Linux: Ready for Prime Time?
The jury is out on whether smaller shops can actually work with open source operating systems.
- By Bill Heldman
I've always been a Microsoft and Windows bigot, and I've suffered for it -- big
time. Bosses have chastised me for always recommending a Microsoft solution
when there were other companies out there whose software often did the same
thing for less. I've paid dearly for being pro-SQL Server when in the camps
of the Oracle faithful. I've been denounced and jeered at by Linux, Mac, Unix
and even mainframe stalwarts. I've suffered the arrogant sneers of software
engineers who simply couldn't accept that Microsoft can write great software.
I could easily roll with all those punches, until Microsoft came out with Vista.
First of all, it's priced beyond of the reach of the average Joe's pocketbook
-- anywhere from $199 to $399, depending on the edition. And is there really
enough bang for the buck to be gained from Vista, aside from the fact that it's
the sexy new OS on the block that lets you spin open windows on their sides
or put in an administrator password at each system change (which Linux requires,
as well), and many other proprietary features?
Microsoft says it's all about the end user and the "user experience,"
but I think Vista is going to confuse the average business-class PC user even
more than they already are. On top of that, Office 2007 is equally expensive,
and I believe it will be equally confusing with its ribbon interface and layered
menus. Training companies everywhere must be thanking Microsoft for the "gifts"
of Office 2007 and Vista.
These factors, coupled with the fact that Vista requires some heavyweight hardware,
make me think Microsoft isn't considering its user base carefully enough.
One of my computer technology students is a huge fan of Linux. He has experimented
with so many different versions that I now have a "Linux lab" crowded
with computers running a veritable Heinz 57 of varieties. I have Damn Small
Linux (DSL), Linux XP, Fedora Core 4 and 6, SuSE 9.2 and 10.2, Mandriva, Ubuntu
in all its iterations, and others we have yet to even try.
This Linux-crazed student also pointed out www.distrowatch.com,
a Web site that tracks the popularity of various Linux versions (although I'm
sure there are tons of other Linux-specific sites out there -- and let us know
if you have a favorite you'd like to share).
Some of the Linux operating systems are nice; others aren't so hot. My student's
enthusiasm, however, got me interested in seriously exploring the viability
of Linux alternatives to Vista.
What if you have
a favorite Windows app you need to run? This is where Wine,
the freeware compatibility layer software for Linux, comes
into play (find it at www.winehq.com/).
With Wine, you can run all those Windows programs you can't
live without on a Linux host. Not everything runs well (or
at all) in Wine, so it's worth testing various apps to see
how they perform. -B.H.
I began my search in earnest with a couple of magazines from the U.K. (Linux
is really hot in Europe). One magazine got deep into the specifics of installing
and running Linux. Another focused on Fedora Core 6, another on SuSE 10.2. Each
magazine came with the installation code on disk. Those types of magazines are
an easy entrŽe for people who have heard of Linux, are comfortable with
operating system installs and want to broaden their horizons.
While I have a lot of experience with Microsoft server and personal operating
systems and application software, in no way do I qualify as a Linux expert.
So I bought both magazines and secreted them home to read from cover to cover.
I was particularly attracted to the fact that most Linux distributions came
with several hundred free software packages -- including OpenOffice 2.0. One
of the magazines said OpenOffice has between 80 and 90 percent of Microsoft
Office's capabilities. It also includes Evolution, an equivalent to Microsoft
Open for Business
After using OpenOffice for a while, I found that those 80 to 90 percent equivalency
claims were both true and false. In some cases, OpenOffice has functionality
equivalent to Office 2007, like the ability to save a document as an .XML or
.PDF file. Other aspects were less equivalent. For example, while Evolution
is much faster than Outlook, it doesn't appear to be able to connect to more
than one account like Outlook can. I wanted to connect to both my Comcast and
Hotmail accounts, but couldn't do so with Evolution.
OpenOffice also lacks an Access-like database, any OneNote clones and a plethora
of other Office features. In most cases, though, if there's something you really
need, there's probably an open source equivalent.
Some of the other attractive software components included were:
- GIMP, an open source photo editor that comes close to rivaling Photoshop
- Audacity, a voice recording and editing package
- Totem, a media player (that simply refused to play .WMV files)
- Nautilus, the Gnome interface's file explorer
There was a host of others, as well, all with goofy names and functionality
similar to what you'd find in Windows. Near as I can tell, the Linux development
camp has been mighty busy, staying up late at night fueled by high-caffeine
drinks and writing application code.
Houston, We Have a Problem
As marvelous as Linux sounded in those magazines, I ran into all sorts of installation
problems. In an earlier life, I played around with SuSE and liked it, so I installed
SuSE 10.2 first. When I tried to install it on my Windows XP box as a dual boot
(which the magazine said is nicely supported), the installation procedure whacked
my Windows boot partition. Apparently, this was a known SuSE problem, but was
allegedly corrected before 10.2.
I tried SuSE on a couple of other machines with no installed OS and encountered
GRUB (a Linux file partition) errors each time. In one case, SuSE installed
without issue and detected but didn't install drivers for either of the NICs
I had installed in the box. I had no desire to poke around Linux forums looking
for legacy NIC drivers when XP detected and installed them just fine.
Next, I tried installing Fedora Core 6. Installation went well, but then the
GUI wouldn't come up. I never got it to work -- not on any box upon which I
tried the install. The command line came up, and I had no idea how to fire off
the GUI after that.
