What do you think would be THE "threshold development" capable to convert 80% of Windows users to LINUX?

[Samsonid]

What would it be?

a) The ability to run ALL existing windows software without loss of performance?

b) An installer software that would allow all users of Windows (including small businesses and corporations) to remove all windows folders and replace them with LINUX while still running all existing Apps on those PCs. Thus minimizing conversion downtime!

c) More extensive driver support?

d) A company to provide centralised support for all Linux issues?

e) Further refinement of the Linux GUI?

What would you add to the above list?

 

[spamsk8r]

Well, I don't think 80% of people have any need for linux...especially if all its going to do is run windows programs. They already have windows for that, why change it if it works for them, and for most people it works just fine. I don't think that the majority of mainstream society will ever switch over to linux completely, for the simple reason that they already have windows, they dont need new and difficult things to learn. Just my 2 cents

 

[Samsonid]

Yep,

Therefore Linux would have to have a niche. Something that would outclass windows in a definitive way.

How about a GUI that has none of the difficulties quirks and annoyances of Windows?

Ohh..ohhh.. I know... what if Apple ported their Linux based OS-X to the PC platform?

 

[Saltin]

Your suggestions might place LINUX in a position to compete with Windows on desktops, but it wouldnt create a situtation where 80% of Windows users would migrate to LINUX.
Microsoft are always going to steamroll the opensource guys on Desktop's because they can afford to advertise. Who's going to pay to get the word out for LINUX?
In a perfect world, word of mouth and a superior product are all you need, but in reality the ability to put your product in people's faces is really what matters.

To be perfectly honest, I don't think *Nix is interested in the Desktop market. It's not thier niche. Maybe I'm wrong, who knows.

 

[ToBeMe]

<< What do you think would be THE "threshold development" capable to convert 80% of Windows users to LINUX? >>


Nothing....................What I mean is, that will never happen (80% switch). Why? Because the VERY vast majority of users have absolutely no need to and would never take the time to learn something new! Like it or not, the majority of users will always be those whom do nothing but check/write email, play a few games, and chat...............they have very little need/desire to upgrade at all and changing to a different O/S is something they wouldn't/couldn't fathom...........

 

[Nothinman]

To get 80% to switch you're going to have to do it without them knowing, because frankly most people don't care what their OS is and the rest don't know there's anything else out there.

If some corporations could switch their desktop, the users who work from home may swtich.
If computer appliances ever find a niche Linux will probably power a number of them and the users won't care.

Microsoft are always going to steamroll the opensource guys on Desktop's because they can afford to advertise. Who's going to pay to get the word out for LINUX?

IBM. There's quite a few Linux commercials from IBM, they're mostly aimed at servers right now but that's not to say they won't move towards desktops in the future.

 

[LNXman]

Agreed. Most people (from the guesstimated 80%) DO NOT LIKE TO LEARN SOMETHING NEW. And when they see that LINUX does not have great support (indicating that the user may have to tweak the O.S.), LINUX becomes very unappealing to use. Besides, Many people who have actually attempted to use LINUX, end up complaining, and so their experiences with LINUX get transmitted to potential other users, who, in the end, decide not to even bother. Sad, but true.
And this applies to any *NIX like system . . .
But, whatchagonnado? Maybe, if LINUX keeps changing its looks like Micro$oft does to make the user not notice the lack of performance. . . lol

Just look at XP. It looks pretty, it has LINUX like features, but it is full of problems. The worst release by Micro$oft, if I have ever seen one. . . Win2K is the best release since 3.11. Even though it has some performance problems, it is more stable than XP.

Development wise, uhh. . . I do not know what Windows has, that LINUX does not have probably 3 different versions of. The only questions I have are: why does Micro$soft tries to look more like some Window Manager in LINUX? Why do M$ developers go to LINUX expos to ask development questions? I was a witness to one such account last August at the Moscone Center LINUX expo in S.F., where a top Win XP developer was asking about a window manager structure, and what were going to be the new features of it. His identity was not discovered until he was there for quite a while taking notes after each question, and asking if there were any demos, at which point one of the deveopers asked for his business card (because he might have been an interested corporation willing to donate), and the developer was shocked to find out who the guy was, to the point that he was laughing (not to his face). Honest! Once he let his other developers at his booth know about him, they all just looked at the XP guy, and smiled. . .
I thought that was pretty funny.

