[Xitami] Status of Xitami ??

[Thomas J. Hruska]

At 10:31 AM 10/6/2004 +0200, Pieter Hintjens writeth:
>Hi,
>
>> I'd like that it is untrue, the only who can deny this are at Imatix... and
>> they seems to be around for a long time here... I'm very displeased of this
>> because i really like Xitami. I really appreciate if anyone at Imatix just
>> say "hello, we are here"....
>
>Hello, we are here.  Busy, but here.
>
>I'm sorry we don't spend more time reading and replying to the Xitami list.  
>It's one of many things we have to spend time on, and inevitably gets less 
>attention that it deserves.
>
>Xitami is, obviously, a project that is fairly static.  We make patches now 
>and then, especially when there are security issues.  The server solves a 
>problem for many people but we cannot compete with the huge resources put 
>behind products like Apache, and we would be foolish to try.
      ^^^
      Buggy.

You missed an important word.  Apache 2 is still buggy under Windows.
Better, but still buggy.  And convoluted.  The five configuration files are
evil convolutions of reality and very user-unfriendly.


>Allow me to explain briefly where we are and how we got here...
>
>When we started Xitami in 1996 the goal was not to develop a web server so 
>much as to demonstrate a way of building servers.  That's why it was open 
>source from day 1.  Unfortunately (for our spare time and personal lives),
we 
>were dragged kicking and screaming into developing our prototype into a real 
>working server.  I can speak for most of the Xitami team when I say that it 
>has always been a real pleasure to see our work used by so many enthusiastic 
>people, but that the flip side has been a support burden that rapidly became 
>impossible.
>
>In 2000-01 or so, we were forced to concentrate on earning money: this was 
>banal and much less satisfying than making Xitami better but the few
licenses 
>we sold for Xitami/Pro did not start to cover even our network costs, let 
>alone work on the product.
>
>I've been asked by members of our team to "officially declare Xitami dead" 
>several times.  I've resisted.  We use the software ourselves.  We use it on 
>websites and in products like smt@ (www.sms-at.com), which is one of the
ways 
>we make a living.
>
>But... Xitami as a living product... it's clearly something who's time has 
>come and gone.  The need for a small, fast, portable web server was
painfully 
>obvious in 1997-2000, and we are proud to have achieved what we did.  But 
>today it's a niche product.

Okay, before you kill off the product, I believe I have finally figured out
why Xitami and ProtoNova are only generating low-volume sales.  Fairly
obvious in 20/20 hindsight.

First off, when business owners want to build a network with Windows, they
typically purchase a NT Server-based solution from Microsoft because
Microsoft tells them they need to and they don't realize they can run a
server on client PCs (e.g. WinXP Pro).  As part of that purchase decision,
they implicitly choose IIS.  The only reason they don't know about other
web servers is because no one is marketing server alternatives to them.
Microsoft doesn't have to advertise IIS, they only have to advertise the OS.

A consumer, on the other hand, will want a cheap or free server.  The basic
Xitami server provides that and ProtoNova provides the cheaper and more
secure route for SSL.  Consumers may typically go for Apache if they hear
it holds 50% of the market share, but will likely look around for
solutions.  Consumers are a terrible market expectation as they are really
uninformed when it comes to software and usually end up loyal to a
not-so-well-written product X.  This group typically relies on friends to
make recommendations for software - also known as the grapevine.

Business owners wanting to use Linux will automatically choose Apache
regardless of the numerous options available.  Linux Apache users are
honestly the most stubborn people in existence who won't change over to
anything else.  You can effectively forget ever getting a sale from them.

Finally, you have tinkerers - people who like to play with things to try
them out.  There are very few of these people in existence and they tend to
hunt down the best products available and constantly weigh their options
with other factors (e.g. budget and time constraints).  I am most likely a
tinkerer as I've used Apache, IIS, Xitami, ProtoNova, Sambar, and 10 other
web servers with my personal preference being ProtoNova.

