Crashdamage

Sounds like devries has the right idea, eliminating one copy of xorg. But instead of uninstalling the xorg-x11 rpms, I'd uninstall the xorg 6.8.2 files you installed from source first and see if that works. If so (or if not) then add Thac's rpm depository to your urpmi sources (see this thread: http://mandrakeusers.org/index.php?showtopic=23130) and do:

# urpmi -v xorg-x11

That will do a pain-free update of xorg to 6.8.2. Thac's got it down.

This looks like a fine example of why I always advise using .rpms instead of tarballs for installation.

linux_learner...cool down a bit. Go back and re-read everything that's been said. aRTee particularly has made some very strong points.

As for me, I *have* successfully installed Debian. Sure, it's good, but I didn't see any major advantage over another good distro. For several reasons, I decided to stay with a rpm-based distro. Anyway, after much playing around with apt and urpmi, my take on the whole urpmi-vs-apt thing is - I call it a toss-up. Properly used, both are pretty fool-proof. The old-wisdom statements about apt-get being the best just don't hold water anymore. Things change. Perceptions often don't keep up.

But that's just MHO, and who am I? I mean, I feel like a fairly experinced Linux user after 3 1/2 years, but I still couldn't hold a real guru's wrist pad. But I think you'd do well to heed an earlier suggestion to check out what Buchan Milne has had to say about urpmi. The guy probably knows more than everyone on this board put together, and you'd never convince him that apt is better. And check out what P.T. Breuer has had to say about urpmi vs apt also (I believe you have a personal problem with P.T., but he actually knows quite a bit and tries to help in his own special way, he just insists on people thinking for themselves). Do some research, but not at Debian or apt-for-rpm websites. Of course they're gonna say apt is better. What else would you expect?

I'm interested in when they'll support WinXP(Home) with Service Pack 2. I have a licensed copy, but when I try to install the demo, it says:

 

Checking CDROM...

loadwinproCD: not a supported Windows CDROM

 

So I guess this is not supported yet. And perhaps never may be... ? Any chance you could ask for this information?

There is already experimental support for XP Pro w/SP2. I haven't seen anybody even mention XP Home...probably nobody involved at this early stage wants to fool with such a crippled OS. I assume it's the only WinNT version you have?

But I'd be willing to bet Win4LinPro will support XP Home at least at some point. I'll post the question and see what Leo & Co. says.

Keep in mind that with Win4LinPro we are now running Win2k/XP, which are both much heavier on resources than Win9x. XP in particular is a real hog, so fast hardware + lotsa memory is a Good Thing. Especially on marginal hardware, disabling as many services a possible will pay real benefits.

I have Win4LinPro giving Win2K Pro 256MB and for my purposes (and with a lot of services off) that's probably fine. That's marginal for WinXP but my laptop is only a Pentium II400 w/256MB and it runs Win XP Pro SP2 (natively) fairly well. But I've disabled a LOT of services to get it there.

If you can get it installed, you can specify the memory allotted to Windoze by adding this line in $HOME/winxp/settings.local:

MRGPRO_WINDOWS_RAMSIZE=256 (or whatever amount you want up to 512)

See the last message in:

http://mandrakeusers.org/index.php?showtopic=21493

Just to update this thread...it got moved to 'Everything Linux' and much additional info is posted there:

http://mandrakeusers.org/index.php?showtopic=23073

I have the 1st public 1.0 release of Win4lLin Pro installed with Win2000 Pro SP4 but I haven't had time to tweak or play with it much yet. Running the trial version right now and it acts like a 1.0 release, but it shows enough promise that I'm buying a license and will keep up with it as it improves. A newer version has already been promised for early April. No slick GUI installer like Win4Lin 5 for Win9x yet, CLI only install, but the Win4LinPro rpm installed and loaded Win2k with no problems.

If anyone shows an interest, I'll keep updating this thread as noteable new features are added to Win4LinPro and maybe give my impressions of it as it develops.

It's becoming clear that there is much work to be done on Win4Lin Pro. The guys at Win4Lin are pretty friggin' good though, and improvement will likely come quickly. Read the statement below.

For those who want to try Win4lin Pro, 2 things to keep in mind:

1. You can install and run Win4Lin Pro alongside an existing Win4Lin 5 installation, no need to uninstall your working Win4Lin/Win98 setup. You can leave the whatever kernel you have installed. Win4Lin Pro dosen't care what kernel you run, but a 2.6 kernel is strongly advised. Don't know yet if you can run sessions of both versions of versions Win4Lin at once. I'd think it might be possible.