----------------------------------------------------------------



<< Ohh..ohhh.. I know... what if Apple ported their Linux based OS-X to the PC platform? >>



Correct me if I am wrong, but . . . isn't OS-X based on BSD and not LINUX?

/edit: added clarification

 

[ToBeMe]

<< Just look at XP. It looks pretty, it has LINUX like features, but it is full of problems. The worst release by Micro$oft, if I have ever seen one. . . Win2K is the best release since 3.11. Even though it has some performance problems, it is more stable than XP. >>


Curious about that staement..........XP is basicly 2K with added eye candy and a few new items as well as SP3 (which is still not released to the public..........but ready). Our tech support expierience has been that our customers with XP installed have had fewer problems than any previous MS release and the Hardware/Software support compared to 2K is astonishing! I've heard so many new XP customers comment that they never have to reboot like they did with any 9X and they hardly ever encounter any errors much less BSOD's.

I recall the release of 2K very well........I remember lack of driver support, lack of software support, and lots of overall configuration problems............tech support during that time was seemingly working non-stop! I agree that 2K has come a LONG way especially since SP2, but, it's beginning was much more marred with problems than XP's at least among the people I talk with...........

 

[LNXman]

That is what is turned out to be, and I don't know why?

XP (or whistler as it used to be called) was supposed to be a WHOLE re-write of Win2K, where:
-*NIX file system was going to be supported.
-WINs was going to dissappear completely, and TCP/IP was going to be the main protocol for networking.
-Performance was going to greatly improve.
-Easier to use.

As a regular desktop end-user. XP has its minimal problems, such as driver problems, desktop issues, etc. . . and most of these simple problems come from UPGRADING from Win2K.
However, from an admin/developer point of view, XP's performace is worse than it was in Win2K. Y? I don't know, but it would be upsetting to me, if I relied on Windows to develop my products. Check this test. And there are many other performance issues with XP. Why do you think XP is only being used as a development OS in a lot of companies, instead just going ahead and upgrading from 2k? Because developers (and admins) know what the problems are. I have yet to see a company with a fully XP core server running (maybe the price is a deterent too). All I have seen are desktops, and test servers.

Well, at least it is easier to use. . .lol

 

[bsobel]

Lets set some facts straight here:

> XP (or whistler as it used to be called) was supposed to be a WHOLE re-write of Win2K, where:

That is not, and never was true. I think your confusing Win2k with the 9x platform. This was a major kernel/os change for those making that move. Win2K to WinXP was always an evolutionary upgrade.

> -*NIX file system was going to be supported.

Nope

> -WINs was going to dissappear completely, and TCP/IP was going to be the main protocol for networking.

TCP/IP is the main protocol, it's quite a pain to get NetBeui support back in.

> -Performance was going to greatly improve.

It's improved, not sure about greatly.

> -Easier to use.

I think they did that one.

> However, from an admin/developer point of view, XP's performace is worse than it was in Win2K.

That is not what we see here, and we do ALOT of Win32 and ringzero development on 2K and XP.

> I have yet to see a company with a fully XP core server running (maybe the price is a deterent too). All I have seen are desktops, and test servers.

.NET server (XP core server as you call it here) is still at the release canidate stage. Maybe the reason you haven't seen it running is it's not released yet. Sheesh.

Bill

 

[LNXman]

<< Lets set some facts straight here:

> XP (or whistler as it used to be called) was supposed to be a WHOLE re-write of Win2K, where:

That is not, and never was true. I think your confusing Win2k with the 9x platform. This was a major kernel/os change for those making that move. Win2K to WinXP was always an evolutionary upgrade. >>



You are correct here. The main feature of XP was supposed to be to improve areas of Win2K, introduce new features, and weed out/ replace others. I had to re-check on that one, cause I was doubting my own post. . .lol



<< > -*NIX file system was going to be supported.

