Why SATV doesn’t use Linux

I’ve been in IT for nearly 30 years, and that has made me jaded towards many of the tropes that one finds in the IT press, such as "Mac Vs. PC", "Should Vista Be Abandoned", and the perennial, "Is <this year/next year/year after> the Year of the Linux Desktop".   I ignore all these memes, almost always perpetuated by lazy IT journalists and editors who know that the ad money rolls in when they get the monkeys to fling turds at each other.

But Alexander Wolfe wrote an interesting article on Linux, 7 Reasons Why Linux Won’t Succeed On The Desktop, and I want to address its points.  Outside of Silicon Valley, there are very few people on the office, or factory floor, with any real opinions on Linux.

Here are a few of Wolfe’s arguments with my observations,  It’s a long essay, so I’m posting this in two parts:

1. Prohibitive application porting costs.

At SATV we have many different specialized applications, what are commonly called "Line of Business" (LOB) applications.  We have a server that schedules our programming.  It runs on IIS and .NET on a Windows 2000 machine.   It will always run on IIS and .NET.  Another companion to that server is used to put "bulletin-board announcements", the graphics and text screens that indicate upcoming programming or the next meeting of the City Council.  That relies on DirectX.

There is no way that either program could be ported to Linux without a lot of pain, if it is even possible at all.  Linux has nothing like DirectX;  until very, very recently (Beryl), there hasn’t been any native 3D support in Linux at all.

This isn’t only limited to software.  Many specialized vendors sell computers integrated with their hardware for specific tasks.  One such machine we have is an Inscriber.  This does broadcast graphics for our studio.  This particular box runs Windows XP.  It has some specialized video and audio hardware in it.   As with the app mentioned earlier, it is highly dependent on the Windows display architecture.  It produces marvelous TV graphics.

It also costs $20,000.

That box is going to have Windows XP for the rest of its life until it dies or we sell it.  Harris, its vendor, will not support it to be configured otherwise and its software is likewise going to run under Windows for the rest of its life.  As it was, I had to play "mother, may I?" with Harris so that I could update it to Windows XP SP2.  (There is no case I can make for Vista on it and no need to.)

2.  The Fanboy alienation factor, or how Linux’s biggest supporters drive away potential new users.

One of our biggest factors in making technology decisions is community.  Salem Access Television is a resource for our community.  More than that, we often decide to purchase things that have a community behind them. 

For example, Apple products are widely used for video production in public-access TV facilities like ours.  There is a community.  There are peers we can go to for advice, both to get it, and to give it.  We have Macs with Apple Studio and iLife for video editing and I’ve spent plenty of time myself with Final Cut Pro, Motion and DVD Studio.

I’m no fan of Steve Jobs, but I would not have it any other way.  Apple works for us.

Similarly,  when I first took on IT duties here in 1998, I had to migrate from an unreliable, failing server from an incompetent consultant.  The old server was running Small Business Server 4.0, and I had briefly considered something Linux-based.

It turned out there wasn’t anything Linux-based that would replace the functionality that SBS offered, but even more than that, there is an active community of people that sell and support SBS.  A friendly community with people like The SBS Diva

I recommended SBS 2000 for our network server, and that went very well.  It was the start of a great relationship that continued with SBS 2003 and will probably continue with SBS 2008.

That’s a contrast with my one exposure to the Linux community.  Around 1998 or so, I was using The GIMP.  The Windows version was new and very unstable so I setup a dual-boot between Windows and SuSE, where I ran the Linux version of the GIMP.

I wanted to get my elderly Mustek scanner working under Linux.  Like most cheap cheap scanners, this used a nasty ISA-card SCSI controller based off the NCR53C400 controller.  I wanted to use SANE, the then-new scanner backend, but needed to compile the driver module and load it on boot. 

I was already well familiar with compiling kernels and updating modules and such so that went smoothly.  But I could not get the NCR module to load or start no matter what did.  It gave an obscure error message in /var/log/system, something about a resource already in use.  Granted, this was in the days of ISA buses, IRQs and IO addresses, "plug and play", er, "plug and curse", but I had gone over my hardware completely.  To use Linux in those days, you had to!

This went on for some time and I made the mistake of venting about it on a newsgroup.  I was very curtly told to "just keep working at it ’til you get it right!"

I decided to look at the source code;  (I’m a CS grad, I can!)  The driver’s author had used a default configuration to be used if there were no parameters specified.  (By the way, there were no parameters documented!)  Unfortunately, the defaults included the particular IRQ/IO port configuration for his particular system!

Two things were wrong there:  Firstly, source code is not formal documentation.  The original developer was at fault for not including documentation (which would have been at most a few pages of parameters) and the kernel maintainers at the time were at fault for not making documentation a requirement when checking in.  Secondly, some parameters–like IRQ and IO addresses–have no good defaults and the driver needed to report that in the log and unload.  There are too many times in Linux where you don’t know what a module needs for options and you have to load it and look in the messages to see if it told you anything.

But that’s probably trivial by now;  I decided to go without my scanner and the driver is not part of the kernel anymore.  But I can never forgive the Linux community both for being rude and tolerating, even encouraging rudeness.  (I saw the guy on the newsgroup later on asking for help for his problem!  Did I mention I hate hypocrites, too?!)

This just has to be a drain on Novell and Redhat who must mediate between these folks and their paying customers.  Their customer service cannot treat their support callers the way Linux advocates treat everyone else outside of them.  I would pay for tickets to see a SuSE tech, perhaps even their lead architect or CTO, say on Slashdot or a newsgroup, "You’re being a jerk.  Stop it.  NOW."

More ranting in part 2.

Disclaimer:  Salem Access Television has no institutional opinion on Linux, and neither does its staff, board or membership. 

2 Comments on “Why SATV doesn’t use Linux”

  1. Rick says:

    "I would pay for tickets to see a SuSE tech, perhaps even their lead architect or CTO, say on Slashdot or a newsgroup, "You\’re being a jerk.  Stop it.  NOW." "
    Oh, that would be so awesome.  I\’d buy two tickets.

  2. Philip says:

    Your post is timely.
    For us, all it takes is a test account on our Small Business Server setup to demonstrate the benifits, and Linux becomes a figment of their imagination.
    And, there is no way to be really charitable when it comes to my experience with Linux fanboys. So, the blog post speaks for itself:
    Trackback: http://blog.mpecsinc.ca/2007/10/unbundling-windows-and-so-called.html

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s