2. There is a a free trial version that's good for 13 days available as .rpm or .deb here:

http://www.win4lin.com/index.php?option=co...=120&Itemid=153

It's the 'normal' full version and runs normally except for a message at every startup or installation until the 13 days are up or you buy a license and register. I intend to try it as soon as I have time and then decide if I'm gonna throw down my plastic for it now or wait a while.

And for the technically curious, here's more tech info from Leo Reiter, Vice President of Engineering at Win4Lin, who is very responsive to questions posted on the Win4Lin-users mailing list:

I am now sending this note explaining the state of filesystem integration in Win4Lin Pro, as well as where it is headed.  I will also give brief background information as to why it's in this state, and why it will be different than on Win4Lin for Win9x.

 

A little bit of background: on the Win 9x product, we achieve full filesystem integration due to the system bootstrapping with DOS.  We simply hook into the DOS file functions and direct them over to Linux.  That allows us to load Windows.  Once Windows is loaded, our specialized IFS (installable file system, based on a normal Microsoft interface) takes over and lets Windows talk to Linux pretty much directly through the use of a "redirector."

 

*Obviously* Windows 2000 is different, otherwise we would have done the same thing.  There is no DOS to bootstrap Windows anymore, instead being replaced by various stages of "loaders" and DLLs that *only* know how to talk to NTFS and FAT filesystems directly on disks.  Microsoft explicitly says that boot-time IFS's (see above) are not supported on 2000/XP.  So to boot Windows, the only alternative would have been to

replace the native loaders with our own versions, and load everything ourselves.  That would have resulted in us being locked in to specific versions of Windows, service packs, etc.  We decided to just let 2000 load itself as a cleaner (and more compatible way to do it.)  Obviously this requires a virtual disk image.

 

The plan was to have an IFS that could store the Documents and Settings directory, under which 2000 stores all user information (documents, profiles, user registries, user application data, etc.) directly on the Linux filesystem.  Unfortunately this did not make the beta cut, and we considered it to be too risky to roll out immediately without a really

good beta test.  We expect to be rolling this out in the next few weeks, as Jim Curtin explained earlier.  We are also looking at adding c:\Program Files to the IFS structure, so that all you will need to have in your disk image is the Windows OS and drivers themselves, as well as the system registry.  This will still mean approxmately 600MB worth of a single image file, but all applications you install and so forth will live directly on the Linux filesystem.  What's more, we will have the capability of making the "system" image read-only, or better said, non-commit, so that viruses and the like will never screw up your base system unless you allow access.

 

Now, we are hoping for 2 things: 1) we will find the magic formula needed to put the entire 2000 installation on the IFS, eliminating the need for the image, or 2) Windows 2003 does not have this limitation, so at least future versions will be better.  Still, in the grand scheme of things, a 600MB image is not that painful to backup these days with large hard disks and writeable DVDs.  Besides, you can gzip it down to 350 or so megs if you want to put it on a CD.  The really good news is that the image itself won't need to change very often unless you are making changes to the system registry or installing software.  What we have today is obviously an interim solution.  We modified the open source samba server to allow it to easily run as a non-root user so it can be launched by the Win4Lin Pro session itself when needed.  On this samba server we enable $HOME and $HOME/My Documents (or MyDocuments or Documents, whatever your desktop defaults to.)  We also enable filtered and raw printing through the same mechanism.  Today your documents will automatically live in your Linux filesystem, which obviously is not the only thing that eventually should, but they are pretty important to be able to share with Linux apps.  You can also easily access your home directory through \\HOST\HOME without having to

set up a Samba server yourself - it's all automatic.

 

As for performance, please keep the following in mind: our development will always focus on functionality first, followed by performance optimization.  So the first cut of the "full" IFS I mentioned above will work very well, but may not be the fastest thing.  We will quickly

optimize it to realize the type of FS performance you see on Win4Lin 5 after the base functionality is rolled out.

 

As for the disk image today, there are 2 things to keep in mind: 1) it's actually very fast, we did a lot of work to optimize it (like caching, etc.)  In fact, it's probably one of the fastest components of Pro right now - we were able to benchmark faster than native speeds using tools like Fresh Diagnose, where we saw it report 7200RPM IDE write speeds on

4200RPM laptop disks for example.  Second, the image grows dynamically, so it won't be just a huge blob on your filesystem unless you install lots of software on it.  Currently it's limited to 4GB, but that will be corrected.  Also, being able to store program files on the Linux FS will probably mean that the disk image should never exceed 600-900MB or so

anyway, so the 4GB limit will become a moot point very soon.