Nope >>



Actually, the very first note about Whistler DID mention the support for *NIX filesystems and NTFS (Sorry I can't find the note for that) However, I understand why they (Micro$oft) may not have done that., or talk about it afterwards, . . and my believe is because Windows NTFS has a MUCH BETTER, and complicated, permission handling mechanisim than *NIX. Although it can be overly saturated and cause headaches in the long run if you don't document what you do, it is still good. That major incompatibiliy is a big obstacle to support both, and it is a b!tch to try to manage both at the same time.



<< > However, from an admin/developer point of view, XP's performace is worse than it was in Win2K.

That is not what we see here, and we do ALOT of Win32 and ringzero development on 2K and XP. >>



I am curious to know what kind of programs you help develop/test. I am sorry, but I am not familiar with ringzero. . . what kind of loads do you test them with? And what kind of tests? Sockets? Threads? Pipes? Data transfers?



<< > I have yet to see a company with a fully XP core server running (maybe the price is a deterent too). All I have seen are desktops, and test servers.

.NET server (XP core server as you call it here) is still at the release canidate stage. Maybe the reason you haven't seen it running is it's not released yet. Sheesh. >>



I am curious, when did I mention .NET? I was just referring to just replacing/upgrading the O.S. to XP in the core servers. Not to update/use some new monopolizing project by Micro$oft. .NET is another thing. And yes, it has not been released yet. . . . I thought .NET was a service, and NOT an O.S? Am I off? Please clarify.

/edit: clarified.

 

[KenCleanAirSystem]

Windows .NET Server=The WindowsXP equivalent to Win2000 Server/Advanced Server/Datacenter. It is still in beta at the moment here's some info.

As far as Linux on the desktop, the only thing I could see converting a significant amount of average computer users would be for a company such as Gateway or Compaq with a large consumer base to start installing it as the default OS. Most average computer users will never even actually install an OS by themselves, and unless they buy a Mac they're stuck with Windows.

 

[LNXman]

<<
Microsoft .NET is the Microsoft XML Web services platform. XML Web services allow applications to communicate and share data over the Internet, regardless of operating system, device, or programming language
>>



After reading about it from your link, I see it is NOT the XP O.S., but a service. Thanks for the link KenCleanAirSystem. I was getting confused.

 

[Nothinman]

Actually, the very first note about Whistler DID mention the support for *NIX filesystems

That's a lot of filesystems, even ufs has ~7 different versions. And frankly I don't see a reason to do it, how many people really put a Sun hard disk in an NT box and expect it to read it? They'd also have to support the different disk labling formats as opposed to PC partitions.

I am sorry, but I am not familiar with ringzero. . .

ring zero means kernel level.

I am curious, when did I mention .NET? I was just referring to just replacing/upgrading the O.S. to XP in the core servers.

XP's server edition will be called Windows.NET Server. Currently there is no server version of XPl, only Home and Pro.

 

[LNXman]

<<
. . .frankly I don't see a reason to do it, how many people really put a Sun hard disk in an NT box and expect it to read it? They'd also have to support the different disk labling formats as opposed to PC partitions
>>



Maybe I should clarify. I was trying to strickly refer to Server versions of the O.S. that were going to support *NIX and NT FS. Development companies that develope apps for NT/Win2K/XP,Solaris,HPUX,AIX will most likely attempt to share that Solaris/Irix/HPUX Disk with an NT box to build their products for efficiency. In my last developing company, we had what it was called a "Net Enterprise" appliance, which had its own O.S. that, according to the company, guaranteed to support exporting *NIX & NTFS partitions with simultaneous permission handling . However, it turned out to be a nightmare because NTFS files would get corrupt to the point which NO-ONE owned the file, and NO-ONE had permissions to touch the file. It was very frustrating. The reason was that when you changed permission from *NIX to a file on an NTFS partition through the appliance, it ended up wiping all permissions, ownership,etc. Thus, corrupting the NTFS file. And crapping out the entire build process. . . This appliance existed before whistler was mentioned. What we ended up doing, was to split both FS's, cause the maker of the appliance did not have a solution. When I read the note on whistler I was happy, because I believed that Micro$oft would be the only one to solve this type of problem. Yet it never happend. At the time, I really wanted to hammer that enterprise machine.



