In the past few months, the debate concerning which software
purchasing/licensing protocol is better, open source or proprietary, continues
to percolate in the TechRepublic discussion boards. On one side of the debate,
we have the camp that distrusts the motivations of the open source community
and dismisses the viability of an economic model that does not revolve around
the exchange of wealth. On the other side, we have the camp who distrusts the
multinational conglomerate software giants and their preoccupation with
earnings and control. Stuck in the middle are people like me, who have no
technological agenda to champion, but instead just want to get some work done
using the best software they can get their hands on.
While the debate will probably never completely die, perhaps
we can help turn the heat down enough to reach a slow simmer by analyzing the
current debate with an eye toward establishing common ground and a rational
Foot in the door
The one significant change in this ongoing debate has been
the increasing acceptance of open-source software as a viable alternative to
traditional proprietary applications by the decision makers who control the
enterprise purse strings. As I suggested in a recent discussion thread:
The question is no
longer whether open source is viable, but whether IT professionals will install
such applications in their respective organizations. This subtle change means
that open source software is now competing on the basis of effectiveness, cost,
reliability, security, and features rather than merely trying to get a foot in
the enterprise door.
However, when it comes to Linux, not everyone is convinced it
is ready for the end-user desktop. For example TomSal argues that:
Linux won’t be on the
desktops soon — maybe a mondo huge companies OR companies that the nature of
their industry means all their users are engineer types; but the average Joe
small business shop with users who need their helpdesk to figure out how to
change a default printer or change a font in Word, EXTREMELY unlikely that
Linux will find a home there anytime soon.
Linux needs to be
friendly to the “AOL” crowd first..then we as IT folks can bring
serious arguments to our CIO/CFO and CEO’s to why Linux is a good switch for
the corporate desktop.
I used to think along similar lines myself. I believed that
Linux required the use of long command line strings with weird syntaxes and
cryptic parameters. Point
& Click Linux, which was recently featured on TechRepublic, changed my
mind. The MEPIS Linux distribution version included on the CD accompanying the
book is about as simple to install as it gets. The “Linux is too difficult to
install” mantra just doesn’t hold water anymore. As apotheon observed:
I find that end users
are even less likely to be able to install a Windows OS properly without help
than certain very user-friendly Linux distributions. SuSE comes immediately to
mind. Fedora and Progeny Debian both use the Anaconda graphical installer,
which is more friendly and easier to use than the Windows installer. MEPIS is
so incredibly easy to install on the hard drive that it beggars the
Even Debian (my
personal favorite distro for a great many reasons), once used as the poster
child for how difficult Linux is to install, has had its installation process
revamped to the point where installing the base operating system consists of
little more than accepting default options.
Anyone trying to make
a case for sticking with Windows based on how easy the OS is to install has not
been paying attention.
Summing it all up
But even if you accept the idea that Linux and open source
software is ready for end-user desktops in many enterprises, you may still have
to overcome other barriers to more widespread acceptance. Security concerns
revolving around the implementation of software are a major concern for any
organization, especially if regulatory and legislative compliance is a part of
the equation. Enterprise decision makers need assurances that the software
deployed within the organization is secure. This is the concern expressed in a
TechRepublic discussion thread started by KaceyR:
In keeping up with the
Open Source software movement, I’ve come across a single, basic flaw.
The only way to ensure
that your executable is as it should be, is to perform a comprehensive review
the source code and to recompile it yourself.
I can, very easily,
set up a distribution web site that contains both the source code and compiled
executables, complete with my own hooks in the executables that will do
whatever I want them to. The typical user will download the executables, maybe
even the source, but will never perform a compile, and I certainly won’t have
my hooks in the source that they can review.
Without a complete
review of the source code and an independent compile yourself, you have
absolutely no assurance that the code you are running matches the source code
that it’s supposed to. Should that code damage or otherwise compromise your
system, what’s your recourse? Rebuild your system.
In addition, if you
have the time and intellect to review and completely understand the source
code, why are you wasting your time downloading someone else’s product when you
can make your own with the same level of effort?
By example, let’s say
you download a copy of Firefox, and it’s been tweaked with a hack that allows
an external user into your system. You’re browsing around the internet and
everything is great, then one you realize that you’ve lost all of your data.
During a post-mortem, you discover that Firefox was the culprit, so you go
after the developers at Mozilla. Oops! The signature of the executable doesn’t
match ANYTHING the original developers have ever released. They’re not
responsible. Time to rebuild your system.
Now let’s say that
you’re running proprietary software and the same thing happens. During the post
mortem you discover the culprit is the ABC product from XYZ company. The file
signatures are compared and, sure enough, they match. XYZ company is clearly
responsible, so they will be inclined to assist you in determing the exact
cause and fixing the problem, as well as you (possibly) having a legal recourse
against XYZ company.
This is both a level
of protection and a level of assurance that the program will perform as expected.
Companies today are
very paranoid (and rightly so) about system intruders and industrial espionage.
With this in mind, why would you turn to Open Source software?