 

As for performance versus VMWare, we have seen *overall* performance for most things to be as good or better.  However, VMWare will beat us (for now) on things like heavy computation, such as MP3/DVD encoding, etc.  We should beat VMWare in disk image performance however, for most things, so *overall*, at least for productivity/office-type apps, we compare very well in most cases.  There are many performance improvements in the works that will close this gap quickly, and you will see these in the next few weeks.  Again, we went for functionality first, then performance.  This is part of the reason we are selling it at a discount until early April.  We believe this product is quite usable (and have been told so very clearly by prospects and many testers), but we also understand that it needs a lot of improvement.  Most of that improvement will be rolled out by early April, when the price goes to $139.  We appreciate your understanding on this, and yes, we do appreciate your feedback.  We will do our best to satisfy everyone, but keep in mind that we are a small company and do not have unlimited resources.  It may take us some time, but we will do our best to make things perfect as quickly as possible.

Posted to the Win4Lin-users mailing list today:

Win4Lin Pro is now available for purchase on our store. We launched a mailing to our hundreds of thousands of Linux users last night informing them that we have made Win4Lin Pro available at a special introductory price of $99.95. This price will be effective through the end of March, at which point the price will increase to $139.95.

 

First of all, thank you for your patience and your support as we have labored to deliver Win4Lin Pro to you.

 

We are extremely proud of, and excited about, this new product. We believe Win4Lin will reach a whole new set of users with this release, helping us accelerate our goal of broader

Linux adoption by the mainstream IT community.

 

As insiders and valued users, we would like to share with you our expectations, give you some deeper insight into the product and, finally, ask for your continued support.

 

Expectations:

 

Win4Lin Pro runs Windows 2000 and XP. Windows 2000 and XP are much heavier than Windows 9x. You will see a performance difference between Win4Lin Pro and Win4Lin 9x. The majority of this difference is attributable to running Windows 2000 versus Windows 9x. That being said, we have identified a number of areas where we can gain dramatic performance improvements in Win4Lin Pro and we will be releasing these enhancements free of charge over the next several weeks.

 

If you have a need to run Windows 2000 applications faster, cheaper and with better integration with Linux than you have found in VMware, then Win4Lin Pro is your right choice. If performance is your highest priority, and your applications run fine on Win4Lin 9x, then you might consider staying with Win4Lin 9x or Win4Lin Home for the time being.

 

The Win4Lin Pro architecture is a combination of our 20 years of product development experience with Win4Lin and Merge, combined with our embrace of the QEMU project. We have been contributing to QEMU for some time now and expect to continue contributing and increasing our sponsorship of this very important open source project. We believe that our experience and value-add represents a significant value to Linux users needing to run Windows in a Linux environment. Integration into Linux host services file system, networking, cut and paste, etc, installation, QA and testing, usability, support, etc are all areas of significant value-add over the base QEMU project.

 

This is a 1.0 product. We will continuously improve the product. We have always valued the passion and expertise residing in our user base. We encourage you to continue to provide us feedback and identify areas you would like to see improved. But, please, we know about certain USB restrictions in both Win4Lin 9x and Win4Lin Pro. These will not be addressed in any short-term releases. Neither will accelerated Direct X. We welcome all other comments aimed at making the product better and more useful to you.

 

The introductory price is being offered in lieu of an upgrade price. Technically, there is no upgrade as Win4Lin Pro is an entirely new product. We are offering the introductory

price to our users as a way of saying thank you, as well as a way of offsetting any issues that you might have with respect to this being a 1.0 product. We are targeting a 1.1

release in roughly six weeks that will focus on performance and compatibility improvements, incorporating improvements from user feedback, as well as allowing the product to run better on less powerful systems.

 

Our product is differentiated from VMware in a number of ways first, of course, Win4Lin is offered at a much lower price. Second, our product is a Linux process, not a VM silo.

This allows us to leverage the host services of Linux and run in shared memory as a Linux process and allow management and deployment of Windows with Linux tools, resources and commands. Third, Win4Lin Pro requires no kernel hooks of any kind even run-time, as with VMware.

 

We believe the release of Win4Lin Pro effectively rounds out our desktop product offerings. Win4Lin Home establishes low-price point leadership vis-à-vis WINE. Win4Lin 9x is the fastest way to run Windows applications on Linux. Win4Lin Pro gives Windows 2000 and XP users the best way to run Windows 2000 and XP applications in a Linux process, integrated with the Linux file system and other Linux environment advantages.

 

FYI - We are targeting release of Win4Lin Pro Terminal Server in the first week of April.

 

Again, we thank you for your continued support. We hope that Win4Lin Pro meets all of your needs immediately. And, as always, we cherish and value all of your comments and feedback on how to make this product more useful to you.