Unlike Vista (which is a GUI OS with a command shell), in Linux you install
a command shell, then the OS installation sets up the GUI. This is somewhat
similar to the old DOS and Windows days. SuSE supports both popular interfaces
-- Gnome and KDE -- out of the box. The GUIs these days are fairly sophisticated,
but at the end of the day Linux is a command-line OS. Gentoo is one of the Linux
operating systems you can only get in a command-shell version
an Eye on ROI
will want to know if the return on investment (ROI) is there
for Linux. Until you can demonstrate to those holding the
moneybags that you've got a money-saving opportunity, you
won't have much support from above.
So, can you really save money with free software? The answer
seems obvious: Of course you can. But consider the following:
- There are significant costs involved in simply putting
people to work changing out the old OS with the new, even
if you use something like SMS.
- Training people to use the new software entails tremendous
operations and maintenance costs. They must be trained to
use the new OS and applications.
- Porting Windows apps may be a huge hassle. Ultimately,
my guess is you'll wind up reinstalling Windows on some
PCs in which Linux simply won't play with a needed app.
- Help desk and PC support costs will increase sharply
at initial deployment time, and will remain there for a
long period before they settle back down.
- If you really need a new application, chances are you'll
find it in the Windows world long before you see it in Linux-especially
if it's an oddity.
- Just because it's open source doesn't mean it won't break
as easily as a Windows app. Open source players mark their
code as "stable" or "unstable," so you
know what you're getting into. That doesn't mean the documentation
is well-written (or even there), however, or that you'll
be able to get updates later.
- Getting support at your level may be tough, as well.
If your Ubuntu server is broken, will you get the level
of support you could get from a Microsoft SPSS engineer?
That's uncertain, but Linux vendors require that you pay
for routine support.
These are just a few of the financial questions you should
consider before converting to Linux. It might be that Windows
isn't that expensive after all. -B.H.
Ubuntu was next on the list. First, I installed an older version (5.2) because
I had the CD available. The OS installed and came up nicely, but after going
through Ubuntu's automatic download and install updates procedure (which was
sweet), my screen resolution went from 1024x768 to 600x400 and wouldn't go back.
There were no other choices. I had giant icons on a 19-inch monitor.
You have to download Kubuntu if you want the KDE interface. I was originally
a KDE fan, but thanks to Ubuntu, I've switched allegiances to Gnome. Think Mac
versus Windows, and you'll have a basic idea of the differences.
At least Ubuntu completed and connected to the Internet. So I tried another
Ubuntu iteration called Ubuntu Christian, which includes superb Web-filtering
software. This installed completely and ran just fine, as did Kubuntu. I did
not try Xubuntu -- a small footprint rendition designed for older, slower computers
-- or Edubuntu, which is designed for young school kids.
It was clear to me that if I was going to run Linux at all, it had to be Ubuntu.
This is the easiest and fastest way to get up and running without a lot of hassle.
It's no surprise that Ubuntu is the top dog on the distro watches.
Linux for the Masses
As a strong Windows advocate, I had one question in mind while I was tinkering
with Linux: Will the typical business user be able to effectively install and
use this OS? Here are my overall impressions:
1. Until all flavors of Linux install completely and fully -- detecting
99 percent of hardware like Windows does -- the average business user can't
or won't use it. They simply don't have the skills to surf the forums looking
for drivers or go through complicated command-line extractions and installations
from a .TAR or .RPM file.
The Ubuntu Add/Remove Applications feature is a huge bonus to the Linux camp,
but there needs to be a lot more work in the area of hardware-detection and
2. The choice of more than one GUI can be confusing. KDE? Gnome? X-Windows?
What's your favorite? Choice is good in toothbrushes, but not so much when it
comes to OSes.
3. The GUI interfaces are different enough to cause confusion as to where
the start button toolbar is located, what browser to use and other basic operations.
Some Linux installations like Ubuntu have two bars -- one on the bottom and
one on the top. Imagine that clueless guy in Accounting who can barely drive
Microsoft Office trying to figure out which button to click in Ubuntu.
4. Most use IE -- not Mozilla or Opera -- because that's what came bundled
with Windows and that's what they know. The choice of Mozilla (Gnome) or
Konqueror (KDE) will be daunting.
5. OpenOffice 2.0 is quite functional, but most will be stymied when trying
to perform basic tasks. However, I do actually think the OpenOffice interface
is better than that of Office. It's more intuitive and in some cases easier
to use with more tools at your fingertips.
6. While Linux will talk to Windows networks and connect to Windows-based
printers (and vice-versa), it's no trivial task to get it to do so. Most
users have little or no clue about Windows networking. Throw in Samba and they'll
check out -- it's as simple as that.
I started out as an advocate of SuSE and KDE. Today, I'd tell you that if you're
going to convert your users, you should consider Ubuntu and Gnome. They're significantly
friendlier, in my estimation. You can also extend the life of those older PCs
by using Xubuntu. And the OpenOffice interface will definitely agree with 95
percent of average users.
That being the case, Linux is 50 to 60 percent of the way toward the realm
of large-scale adoption. LAMP -- which stands for Linux, Apache, MySQL and PHP
-- servers are already making headway into the datacenter, pushing out many
Windows Server installations (see "Linux
Gains Windows Muscle," July 2007).
If Linux developers can match the sophistication of Windows with seamless installation,
detecting nearly every hardware component and running numerous applications
side by side, that would make a tough run for Windows in the future.
Until you can somehow get across to the average user that you've got something
simpler, easier to use and more capable than what's currently in vogue -- plus
it's free -- you won't see large-scale adoption. For example, my wife, who's
the CIO for a large government department, says it doesn't matter if Linux rocks.
She's simply not going to go through the pain of reinstalling all the PCs and
training the users.
As Walter Cronkite used to say, that's the way it is. (And by the way, this
article was written entirely in OpenOffice Writer.)