Roy Schestowitz wrote:
> Free software may kill some software firms. So what?
> ,----[ Quote ]
> | Conclusion
> | It is obvious to anyone that free software is gaining ground
> | everywhere, and that this is an evolution which will only
> | continue in the future.
This is actually the FOURTH paradigm shift in how information
technology was marketed.
In the first generation systems,the computers were huge, expensive, and
required constant support and maintenance from the provider. The
computers, software, and hardwer were LEASED and the support was part
of the huge lease contract.
In the second generation, customers wanted to own their computers.
Vendors charged huge price tages, because they knew that the price
would include the support. You might pay $10 million for an IBM
Mainframe, but you had at least one full time person sitting in an
office in your building, who could bring in anyone needed, any hour of
the night or day, who could access the system remotely, or could fly or
drive to the customer site in a matter of hours.
In the 3rd generation, hardware was relatively cheap, and software was
expensive. The softare price included some support for installation
and configuration, and then you would pay additional funds for
customization and production environment support.
In the 4th generation, the hardware is very cheap relative to the cost
of the whole project, the software is free or very cheap, and the
consulting and support are purchased independently. This allows the
customer to match/balance his budget and needs for the optimal blend of
functionality, customization, and support. Keep in mind that a major
project might cost $1 million in hardware/software, $5 million in
consulting and customization, and $50 million in production environment
support. In 2005, Linux generated nearly $16 billion not including
production environment support.
Linux, OSS, and industry standard software based on publicly documented
standards, help to reduce the actual costs on these customization and
support efforts. Even where proprietary software such as WebSphere, or
Rational, is used, a key function is to reduce the time and effort
required to generate industry standard solutions.
Microsoft is gradually making the transition from 3rd generation
royalty based revenue models to 4th generation service based revenue
models, but customers have not been terribly impressed with Microsoft's
"service". Many do like the MSDN package for developers, but the
equivalent package for production or GA products has left much to be
desired. Both programs can cost as much as $100/user/month or up to
$1500/year, possibly even more, but the support hasn't been all that
great, compared to the innovation that seemsto be just "stock in trade"
for Linux/OSS.
> | It is just as obvious that this
> | evolution will be a problem to people who have high stakes
> | ^^^^^^^^^^^
> | in proprietary software.
> `----
> http://www.freesoftwaremagazine.com/blogs/free_software_may_kill_some_software_firms_so_what
> http://tinyurl.com/yhxwfj
>
> This explains the presence of Microsoft shills in this forum. It's a matter
> of self defence and interest, not logic and honest. IOW, all the criticism
> of Linux is invalid; it's about one's survival.
Absolutely! Microsoft is severely threatened by this whole program,
because it's such a new paradigm. Its' as foreign a concept as the
Quartz watch was to Swiss watch makers, as customized cars were to
Henry Ford in the 1920s, or as TCP/IP was to Novell.
In this group we have a number of professionals who have based their
entire careers on Microsoft's most proprietary technology, and have
never bothered to take the time to learn the published standards and
OSS technology and applications. Some are "one trick pony" types who
have spent years doing Visual Basic interfaces to SQL Server databases.
Mind you, there is nothing wrong with this. There are COBOL
programmers who spent the latter portion of their careers writing CICS
programs for DB2 databases. Some of them retired shortly after Y2K
(some actually came out of retirement for Y2K), others learned new
technologies including VB and Java.
There are people who have spend 7 years doing nothing but
MQI/MQSI/WMQI/WBI-MB, and their skills are very much in demand, but
they have had to maintain and upgrade skills almost every year to keep
up with the software that more effectively meets more of the needs of
our customers.
Many of these people who have built entire careers out of programming a
particular application component such as VB for DB2, have honed their
skills and abilities to the point where they can mouse around faster
than they can type, and can reuse components they have written in VB or
C# and quickly integrate them into new projects. This is good.
When Y2K ended, around March of 2000, there were a lot of people who
were experts in Windows and COBOL programing who suddenly found
themselves going to interviews for jobs as web programmers. Many had
read the Microsoft propaganda and tried to claim that they could do
anything that UNIX/Java programmers could do using Windows/VB. Some
were so good that they really could whip out a nice pretty user
interface in a day or two using Microsoft's tools. The irony is that
they were often asked to generate an XML request which would then be
sent to a message queue (often MQSeries) where a UNIX/Java system would
do the integration to the corporate servers, suppliers, customers,
bankers, accounting, inventory, and shipping.
In other cases, Microsoft often provided a nice user-friendly front end
for such functions as stock quotes. They could query transactions sent
to the actual transaction processing syste and collect the statistics
into local SQL Server tables using ODBC. The cached tables could be
searched without putting additional load on the main database. Since
you only needed a fraction of the information, and you could re-read it
as often as needed, it wasn't a big deal if the server crashed and had
to be auto-rebooted. If you had enough of these boxes running in
parallel, and used a UNIX based load balancer, you could get pretty
good perforance a a lower cost than trying to read directly from the
production trading system. Microsoft is pointing to a similar solution
that has been deployed in the London exchange.
Of course, these solutions weren't cheap. Windows 2003 licenses are
normally very expensive, often costing as much as $20,000 per processor
(negotiable of course), and SQL Server licenses have similer site
license fees (CALs just aren't practical).
Linux would normally use MySQL or PostgreSQL to do similar tasks, but
my guess is that the London exchange priced their Linux solution using
a premium database such as DB2 or Oracle. I suppose if someone really
dug, they could confirm or disprove this.
The problem is that people who do pretty much the same type of
"assembly line" task year after year are often threatened by change.
If a Windows/VP/SQL Server user suddenly had to learn Linux/Unix, Java,
and industry standard SQL and modelling tools, it might mean losing the
reputation they have worked so hard to establish.
If I wanted to get a really quick and pretty interface using a
lightweight Windows based solution, I'd probably want someone like DFS
to do it. I might even have him generate XML instead of SQL (something
that is practically automatic with Visual Studio), because I know that
he could probably do it in less than a week, and even though his rate
is high, if I have a good specification (which probably took 2 months
to define and confirm), his "quick and easy" approach would be a really
great way to get a good user interface quickly.
Ironically, I really do hope that people like DFS do join the
Linux/Java/AJAX club soon, because that same skill that has made them
so good at doing Windows GUI interfaces would be even MORE in demand
for the Linux/Java/AJAX solutions.
Unfortunately the very thing that makes them enjoy doing the same core
tasks over and over again are what make it so difficult to learn new
technologies quickly and efficiently.
Linux/Unix people tend to be "bleeding edge" people who are pushing the
envelope of performance, stability, security, and supportability, but
once the puppy comes together, they turn the work over to support
organizations. The support organizations often don't even uses the GUI
interface, they can diagnose problems in a *nix system using the
command line interface, and from that point of view often have to ask
which kind of system they are using because the UNIX and Linux servers
look almost identical. They usually remember when they try to run an
AIX command on Linux, or an AT&T SysV command on Linux, or try to run a
Linux command on AIX or Solaris.
|
|