This answers the questions about USB, DirectX, etc. It should be noted that At $139.95 after April 1st, Win4Lin ain't cheap, ($99.95 ain't peanuts either) though VMware is still more at $189.00, and it will very likely still kick VMware for speed.

So, what's wrong with using the provided source package? Why do you need an rpm?

One thing I learned the hard way foolin' with Linux is it's always best to install a package designed for your system i.e. rpms for rpm-based distros, .deb for Debian-based, etc. Problem is, if you install from source the package manager has no knowledge of what was installed. Apt, urpmi, whatever, can only properly keep track of and resolve dependencies and conflicts with what it knows about. So stuff installed from source (unknown to urpmi) can cause problems later with upgrades, installation of other software, etc. Chances are it won't, but it can, so why risk it if a package is available that will let urpmi do it's job?

Hey Hey!! I have clouds!

I made the full path edit in the script and ran it manually and it worked. Made a couple of other pathway edits, renamed the original 'default' config file in /usr/X11R6/share/xplanet/config to 'default.old', and everything works. Even the starmap file is working. I can't see the stars with the default screen sizing, but the moon comes into view so I think I'll leave the size as is, at least for now.

Now I'm just hope the crontab I put in runs the 'download_clouds.pl' script properly, but my logs overnight say it was running every 3 hours so I think I'm go to go.

Xplanet just makes such a cool background, especially with an clean, iconless desktop and transparent apps. My dad is here and just saw it running and it really knocked him out. Now he wants me to do it on his Win2k box. At 84 he's not up to tryin it himself. Ugghhh...seems setting xplanet up on 'Doze is kind of a bear.

Many thanks for the help!

THAC deserves a medal, his RPMS somehow clean out any errors, im so glad I found his repository

Yeah, his rpms are terrific. He's a friggin' Swedish Texstar. Use his repositories and a trouble-free upgrade to xorg 6.8.2 is just a 'urpmi -v xorg-x11' away. Or, if you urpmi his latest rpm of Firefox, it also does the x11 upgrade at the same time. Someone really ought to add his stuff to the easyurpmi link.

But no need to reboot, at least not just to upgrade x11. A logoff and restart of X is all you need. I'd think the new ATI drivers would be picked up then too, but maybe you're right a reboot is needed for that. I don't use an ATI card, so pay no attention to my rambling. I still use the same ol' 32MB Savage4 card I've used for 6-7 years, which runs glxgears perfectly at a screaming 140fps. But hey, I don't do games and it's always rock-stable, the most important thing to me.

I'll get to find out how good the new ATI drivers are next weekend. A friend and I are building him a quick new box with MDK10.1 and an ATI 9800 card. He'll give it a workout!

Thanks devries! I'll make the changes and try again.

Update: Still no go, but now I can't tell if it fails 'cause I can't get a cloud map or if it's still something else that's my fault. Supposedly, it's all set up, cron job and all. I guess if I login in a couple of days and have clouds I'll know I got it right.

exit command: killall xplanet

DUH! Why didn't I think of that??? I gotta be gettin' senile...

Still trying to get clouds...

I put the xplanet setup file (/default) under ~/.xplanet/config/default and use this startup command to set some options and tell xplanet where to find the config file at startup:

xplanet -config ~/.xplanet/config -transparency -random -starmap /usr/X11R6/share/xplanet/stars/BSC

Made a minor edit to the 'default' config file to tell xplanet where to look for 'clouds.jpg' (put under the "Earth" section of the config file as it should be so xplanet doesn't try to draw clouds on everything):

cloud_map=~/.xplanet/clouds.jpg

And edit the 'download_clouds.pl' script to put the downloaded 'clouds.jpg" file where I want:

# Set options:
#  - where to save the downloaded file (default is current directory)
my $Filename = "~/.xplanet/clouds.jpg";

Now to add the 'download_clouds.pl' script as a cron job. But I can't find kcron anywhere. KDE is messed up, and I don't want to bother to fix it. Basically, I just don't give a squat about KDE. So I just added a user cron job the old-fashioned way:

$ crontab -e
0 3,6,9,12,15,18,21 * * * perl /usr/X11R6/share/xplanet/scripts/download_clouds.pl

...which wrote the crontab in /var/spool/cron. OK, I should be set, right?

Still no clouds. Hmmm...maybe it's not finding the cloud file, so I try running xplanet from a terminal to check for error messages. But xplanet starts without complaint and no errors on exit.

A check in ~/.xplanet shows no cloud file has been downloaded there. Jeez...no wonder I still don't have clouds.

