressed AVIs, Microsoft MPEG-1, and the later versions Windows Media and MP3? That's diddly-squat. Still, DirectShow is a much better architecture because of it's chain-linked system. Filters are not few and far between.

There is an equivalent: libSDL. In fact, when compiled in a Windows environment, SDL acts as a wrapper to DirectX. So if you decide to program for SDL in the first place, you can run that code for just about any platform.

The start menu on most distributions I've seen is an unorganized mess.

I wouldn't know.

What to do when Xwindows fails to initiate? Who knows...

Having an automatic fallback to the VESA driver would be nice.

Plus it REALLY needs an easier way of mounting drives. Why in the damned hell is my SATA drive named as some SCSI/RAID/or whatever device (/dev/sda I think it was)? How was I supposed to know my Windows drive was /dev/sda? Where's my disk manager? I mean a graphical one, not fdisk print.

The combination of linux-hotplug, udev, D-BUS, and HAL will "standardize" automatic mounting. HAL spec has a diagram for how the system works.

The other thing that is in Linux's favor now is driver configuration, IMO. Linux sets up a lot of your devices automatically. My printer worked, sound worked (on my P4 machine), video worked at full resolution (though unaccelerated). None of that on Windows.

And we can do even better, if those hardware and chipset companies fork up the specs to their hardware.

 

[drag]

Originally posted by: bersl2

Originally posted by: drag

thus i'm not terribly excited by the imminent releaes of 6.9/7.0, i'm far more excited by the prospect of what they can achieve with 7.2 given that they have this new modular environment that should result in faster feature adoption and performance.

Modular X is going to be great. More rapid updates with Debian and smaller updates and bugfixes is nice. It's a pretty big deal.

After they get 7.0 out the door I hope that developement of XGL picks up more. Get rid of this multiple driver crap and move it to a OpenGL api. That'll help move it off of the hardware, get the drivers in the OS itself rather then in the X server, and make it so that it can run completely in a usermode and close a big security hole.

That (and combined with modularity) should allow it to more easily adapt to multiple operating systems, increase security, increase performance, allow it to more easily use things like the excellent Linux hotplug (and related userspace apps/systems) support for things like mice, keyboards, joysticks, tablets (which it has absolutely no support for now) and support actual multiple users on a single computer with no weird hacks.

If you think about it a X Server should be a rather a insignificant program (in terms of complexity and size) in a larger OS rather then practically be a entire OS itself like it currently is.

Well, true X via GL is a long way off. EXA has the advantage of being ready now and having financial backing (from Trolltech), so not everyone is convinced that Xgl needs to be a high priority at the moment. This (and also because he wasn't getting paid for it ) is why one of the two developers, Jon Smirl, stopped his work on Xgl. He subsequently wrote up a piece on his view of the future, and an (at times) very ugly thread ensued. A lot of it is over whether further EXA development duplicates functionality that OpenGL already provides.

EXA is just a bandaid. It's simply a system to extend the current driver model for X for another couple more years. It's not a real solution...

One thing to think about is that currently the only cards that have EXA drivers have perfectly well working OpenGL drivers. Cards that don't have OpenGL drivers can use software MESA. Cards that don't have OpenGL drivers don't have EXA drivers for the simple reason that they don't exist. That is that OpenGL system is more likely to work on more hardware right now. It may be slower, but it's better then non-functional.

Also OpenGL works now thru mesa, on a whole host of other operating systems and weird hardware which EXA doesn't.

EXA doesn't solve any problems like moving X out of requiring root access to run. It doesn't solve the problems that every OS that has to support EXA has to support it's own set of display drivers and X's own display drivers and make sure they work together.

If you don't beleive me, then read what the developer of EXA has to say about it

You see with OpenGL all a OS has to do is have it's own OpenGL stack. It can be OS X, Windows, Linux with DRI, Linux with Nvidia drivers, Linux with mesa, FreeBSD, OpenBSD, etc etc etc. Just as long as they have the stack they can handle the display in any way they feel like. With EXA they have to port the EXA drivers to their OS and use X's management of the hardware which will probably be sub-optimal compared to the way the OS itself would like to handle it.

One thing that you don't know is holding back development is memory management by the Direct Rendering Infrastructure/Direct Rendering Manager. Apparently, it's in an abysmal state right now and needs to be rewritten once X is modular.

DRI isn't part of X. It's direct rendering infrastructure for Linux, a seperate project.. It works closely with X developers though.

For instance there is Mesa-Solo, which is a standalone 3d system that doesn't run on any X server.. it runs directly on the framebuffer just fine. DRI drivers are based on Mesa and simply accelerates what parts in hardware that it can.

The trouble with Dri isn't memory management, as far as I know, it's the fact that card manufacturers are hiding card details in a way that it makes it impossible to write drivers for their hardware.

The memory management issues I think your refering to were probably related to one of the developers of XeGL dropping out.

He was pissed because he knows that XGL is the real solution to X's problems and EXA was a temporary solution to avoid those issues for the next year or two. If people devoted resources to XeGL we would of had a fully OpenGL-driven X server in six months to a year. But currently people are working on getting the sorely needed modular X server, which is behind scedual, out the door. (modular X server is nessicary for XGL realy though)

Plus having EXA was the nail in the coffin for him. It was further dividing resources and pushing back XGL developement by another year.

keep in mind that the only EXA drivers that were aviable are for cards that already had perfectly working DRI OpenGL drivers. All these new drivers would have to be written for cards that had good OpenGL drivers and good XAA drivers.

For future reference note that the original XGL X server is not a standalone X server. It is a 'nested' X server similar to OS X's x server or xnest. It needed some of the management from a standard X server to work (such as resolution and such).

XeGL is based on that work but was designed to include the some of the ES extensions to the standard OpenGL API.

OpenGL-ES was designed to impliment a subset of OpenGL stuff for embedded platforms that lack the hardware resources to complety run a full OpenGL system. Specificly the XGL developers needed to take the stuff they wanted was the ability to send commands to drivers to set things like refresh rates or set screen resolution, not make it run on 'lesser' hardware. It would require a full OpenGL implimentation just with some part of the (already existing, not X propriatory) ES extension.

XeGL works _right_now_. We have a fully OpenGL driven X server _right_now_ in the form of XeGL. It runs on a Linux framebuffer and depends on ATI 9200 video card... but it works and it works right now.

The trouble is is that Xegl's memory management bits and lots of other details needed work and it wasn't going to get done anytime soon with 2 developers working on it.

If you are interested in the future of the Linux desktop, reading the mailing lists hosted on freedesktop.org every so often is one thing you'll want to do. HAL, D-BUS, Cairo, and X discussion all happen there.

Yes I know. I've subscribed to several mailing lists for some time now.

If you have a link to were people are discussing Mesa/Dri's memory management issues I'd like to see that. (serious about this)

Here is a very interesting article about the current state of Linux graphics. It's by John Smirl, the fellow who dropped out of Xegl for the time being to persue more fruitfull pastures.
http://dri.freedesktop.org/~jonsmirl/graphics.html

The "other" XeGL developer and creator of XGL is David Reveman. He is a busy contributer to the X.org project and is also the developer of Glitz which is the OpenGL replacement for Xrender (which EXA is designed to extend for a couple more years), which is also going to be used as the backend for accelerated Cairo graphics. Right now he works for Novell (ya novell!) which hired him specificly for his work with XGL and such.