Now of those four general groups of people, which are most likely to buy
Xitami, ProtoNova, or one of the host of other servers out there?



If you answered the tinkering group, you are correct and have sound
business sense.  Targetting a group of tinkerers is a great way to get
locked in a market niche and lose a business.  Not one of those other
groups will buy any web server let alone switch without a sound business
reason...and that means addressing the bottom line.  You have to give the
people who make purchase decisions a really good reason to buy any given
product.  If they already have a web server and it _seems_ to be serving
their needs (regardless if it is causing the business millions every year),
they won't consider bringing anything different into the mix.  So, the
three largest areas where a web server could potentially be used are out
(with the occassional exception).  That's 80+ percent of the market share
right there.

This explains why Xitami and ProtoNova are not the best success stories in
the world.  Now, before anyone kills off either one, I have one last bit to
say.  There is a growing undercurrent of users of IIS and/or Apache under
Windows that are getting fed up.  Not quite to the point of hunting down
alternatives, but dangerously close that Microsoft is worried.  Of the
servers I present such users, people are most intrigued by ProtoNova but
are ultimately turned off because of _migration issues_.  That is, they
_want_ to move away from IIS to ProtoNova, but _can't_ because of the time
it would take to switch.  Remember, time equals money.  This means that
ProtoNova is dangerously close to swiping a large chunk of the market share
to be had and giving Microsoft a run for its money.  From a technical
standpoint, all that has to be addressed is IIS to ProtoNova migration to
make ProtoNova succeed.  There are other issues as well, but that is the
biggest technical hurdle.  Xitami needs a ton of work before the same thing
could happen.

Why am I sharing this information with direct competitors?  Because iMatix
has effectively stated that they really do not want to support Xitami any
more.  Therefore, the ProtoNova development team would love to take this
obvious burden away from iMatix.  ProtoNova was _designed_ for Xitami
users, meaning migration to ProtoNova from Xitami is simple and should only
take an hour or two for really complex setups.  There are, of course, some
differences, but out-of-the-box, ProtoNova functions in an almost identical
fashion to Xitami except that CGIs execute much faster under ProtoNova and
you never have to restart the server to make configuration changes (e.g.
dynamically bring Virtual Hosts and servers online/offline).

Xitami is 3 years behind ProtoNova in development and 4 years since its
last true update (emergency 1-5 line security fixes don't count).  The
logical decision, if you are concerned about Xitami, is to just move to
ProtoNova.  You are guaranteed the following or your money back:  If you
use ProtoNova and ever have issues with the server itself, you can be
doggone sure they _will_ get fixed - ProtoNova is a well-supported product.
 Additionally, if you need a feature added, ask and you shall receive.

For those who want a brief overview of ProtoNova.  ProtoNova has LRWP (1.0
and 2.0), Java Servlet, ColdFusion, PHP, Perl, Nuclear Vision, and custom
CGI support along with a complete interface for making your own custom
add-on modules that is much more secure and versatile than LRWP.  It uses
an ODBC-compliant database for logging almost everything.  It is massively
scalable and load-balances everything from memory use to Remote Module
management.  (Remote Modules are how CGIs, LRWP, etc. work).  It also has
features such as never having to restart the server (or sending the
HUP-equivalent signal) and automatic DoS detection and prevention
mechanisms.  It only runs on Windows at this point (there is a Linux
version, but it is not ready for production), but the goal of the server is
to be capable of carrying any load on any PC while being smart about
today's Internet security issues.  In relation to security, Apache will
return a 403 Forbidden error to the user if they try to access an illegal
page - ProtoNova returns a 404 Not Found error to the user.  Neither are
desireable, but which one will most likely keep a hacker out of restricted
areas?


Thomas J. Hruska
shinelight@shininglightpro.com

Shining Light Productions
Home of the Nuclear Vision scripting language and ProtoNova web server.
http://www.slproweb.com/