<<
XP's server edition will be called Windows.NET Server. Currently there is no server version of XPl, only Home and Pro
>>



I thought I saw an XP-server version out already without .NET capabilites. Maybe I got confused by the pricing of the current versions . . . Thanks for letting me know. But .NET is still a service that will come with the XP server version only? right? I don't want to confuse myself . . .

 

[KenCleanAirSystem]

My understanding is that it will be marketed as Windows.Net Server. There is no WindowsXP Server, just Profesional and Home (discounting the Corporate versions as they are essentially the same as Pro), and there never will be. Other than the name and the inclusion of .NET all the versions are basically the same as their Windows 2000 server counterparts.

 

[bsobel]

> I am curious to know what kind of programs you help develop/test. I am sorry, but I am not familiar with ringzero. . . what kind of loads do you test them with? And what kind of tests? Sockets? Threads? Pipes? Data transfers?

I work for Symantec, at one point or another I've been involved with most of the Windows development here from Norton Desktop on Win3.x to the latest NAV/SystemWorks/NIS/etc. Ringzero, as someone else mentioned is kernel level. We do extensive performance testings on all aspects of the OS, we just aren't seeing the performance loss you stated. It could certainly be machine or configuration related, I won't claim we've tested on every variation. But I don't believe the assertation that XP is slower than 2K in general.

>I am curious, when did I mention .NET? I was just referring to just replacing/upgrading the O.S. to XP in the core servers. Not to update/use some new monopolizing project by Micro$oft. .NET is another thing. And yes, it has not been released yet. . . . I thought .NET was a service, and NOT an O.S? Am I off? Please clarify.

A couple other people answered this also, .NET is a branding. Their are .NET services but the next releases of the server operating lines are to be called Windows.NET server. These are the versions based on 'xp in the core', my point is that it doesn't suprise me you haven't seen these deployed as they have not been released yet. To further confuse the issue expect to see just about everything coming out of MS called something.net for the next year or two.

Best,
Bill

 

[Nothinman]

Development companies that develope apps for NT/Win2K/XP,Solaris,HPUX,AIX will most likely attempt to share that Solaris/Irix/HPUX Disk with an NT box to build their products for efficiency.

To do that you don't need to support the host unix filesystem, just a network filesystem like NFS or SMBFS. Install Samba on a unix that supports ACLs (all commercial ones that I know of do, Linux with XFS does) and you can share files easily and you can even manipulate the ACls with Explorer on the NT boxes. Or going the other way MS has a NFS client for NT, not sure about the permission support though.

Supporting the filesystems themselves is only usefull if the partition is in the same local PC as the OS wanting to read it.

 

[LNXman]

Nothinman


<<
To do that you don't need to support the host unix filesystem, just a network filesystem like NFS or SMBFS. Install Samba on a unix that supports ACLs (all commercial ones that I know of do, Linux with XFS does) and you can share files easily and you can even manipulate the ACls with Explorer on the NT boxes. Or going the other way MS has a NFS client for NT, not sure about the permission support though.
>>



That is absolutely right, however like you mentioned, permissions are a whole other shabbang, and that is what causes the problem between administering both NTFS and *NIX partitions simultaneously. Lot's of corruption occurs. . .



<< Supporting the filesystems themselves is only usefull if the partition is in the same local PC as the OS wanting to read it. >>



That is what we initially had, then we tried the "Net Enterprise" appliance, only to go back to what we had initally set up 3 months after, due to tremendous corruptions occurring with files' permissions. . . Off the bat I did not like the appliance, it was a head-ache!!! But I hoped it would have worked . . .

-------
bsobel

Thanks for the info. I think I have understood what to expect to from XP server. . .errr. . .Windows.NET. . . I just have to begin playing with it to familiarize myself with it.

 

[Davegod]

I figured it might be useful to get the view from someone who has never had the oppertunity to use Linux, but has read about it some. That person is of course me

I'm talking about Linux trying to go mainstream, which I think is the purpose of the question. It's no good talking about servers etc, it is what businesses are using as their standard workstation OS and what people have at home.