So I try manually running the script from a terminal:

$ perl  /usr/X11R6/share/xplanet/scripts/download_clouds.pl
Using http://giga.forfun.net/clouds_2048.jpg
Downloading...
Download not available, trying another website

Using http://enekyn.com/xplanet/clouds_2048.php
Downloading...
Download not available, trying another website

Using http://myhome.hanafos.com/~hyoungkee/cloud_data/clouds_2048.jpg
Downloading...
Download not available, trying another website

ERROR: Tried to download the file 3 times, but no servers could provide the file

Repeating the script tries various the various sites listed in the script but never finds and downloads the cloud file.

So anyone know what's up with this? Have I messed up on syntax or something? Or is today just my lucky day and the cloud images aren't available today?

But hey, at least I haven't trashed perl again somehow.

BTW, anyone know an exit command for xplanet?

Beta testing went very well, and it appears to be ready and will be available on Feb 23.

For more info:

http://www.win4lin.com/index.php?option=co...d=64&Itemid=125

http://www.linuxgazette.com/node/9949

Cost is $99.95. There will probably be a cheaper deal for current users wanting to upgrade, but I have no info on that yet.

This should be a killer product and take away a lot of potential customers from VMware since the main reasons for most people to ante up the $300 for VMware instead was it ran not only Win9x, but Win2k/XP and no kernel patching was needed. Win4Lin Pro removes those objections and keeps the terrific speed and stability of Win4Lin, which VMware never could match. For many who must run Windoze apps - period, no choice - and Wine or Crossover Office can't cut it, this is a major improvement over Win4Lin 5.

As for me, I really don't have a good reason to upgrade (well, depending on just what features it adds), but probably will pretty soon anyway just to support Win4Lin. They've long made a terrific product to help Linux converts and have the best customer support I've ever encountered.

I'll post more as I find out, such as does Win4Lin Pro add long-awaited support for USB and DirectX?

[moved from Emulation by spinynorman]

I finally gave up on fixing perl. Who knows what happened to screw perl up so bad. But I was wasting more time trying to fix it than I would reinstalling, even though that's an awful lot of work on my box. So I've already done the basic reinstall, but still have a couple of day's worth of config and stuff to do. This is a pretty complicated installation for a desktop workstation. Anway...

change the first line of the script to where perl is installed.

I'll change the path in the 'download_clouds.pl' script and set it up as a cron job WITHOUT using Webmin. Anything else I should know, for instance where you put the script, etc. I want to make no mistakes this time...

Let urpmi sort it out for you.

urpmi is broken (see my post again) so that's not an option.

you need to change the first line of the script to where perl is installed. I have /usr/bin/perl. Perhaps you have /usr/local/bin/perl.?

Even after reinstalling the perl package (see prior post agian) there is no /user/bin/perl, /'usr/local/bin/perl, etc.

which perl

no perl in (/usr/bin:/bin:/usr/local/bin:/usrX11R6/bin: etc,etc)

Please tell me I don't have to reinstall...

Have you made the file 'executable'? What do you get when you type perl download_clouds.pl?

I did chmod 'download_clouds.pl' script to executable (100755).

$ perl download_clouds.pl

was resulting in something like:

 $ /bin/bash: bad interpreter:  no such file or directory.

This happened whether root or a user.

However, and sorry this is going way off-topic, but things have gotten MUCH worse. The machine does boot, including starting X. For ease of use for other users on this box, I use the graphical login manager. It fails - acts like it's gonna work, then just kicks back to the login screen, no matter the WM you choose. I can start Fluxbox by opening an aterm and starting Fluxbox from there. Of course, I am able to log into failsafe or just boot to a console. But it many things are broken, like MCC, GnuCash, OO, etc. urpmi is broken, but rpm works.

During the boot sequence there's some error messages along the way about Shorewall failing (no biggie there, I can deal with that later) and

/etc/rcd/5/s15cups/blahblah

and:

/usr/bin/perl: bad interpreter: no such file or directory.

That's the key error. In fact all the broken apps are related to various errors like:

/bin/sh: /usr/bin/drakconf: /usr/bin/perl: bad interpreter: No such file or directory.

Oddly, I also get a bad message during shutdown about:

stopping harddrake excvp no such file or directory.

Never saw that before...but must be related to the various application failures.

All this is my own fault for f*sking around where I didn't quite know what I was doing.

Basically, I think it must be the result of me stupidly tryng to setup 'download_clouds.pl' as a cron job. I was doing this in Webmin, partly 'cause it's easy but mainly 'cause I just happened to have Webmin open anyway when I thought about doing it. It failed to save the new crojob, giving me a message showing the contents of an unspecified file beginning with the line:

"Do not edit this file. Reinstall the master"

or something like that.

