IBM Desktop Linux Pledge, One Year Later

Blue writes "It's been more than a year since the bold announcement from IBM that they planned on dumping Windows for Linux throughout the company. InfoWorld is reporting that not all is well with IBM's desktop Linux push. What went wrong?"

Comments

IBM wrote a redbook on the topic

[Dancin_Santa]

The redbook is about 200 pages, but it talks mostly about the migration of desktops by discussing server administration techniques rather than focusing on enabling users to upgrade painlessly.

Linux (nay, any OS) migration is tough work for the administrators *and* the users whom it affects.

It's not a surprise that they weren't able to do it.

How Disappointing

[bigtallmofo]

If IBM of all companies is developing their internal applications to require Internet-Explorer dependent technologies like ActiveX... What does this say about their commitment to Linux?

Hopefully this is just a case of a huge company's left hand not knowing what the right hand is doing. But still, this is very disappointing.

Re:How Disappointing

[fm6]

Linux isn't the only platform that ActiveX doesn't run on -- it hardly runs on anything at all.

Except that "hardly anything" is 95% of the users!

This story isn't about Linux, it's about how IBM fucked themselves by not thinking.

Sure, they've done stupid shit in the past. Before Gerstner, upper management even refused to use email. But that's kind of beside the point. We're all stuck with an overdependence of Microsoft products. IBM, at least, is trying to make the change. And the difficulty of doing that is what we should focus on, not pointing fingers for past mistakes.

Re:It just won't work.

[tibike77]

For a "FP attempt", the poster sure hit the right spot (by accident?)

If you read the article, the main problem is that some frequently used internal IBM applications only run (so far) on IE. So the users end up using Wine or VMWare (heh, sounds like my apartment mate here).

So, the problem it's not that it won't work at all, it's just that *somebody* lacked the foresight to migrate all day-to-day applications to that "customised Red Hat Linux distribution" IBM uses. And now, the problems finally hit the fan.

People don't fear change

[Kipsaysso]

People fear loss. What is stopping people from making the switch is the fear of losing the control of their computer that they have taken so very long to cultivate. Not to mention all the internal documents that probably have been created over the last few years using the .doc extention. Who wants to go through years of porting old files?

Re:It just won't work.

[Anonymous Coward]

They don't care whether it's Windows, Linux, FreeBSD, or whatever. They just want to get work done. If they OS enables them to do so, they'll take it.

Therefore, it's just the GUI, and you can make it to as close to Explorer as you want.

Re:what went wrong?

[Stevyn]

Well, the article talked about problems of trying to convert people used to IE and Lotus over to a platform that doesn't run that software well.

What they should have done first is switch everyone over to firefox/mozilla and find an alternative to lotus that runs on windows and linux. Then after people are comfortable with that, try to switch to linux.

I think a switch of this magnitude must be done slowly, or else tech support isn't going to know what to do or be able to handle the "where's my icon" problems for tens of thousands of people.

Problems are normal

[Neo-Rio-101]

Well, they made a bold push to do this by the end of 2005... so they have 1 year left to beat their own goal.

I'd say the problems they are experiencing are due course when migrating to a different platform. Sure, IE wrecked browser standards and many people had to play fiddle to Microsoft and write IE compatible pages. That's going to take a while to fix and there will be problems.
Porting some of their other apps in going to be a long and painful process.

I can't see any transition to another operating system being a smooth transition. For sure there WILL be DIFFICULT problems for them.

This is just part and parcel of development, and not a reason to give up on Linux AT ALL.

Good news is that Linux will be a better platform once their present-day trials are over, and they complete what they set out to do.

Maybe 'cause they can't read Slashdot

[Pedrito]

Here I am using Firefox to TRY to read the comments of other posters, but the comments bleed over into the sections list and comment separator bars are partway over the text above and below and I can't tell what people wrote.

I would think that Slashdot, being such an open-source advocate, would at least make their page render properly with the most popular open source browser.

But if Slashdot can't be bothered to do it to their page, which is their entire business, how can people expect IBM to do their web-based internal help support which isn't really a source of income for them?

Internal standards

[tjwhaynes]

If he wants open standards based computing, then he had better start mandating that all web applications are developed so that all relevant browsers can use them.

Actually thats precisely what the internal standards do say. I find that waving them under the nose of the respective developer is quite instructive. And occassionally successful - several web apps which used to be IE only work seamlessly under Firefox/Mozilla these days. The number of web apps which are IE-only is shrinking fast.

Cheers,
Toby Haynes

Re:It just won't work.