If i knew I could get Linux, and all the programs i need/want to use works on it with comparable performance, then i would be happy to convert to Linux the next time I decide i need to get a newer OS. As it stands i'm using W98se, i feel no need as yet to upgrade even to XP. A central problem is game compatability, that was my reason for choosing 98se over for example WIN-ME or WIN-2000. If i was doing it now, i would be torn between W98se or XP, because on the one hand i want max stability and game compatability on a PROVEN o/s, on the other hand i would be unsure because i wouldnt want to spend money on an old o/s when an upgrade to XP would be likely eventually, i.e. once i feel it is proven enough to be stable and compatable. You can apply the same principles to Linux.

do not discount the influence gamers have on development of new technology. Who is it that is most likely to be buying the spanking new CPUs? businesses should be going for the best bang-for-buck, which are definately not the newest CPUs. Home users generally are happy to lose out on 15% performance for 30% cost reduction. Look at the video card development and its quite obvious. The same applies to o/s, the main people who buy things first, and hence allow both more development and initiate the produc life cycle on computers are gamers. Make it so games are all compatable with linux immediately upon release, and with either equal or better performance and Linux goes mainstream.

Quote: "
What would it be?

a) The ability to run ALL existing windows software without loss of performance?

b) An installer software that would allow all users of Windows (including small businesses and corporations) to remove all windows folders and replace them with LINUX while still running all existing Apps on those PCs. Thus minimizing conversion downtime!

c) More extensive driver support?

d) A company to provide centralised support for all Linux issues?

e) Further refinement of the Linux GUI?

What would you add to the above list?
"

- easy answer, ALL. If you want Linux to overcome windows, it has to be better in almost every way. Basically all people really want is a more stable stable windows, with some people adding "and not from microsoft".

 

[Nothinman]

That is absolutely right, however like you mentioned, permissions are a whole other shabbang, and that is what causes the problem between administering both NTFS and *NIX partitions simultaneously. Lot's of corruption occurs. .

I also said if you use Samba and a filesystem that supports ACLs it's not a big deal and works rather well. The only thing that might not work well is copying files from an NTFS volume and preserving ACLs because right now I don't believe samba knows how to map NT SIDs onto the ACLs, but just working with the files wouldn't cause any corruption.

 

[LNXman]

<< I also said if you use Samba and a filesystem that supports ACLs it's not a big deal and works rather well. The only thing that might not work well is copying files from an NTFS volume and preserving ACLs because right now I don't believe samba knows how to map NT SIDs onto the ACLs, but just working with the files wouldn't cause any corruption. >>



Have you ever created an automated/synchronized build environment for ALL types of O.S.'s? I ask this, because you don't seem to account what kind of process is involved in building apps from pulling source to packaging (not to mention the pesky '^M'). . . Let's just say, you HAVE to manage permissions, and you have to move files around from different machines (*NIX or NT) . . .And unfortunately, all kinds of overlapping processes we used would corrupt the files one way or another when everything was attempted to be managed centrally.
And yes, we used Samba the entire time, but Samba does not manipulate the build, each O.S.'s filesystem apps do. Ultimately, like I mentioned, we just decided to divide the builds into 2 separate locations, managed by different native machines, just how we had initially started. Then everything was peachy. It was an interesting experience.

 

[Nothinman]

Why didn't you use something like CVS?

 

[Insidious]

Not to be disrespectful, but it is not the job of applications to run the Operating System. It is the other way around.

If Linux is ever going to become popular to the "mainstream"... the job is pretty straight forward. It has to run applications, run them correctly, and run them without the need for an IT credential bearing operator.

The knowledge necessary to really use Linux is self-limiting. To the "mainstream", giving them Linux is like giving them a car but not telling them it needs gas. It won't work for them..... so why would they use it?

Please note this post is not directed at ANY of the posts above it... it is just a reflection of my frustration that this industry just can't seem to understand that if it doesn't work "out of the box"..... IT DOESN"T WORK!

 

[LNXman]

<< Why didn't you use something like CVS? >>



It was not upto me, but the Engineers, and Head of the Software Engineering to decide the Source code repository. It just boiled down to support, which the engineering folks take very seriously, and did not see it with CVS. I was just a pawn, getting paid well to suffer . . .lol

Support is what hurts many good *NIX software out there . .

/edit: added comment.