I attended an executive briefing yesterday that was very interesting. Dr.
Eugene H. Spafford, a professor at Purdue University who sits on the
President's Information Technology Advisory Council, had many interesting
and valid points, one the biggest ones being his "observation" on the
default installation of unneeded services (Windows being the main culprit
mentioned). These services are a big security risk, especially for those
that are what we know as "newbies." These types of end users have no idea
what these services are and most likely will never have a use for them. 

One of his other "observations" was from a developers stand point. He
pointed out that unfortunately most systems are designed with the virtuoso
in mind (this includes operating systems, applications, even hardware
interfaces such as web based firewall interfaces). Most end users are at a
low level when it comes to computing. This means that developers should
always be thinking of end users as low level, not as virtuosos. He also
pointed out how most developers think about security after release, not
during development. He stated that security should not be viewed as patch
management, but should be taken care of from the very beginning stages of
coding.

He made some other points about being needlessly bound to legacy systems,
user's lack of understanding statistics (look at tobacco use, airline travel
vs. driving, etc), and gave an example of using the wrong requirements to
make software decisions: The US Navy completed a study of MS products and
found them to be insecure. A year later they announced they would be using
Windows 2000 on one of their ships to the lower cost, not security. It makes
you wonder. Lowering the total cost of ownership for a product, device, or
application is not always the most important thing, especially if one of
these things has a security problem that can lead to costly damages. Also,
releasing a product before it is secure simply because you wish to meet a
deadline or make profit is a bad practice (once again, Windows). (I should
mention that Dr. Spafford has a working relationship with Microsoft and
visits them quarterly, yet he still points out their flaws)

Every one of the speakers on hand agreed that the best practice is to
disable all services on install and only allow those that will be used. I am
not sure if this is the answer or not. I personally like the convenience of
having some things ready to go, but I know what to look for; others may not.
I think ultimately the Mepis team must do what is best for Mepis. It is
impossible to please everyone, but it is possible for the team to create the
best distribution that they can create, which includes creating a happy
medium between security and ease of use, includes timely updates when
needed, and includes good user support. I think that Mepis as a whole has
met these requirements.

Here is a link to Dr. Spafford's site:

http://www.cs.purdue.edu/faculty/spaf.html

Joe


_______________________________________________
Dev-mepis mailing list
[email protected]
http://mepis.org/mailman/listinfo/dev-mepis_mepis.org

Reply via email to