After quite a bit of Googling and fooling around without finding a really relavent answer about what do do about the perl errors about "bad interpreter...", I found the above mentioned file in /tmp/.webmin and just deleted it. Then I checked etc/crontab, /etc/cron.d, etc/cron.hourly - everywhere I could think to look for any reference to the failed cron job and found nothing. Then I uninstalled Webmin, hoping that would be the end of that. Unh-unh.

Just for grins, I reinstalled perl, but no help, not that I expected it would.

I've also deleted 'download_clouds.pl' from where I'd put it. It is no longer anywhere on the system.

Bottom line is I've seriously broken my system and don't know how to fix it. That's the bad news. The good is I bet it can be fixed fairly easily if someone can tell me what to do. At this point, I'm stuck, and reduced to posting this from my laptop and not looking forward to reinstalling at all.

I haven't been able to get cloud maps working with Xplanet since I upgraded to 10.1OE. I downloded Oliver White's perl script 'download_clouds.pl' for getting maps from random servers, but I can't seem to get it going and there's really no specific instructions on his website.

Also, I start Xplanet with these options:

$ xplanet -config /home/tim/.xplanet/config -transparency -random -starmap /usr/X11R6/share/xplanet/stars/BSC

...but I don't get stars in the background like I did before either. Obviously, my brain is not engaged and my smarts are slipping...Can anyone give some step-by-step info on how to do it?

I'm no networking guru and I'd have to do some checking, but I think there's something not right with the second line, something with 'POSTROUTING' part doesn't seem right.

Why not try using the Shorewall module in Webmin? I know, "What? - a friggin' GUI?" - but it really does make setup of IP masq and NAT in Shorewall easy.

I'm running GnuCash 1.8.9 from the 10.1 Power Pack, but not the 64-bit edition. No problem here.

Other than 'install all the fonts you can find' I don't kow what to tell you except you should try posting your question on the GnuCash users mailing list. Pretty good help there, often from the developers themselves, and you should have an answer shortly.

GnuCash is great. Once I found it, I knew I'd be able to kick my M$ habits. Used it trouble-free for 3 1/2 years now.

After reading all the thoughtful comments here, I still say the 4 points I made in my post on page 3 of this thread wrap most of this up, and that most comments since are a further (but good) expansion on those points. I could quote and reply endlessly to what has been written here already.

Here they are again, in a nutshell, but with added comments 'cause you guys have made me think a little more about this:

1. I said we've gotta have pre-installed Linux systems in the stores, with a choice of software.

I'll add that Linux has also gotta have support from hardware manufacturers. Sorta chicken-and-2 eggs situation. I think games are important, but not the critical factor in Linux adoption for the majority of users. With Linux numbers, the games will come.

But it won't do much good to have people buy a Linux system and find out the new printer-scanner-copier or mp3 player they bought to go with it is Windoze-only. That buyer will never get as far as playing games. That's now a pissed off customer, cussing his choice of Linux - taking the Linux box, not the all-in-one unit - back to the dealer and going back to the Dark Side with another XPee box that "just works".

2. A majority of businesses cannot use Linux desktops because of custom, propreitary Windoze software.

I'll add here that if people don't get used to seeing Linux at work, they will naturally be less inclined to try it at home. And I see no Linux answer to lotsa Win-based business software happening for a very long time.

Like I said before, running this stuff with Winshaft as a guest OS via Win4Lin has some usefulness. Win4Lin is really very good and the situation there is improving with Win2000 and XP support coming very shortly, and with no need for patched kernels anymore. VMware simply not a contender - much too expensive and too slow anyway. It was never even intended for such purposes. But either of these methods is an added expense to Linux adoption, too complicated and just too much hassle for most businesses/users.

The only real solution I see here is a very, very capable version of Wine - kinda Crossover Office on steriods, SuperWine - so existing apps can be installed normally and run perfectly on Linux workstations. This is CRITICAL, the importance of something like this to widespread business (and to many home users) Linux desktops cannot be overstated. It would necessarily have to include support for .Net, VBS and DirectX for many such apps to run. Problem is, if such a version of Wine were available it would of course reduce incentives to make more native Linux software...sorta back to the chicken and its eggs...but if enough Linux machines get out there, IMO this would slowly change and native Linux apps would eventually win out.

3. Internet Exploder is still a absolutely necessary evil for many businesses and individual users to use sites they need due to idiot webpage designers.

This may be the most difficult problem of all to overcome for many users. Like I said, my business is a good example of IE lock-in, and I see no solution for a long time and believe me, I've looked. It's not just changing the websites themselves, but there are web-based apps that companies have made large investments in programming, distribution of the software (both custom IE plugins and desktop apps for working with websites) and training. This is gonna be very tough, probably tougher than overcoming the desktop applications problem. Again, a SuperWine to run a fullly-enabled IE is probably the only answer. We need to go balls-out on Wine development. Those who claim "cop-out, stay native-only" are not facing the realities of the situation.