In a nutshell, this argument is probably the main reason
more companies have not implemented open source software solutions en masse.
Whether it is deserved or not, the perception has been that open source is not
as robust as proprietary software—that, when it comes to support and
accountability, there is no single entity whose proverbial feet you can hold to
the fire. Backers of open source software have their work cut out for them when
it comes to overcoming this lingering stigma.
Perhaps the best answer to the “is open source software
secure” question posed by KaceyR was
expressed by bhoule:
Several years ago, a
long time commercial database product (Borland InterBase) was turned over to
the Open Source community. Only after its public release was it discovered that
an unpublished “back door” previously existed to bypass the security
of both the DB and the OS it ran on.
How can you be sure
that the same situation does not exist in other commercial products? Without
access to the code, how will you ever know? With open source, at least the
possibility exists that nothing similar will sneak by. But does this mean that
everyone should eschew commercial software? Certainly not.
Open Source is not a
cure-all, nor will commercial vendors ever drive the open code model away. The
decision to implement one vs the other is a function of economics and trust. Do
you trust the open source community more than you trust your vendor? Is the
long-term TCO of open source greater or lesser than that of commercial
programs? There is no right answer. Neither camp is 100% correct. It boils down
to an individual’s decision in specific situations.
But this concept of
individual choice and responsibility brings to mind another trap that KaceyR
falls into. The claim is that the only way to know if an open source product is
safe is to personally perform a code review. While speciously correct, the
argument neglects the power of community and the monitoring and reporting that
the collaborative open source environment provides. If living in a vacuum, open
source exploits — theoretical or otherwise — would have a field day. But the
reality is that with millions of eyes in the open source community, such an
exploit would not exist for long. One could be paranoid and demand a
self-performed code review on every single open source app, but then I would
expect such an individual to be testing their gasoline before filling their
car’s tank just to be sure it wasn’t ethanol or diesel!
Lastly, KaceyR’s legal
liability hypothetical was a bit of apples/oranges. A checksum-validated
commercial product was compared to a checksum-failed open source project. Let’s
turn that around and compare like items: a commercial vendor is no more likely
to support a checksum-failed product than an open source provider is. As for
the checksum-validated converse, while it may be technically true that legal
recourse would be possible in the commercial situation, reality dictates that
no-fault EULAs and the long legal process would render the point moot. Legal
indemnification of commercial software serves mainly to bolster the
Trust in whom you feel
appropriate. It seems clear from the comments that KaceyR will live happily in
the commercial world. Zealots, meanwhile, will run with Linux on their desk,
Opie on their PDA, and Freevo on their DVR. The rest of us will continue to
strike that balance of economics and trust.
The question of secure software is very compelling at first
blush. Fear, uncertainty, and doubt are difficult obstacles to overcome,
especially in organizations not culturally attuned to change. Tried and true
may cost more over time, but they are a known quantity, bringing with them the
comfort of predictability. But when you get to the core of the matter, what
makes software secure; how much assurance and trust can you have in anything
software related? Bugs, bad installs, and broken features are all just part of
the experience for any IT professional in an enterprise setting. Things happen.
In the final analysis, it comes down to service and
confidence. As an IT professional, when making any software purchasing
decision, you must consider the quality of the service agreement (if there is
one) and use it to assess your confidence in the overall viability of the
software in question. This applies whether the software is proprietary, open
source, or developed in-house.
As t_mehta put it
in a discussion post:
- Trusting any software is a question of
faith and testing–Can we trust any/ all proprietary software to be
absolutely perfect–nope, same is true for OSS front–no difference here.
- Will finding whose responsible for a coding
problem in OSS or in MS or Borland etc. solve the problem in my plate now
– nope, you’ve got to wait for someone to fix it, who pays that guy is
irrelevant – no difference here.
- If you are a corporate with service
support–commercial OSS support and proprietary support is equally
good–no difference here.
- If you chose to download & install
software with no support identified in advance–e.g. freeware on MS or
Linux–your support options are the same–NIL, if things go wrong its your
funeral–no difference here.
- Legal recourses are the same if you
have the same support agreements in place–no difference based on
platforms (OSS/ Proprietary).
Overall, the argument
put forth is axiomatic since in simple words it says proper service agreements
are useful to customers irrespective of the source/ ownership of the software
installed – that’s like saying 2 + 2 = 4, we all know it, so what’s the big
deal–I don’t get KaceyR’s
issue–lets compare unsupported Proprietary software with supported OSS
software–give me the supported software any day, who cares about its origins.
I say, based on the
lifespan you are looking for use whatever is functionally appropriate and
cost-effective (over the lifecycle) IRRESPECTIVE of it belonging to the OSS
world or proprietary–period.
I think this is really what it comes down to: The amount of
code and sophistication of features incorporated into application software
means that no program can ever be truly perfect. Due diligence will be required
no matter how the software is acquired. The real debate between proprietary
software and open source software lies not with the often politicized method of
delivery, but with the features offered by the software itself. The really
fruitful debate should be focused on the merits of the software, not what was
paid for it and to whom.