Hi Logan,

First, thanks for addressing the list.  I think it takes integrity and guts
to respond rather than just run away.


-----Original Message-----
From: Logan Harbaugh [mailto:[EMAIL PROTECTED] 
Sent: Monday, November 24, 2003 8:05 PM
To: [EMAIL PROTECTED]
Subject: [SAtalk] An Open Letter to the SA-talk forum


To all concerned, I apologize for the apparent maligning of SpamAssassin in
my recent article in InfoWorld. In my original article, I stated that I used
the 2.44 release of SpamAssassin for two reasons - because it was the
version shipping with the latest release of Red Hat 9 and because it would
illustrate how much the state of the art has changed in the last year or
two. This explanation was condensed in the finished article by copy editors,
which is beyond my control. This will be covered in the letters to the
editor section of InfoWorld so the rest of the world will know that I did
not deliberately use an old version of SA to show it in a bad light against
commercial products. I plan to review the current version in an upcoming
article, and I am sure that it will perform better. 


LG >>
Unfortunately for your readers, illustrating how much the state of the art
has change imparts the impression that SA is static.  You need to better
express that state of the art for SA has changed too.  The manner in which
you chose to include SpamAssassin (SA) was wrong regardless of your
explanation.

But this strays from the real point.  You know where the web site is and you
should have obtained the latest version and you should have tested it againt
the problem that it really solves.  It is not enough to just compare all
these products as they solve different problems.  Your article does not
properly address this.  For example, I own stock in Postini and I love their
service.  I have a customer that subscribes to the Postini service and I
would not even try to talk them out of it for SA or any other solution.
Postini will not only protect against spam and viruses but it will also
place the infrastructure costs on Postini rather than being incurred by the
subscribing customer.  As bursts like Klez and Swen come along, machines
with greater capacity need to be able to handle the bursts of traffic.  So
Postini customers need not over design their infrastructure.  Additionally,
Postini customers do not need Email MTA/SMTP experts.  They can continue to
utilize groupware specialists.  Now I can go on with other factors but I
really don't need to to make my point.

Now on the other hand, companies that utilize SA satiate a much different
problem set.  SA is not a product nor is it an anti-spam gateway product.
SA is a tool that can be utilized within custom solutions.  SA can be used
on the perimiter, it can be utilized in a service, it can be used on
mailservers that contain the mailstore, etc.  I also have customers that
will not use Postini because the Postini model does not address their needs.
Postini can not be customized in any fashion like a solution that utilizes
SA.  And while we who use SA know that it's tagging accuracy can be
configured to compete with any product on the commercial market, you failed
to convey the reasons.  Out of the box, an *up-to-date* version of SA is
very effective.  However, I think everyone does agree with you that it is
not a hands-off solution . . . at least until it is configured properly
within the target solution.

Although you allude to an understanding that the solutions address different
problems or criteria that addresses a problem set, you really portrayed the
products as if they are equivalent.  Each product, however, is really a
unique solution that addresses a spam problem.  Each solution is optimal for
different environments.  So the real question is not how easy to
administrate, or how effective is it, or what has the best support, or etc.
While the aforementioned points are important and necessary, the question
that best needs answereing is what environments can best be served by the
specific products mentioned in your article.
<<LG


Regarding some of the other comments that have been made, some of you have
said that SA is not hard to install, taking no more than an hour or two to
download, install, configure and begin using. That is consistent with the 10
times longer number I used, because the other installation and configuration
times were all around 5-10 minutes. You have said that an experienced Linux
administrator doesn't find SA difficult to install or configure, and that
additional functionality such as user-accessible white lists can be added,
either through additional open source software or by writing scripts or
programming to extend the functionality of SA. That's true, but not really
relevant, unless there is a distribution that contains all of those
features.

LG>>
Of course it is relevant!  This is exactly the point of SA.  SA is a tool.
Even in the most basic configuration (which is dependent on the MTA chosen),
SA will take on different faces.  SA is not the final product but a tool
chosen for use within the final product.  Comparing it to other solutions
needs to address this.
<<LG


You have also said that I should have taken into account the fact that it
doesn't cost anything before making statements about it being harder to
install, configure and manage than the commercial products. SA does cost -
but in an administrator's time rather than money, which I did say in the
article.

LG>>
But the point is not really clear.  You are comparing SA out of an old box
to current versions of other products.  The comparison by nature is not
correct without mentioning that SA can be used in other products like
SpamKiller by McAfee.  SpamKiller is a descendent of Deersoft SpamAssassin
Pro.  To make things equal, you would need to compare SpamKiller to
BrightMail.  SA is just the tool utilized within other products.
<<LG


The same is true of support - while you may get faster or better support
through this group than you get with commercial software, there's no
guarantee that you'll get any support at all - and most organizations will
find that hard to live with.

LG>>
I would certainly agree that some organizations would feel uncomfortable
with no non-commercial support.  I don't know why though as support from
some large companies I work with is a coin toss as to whether I get the
problem solved or not.  Others have gone through cycles of good/poor
support.  I don't really know that "most" is necessarily true.  Might be a
good article. ;)
<<LG


So, when I review the latest version of SA, you can expect performance to be
better, but I will still look closely at installation, administration,
updates, maintenance, reporting, granularity of management, and end-user
features for SA, just as I will for any other anti-spam packages I review. 

>>LG
And so you should.  However, please also mention the nature of SA and where
it may fit.  It is really important that your readership understand what SA
is and what it is not relative to commercial products.  This list would
obviously not exist if SA did not have it's place in the world.
<<LG


Again, my apologies for creating a story that distressed so many of you. I
do try to create balanced reviews that reflect the pros and cons of all the
products reviewed. 

>>LG
Thanks again for responding Logan!
<<LG


--Larry



-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?  SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to