4. The average user is going to use what seems familiar to them and is unable to do even very simple computer tasks. This effectively blocks migration to other platforms.

Well, this has been discussed for pages in this thread. Simply put, if Linux distros really want to win over home Windoze users, it will have to be just as simple (or even more so) than Windoze. Happily, in many respects it already is, but the problems users have with things like installing video drivers show it still has a ways to go.

For new Win>Linux converts, full environments like KDE and Gnome will, unfortunately, have to an offer an option to nearly duplicate Windoze in look, feel, layout, names, etc. right down to Start, MyComputer and MyDocuments. The transition will need to be nearly seamless. Better may not be better if it's different, 'cause for most users Different is Bad.

I think there's no question that right now Linux generally isgood enough. I've used it exclusively at home for 3 1/2 years now and it's obvious it is for me. It's a matter of getting a few things to point-and-click simplicity so Linux can reach a critical mass of users, having a large enough base of "Joe Blow" users and businesses to force support from that 3-headed monster of software and hardware manufacturers and web designers.

Last, some side notes, sorry if they're a little OT. To me Mandrake is still the best overall distro, at least of those few I've tried. 8.2 was an anvil. 10.1 seems rock-stable, and pretty much worked right off for me. I have only a couple of details to sort out, kdeinit hanging being the only serious one. It's still a nice mix of being both hands-on enough to use much like Slackware if you know how, and yet GUI-driven enough for many XPee refugees to get a handle on it. Hopefully it will continue this and not go the way of Suse or Linspire, but it's starting to lean a little. I'm sure it's pretty hard for the developers to balance the wants/needs of two very different types of users while maintaining a true open source philosophy.

Oh, and I've had no problem with XP SP2, it's great. Just gotta install your software *after* SP2, then disable about 15-20 unecessary services, that's all. And keep your AV up to date, run WinshaftUpdate every week or so, and run AdAware and Spybot Search & Destroy every week. Don't forget to keep your firewall up, defrag now and then, maybe run RegClean occasionally...damn, be careful of those email attachments!! Then Wincrap runs OK. It rocks.

Nothing to it really...it only takes about 10 times as much time and effort as keeping my MDK box right...or my Netware 5.1 server than hasn't been rebooted in a year, or the RH 6.3 router box that hasn't for over 3 years...

I've used VNC on both Windoze and Linux a lot the last couple of years with no problems. Fast and stable. But you didn't say what kind of connection you are trying to VNC through. LAN, internet...? If a LAN with static IPs, try the previous suggestions. If you are trying to log into a box using a DHCP connection (like most cable services for instance) then you have no static IP to connect to. You'll need something like GotoMyPC.com to provide a static IP.

artic:

this (requirinjg IE) is not really a problem of the internetbrowser you are using but the webpage-coder who was lazy/ignorant. everything that can be done with ie can be done with opera, netscape, mozilla, firefox,... it just needs to be coded.

I partly agree. I disagree that anything that can be done with IE can be done with other browsers. Opera, Firefox, etc will never (thank God) do DirectX, VBS or custom-made IE-only plugins.

I totally agree that it's lazy/ignorant webpage coders. No browser should ever need use any of that crap. Any coder that uses it on their pages should have a Mandrake CD stuffed up his butt. But the reality is they do, so IE6 is, at least for some of us, an evil necessity not about to go away soon.

...you and almost everyone else underestimates the ability of people to learn some things.

Nah, you misunderstand my point. Most people certainly aren't that stupid, they have the *ability*. What they lack is the *willingness*. Short supply of that.

how can such people survive? how do they manage to open a can, make some coffee, how to use a spoon (and for what!?!) and how do they change clothes? oh my god....

Hey, day-to-day life is pretty challenging for a lot of people...I struggled for years about which end of the toothbrush to grab...

BTW, I stand by what I said about most Windblows users not being able to do simple tasks like create/copy folders/files. Betcha if you put 100 randomly picked, ordinary Winshaft users (including those that who claim they use computers everyday at work) in front of a simple Win98 box and asked them to turn it on, create a folder, put some files in it, and copy the folder to a different location, over 50% could not do it without help. Users like that, who don't want to *learn* anything, "Just tell me what to click to get my email..." are what Windoze went after and successfully got. Well, good for you, Billy Bob Gates, and how about a smalll loan?