[Anonymous Coward]

So, the problem it's not that it won't work at all, it's just that *somebody* lacked the foresight to migrate all day-to-day applications to that "customised Red Hat Linux distribution" IBM uses.

The problem started before then. It's that *somebody* lacked the foresight to procure software that didn't lock them into a particular vendor.

Re:Tried to drink the blue koolaid…

[dedazo]

We recently had one of our in-house developers set up a Linux workstation (Debian unstable) to test a PHP/MySQL application that was primarily developed for Windows. He was gung-ho at first, and then as the weeks passed and no updates to the bug tracker were made I went to talk to him. He was frustrated beyond belief. He couldn't get Apache to work quite right and he was having a lot of problems getting MySQL to behave, never mind the fact that he had to use a plain text editor to work. CVS (running locally in his box) was a mess.

Instead of fixing things for him I showed him what the problems were to the best of my ability - I'm a 6-year Linux "power user" if you will but far from an expert, so the rest of the stuff we just Googled. Eventually he got to the point where he was coding, doing his daily check-in and builds and demoing the app, which is working pretty well now. His changes are then backported to the main Windows trunk to ensure everything works. We're thinking of also supporting BSD at this point.

It just takes time. Even the most savvy PC users will have problems. I can't believe IBM would be that different, especially when trying to move their less technical folks to a brand new environment. Just give them time.

Re:more than insightful

[Zontar The Mindless]

I had my current desktop system built for me in November. Came with Windows XP, which I promptly got rid of in favour of Windows 2000 Pro. In spite of the fact that all the hardware is commodity stuff, and the integrator had thoughtfully supplied me with nearly all the necessary drivers for Win9x/NT/2K/XP, it took me about a full day to install Win2K from scratch, run Windows Update, install drivers, configure hardware, *reboot about 35 times*, install software, etc.

This past weekend, I ditched Windows for SuSE 9.2. The installation itself, including time for running the SuSE Online Update to get their latest patches, was somewhere between 90 minutes and two hours. Number of reboots: *1*. (Yes, that's a "one".) At the end of this time I had a completely functional system, and all hardware had been autodetected and configured correctly. Number of drivers I had to load manually: *zero*. Number of trips to vendor websites to obtain drivers: *zero*. Number of apps I had to install separately before I could get any useful work done: *one* (BitKeeper). Number of apps that came with the distribution whose Windows equivalents either cost bucketloads or simply don't exist: *dozens*.

In fact, it wasn't until a day or two later that I printed out something, then realised that I'd not ever installed the printer. It was just there and ready to use. On Windows, installing that same printer took about an hour of loading drivers and fussing with the configuration to get it working properly.

I had/have exactly three issues:

1. KsCD didn't produce audible output until I told it to use direct access mode.

2. I had to change permissions on a bunch of email and other working files that I copied over from the Windows partition before Mozilla Mail and other apps could access them properly.

3. BitKeeper's not sending commit emails to our developer list. I'll probably have to get a bit of help configuring it and/or sendmail.

Executive summary:

1. "Everything just worked" rating >= 99%.

2. Anybody who says that Linux isn't ready for the desktop is misinformed, or a liar.

BTW, I still have two Windows boxes on my LAN (at least one of them will be getting converted to Linux as soon as I have time for it). When something goes wrong with one of them, I drop in a Knoppix CD to find out what's wrong and fix it. And when I use one of them remotely, the interface already seems clunky and counterintuitive, in spite of the fact that my reflexes aren't yet properly trained for the Linux desktop.

How much do I miss Windows? Well... I'll probably finish copying some miscellaneous files over from the the Windows partition and reformat it this weekend.

Re:It just won't work.

[Ogerman]

sounds more like somebody took the most cost-effective option available that did the job at hand

Designing with web standards is the cost effective option. Obviously they hired the wrong people to write their internal apps (or had nobody to give them proper guidance) and now they're paying the price. It is almost always more cost effective long-term to maximize your future options. Flexibility yields efficiency. In the case of web standards, it's cheaper even in the short term to develop with strict conformance because this methodology gives you a way to test the results. "It looks right in IE" works until something breaks or IE gets updated.

And cut the crap about "zealots." If you don't know what you're talking about, don't even bother posting.

Re:It just won't work.

[cofaboy]

Can't help it have to bite sorry folks.

How many different desktop distributions would you use across an enterprise? Oh just the ONE, oh look that means all the apps that you need will just work!

Retraining for the desktop is not that big an issue, a quick class with every body in that building section and hey presto one week later they'll have forgotten about that, sorry whats that other office suite called?

Retraining costs = done once.
License costs = forEVER