But real-world computer "expertise" like that makes migration to Linux or Macs seem awfully difficult, no matter how easy it installs and configures itself.

UrpmX is 10 times what it was 2 years ago.

Is apt still better? I think so, but not by much.

urpmi has always worked very well for me. I agree there's not much to choose from between apt vs urpmi anymore.

One reason I've stuck with Mandrake is urpmi and that I still think rpm's will win out as the default Linux ".exe" equivilent of the future. For example, I just bought Symantec's pcAnywhere 11.5, partly because it offers Linux support now (to a point anyway). It only supports installation on "Red Hat Linux, Red Hat Enterprise Linux WS/ES, SuSE Professional, or SuSE Enterprise Server" - all rpm-based distros. I don't see anything changing the situation. More likely, eventually .deb or source-based distros will eventually need to offer rpm installation, as least as an option, to support commercial Linux software.

(And no, I haven't tried to install pcAnywhere on MDK 10.1. I use it to connect to Win hosts at my office. Has Java Browser-based support only for Linux remotes right now, and I don't need or want to install and run the Linux host application on Linux here at home. But I may try just to see if I can.)

I recently thought I was going to have to finally reinstall mandrake after what??? 2 years

I ran the same 8.2 install for almost 3 years, through many, many upgrades and modifications. Never became unstable or f*scked up in any way despite my not knowing what the heck I was doing much of the time. I could've used the same 8.2 install happily for another 3 years, but decided it was just time to make a move to 10.1, go to Xorg, 2.6 kernel, etc. My club membership let me ftp the 10.1 Power Pack, so why not? And, like going from Win98 to Win2000, it has been a very worthwhile upgrade, if not really necessary.

But to get back to the thread subject...I think there's 4 things holding Linux back from catching on with desktop users, mostly unrelated to Linux itself as far as it's quality or capabilities. Not that any of the rant below is new, but in no particular order...

1. You can't run to Best Buy, CompUSA, etc and grab a desktop or laptop with Linux pre-installed and assorted software to go with it. So most people are unaware they have another option, except for a few who know there's something called a Mac. Most people think of Apple as the Beatles' record label.

2. Business users have a large investment in licenses for Windows, Office, etc, but that's not really the problem. It's the much larger investment compaies have in custom-made Windoze software, including my company's, as an example. Our proprietary managment software is probably *never* going to be ported to Linux. Emulation such as Win4Lin is the only solution for these apps. That's OK to a point, but there's other problems...

3. Many businesses, and more all the time (like mine, again, for example) have to do business over websites which *require* IE (6 SP1 or later) in order to run features using DirectX, VBS, or IE-only custom plugins. Firefox, Opera - no matter how much better they are than IE - are simply not an option. Gotta have Windoze and IE6. This is the unsurmountable problem in moving my office desktops to Linux.

4. Linux and Mac both have the same problem trying to get people to switch. People are comfortable with what they're familar with, and Windoze is what people know, at least what little they do know. Keep in mind that most Windoze users don't know how to do even simple things like create and copy folders. It was very hard for most users just to learn doing the most basic computer things, like how to turn it on, go on the web or do email. They don't want to fool with starting over with a new system, even if such doing simple tasks actually are functionally basically the same (click "Check email" button, click "Bookmarks" button, etc). It's too scary to them. So they stay with Windoze, IE, Outhouse Express, stuff they at least know the name of and where a few buttons are.

Think of it as kinda like changing from a car with a steering wheel and pedals to one with a joystick to steer and buttons for brakes and gas. It would be a better system, offering easier, quicker, even safer and more accurate control of the vehicle, but how do you convince people who've used a steering wheel and pedals all their life to buy newfangled joystick cars and learn to drive again? Just too scary for most. They'd rather stay with the familiar than learn something better.

All that said, I still think Linux is definetly the future, we're just gonna have to be really patient. The rest of the world will forge ahead of the U.S., while we try to legislate the M$ monopoly (a U.S. company with global domination) into permanancy - which will of course fail and fail ugly, costing the U.S. in ways we can't yet imagine...

But then, what do *I* know?

Checkinstall is a great tool, but not intended for building rpms for distribution. It will yield rpms that you can fully uninstall or reinstall on **your system**.

If you really want to get into making rpms for others more work is necessary. Check out the link Tymark gave or do some Googling - you'll find plenty of info is available.

An update for those who care...

Netraverse has put out the call for beta testers, with testing to begin in 2 weeks. Win2000 will be fully supported 1st, but XP (at least some of the many versions out there) is supposed to be working pretty well already, and will also be fully supported soon.

NO MORE PATCHED KERNELS NEEDED! This will run on the stock kernels as suppled by your distro of choice. Just install and go...