London Stock Exchange to abandon failed Windows platform

do you remember this article:

While eating lunch, I was flipping through eTimes, and found this in the insider front cover.

http://www.sandelman.ca/mcr/humhour/microsoft-gets-london-stock-exchange1.jpg

Blogger

This piece, like any written by Nichols, is pure propaganda and a joke in terms of jouralism. He's as much a journalist as I am the Pope.
Like RMS, he'll get extremely fat and fade into oblivion with the rest of the relgious zealots who have no place anywhere near anything but perhaps a very pitiful comic book.

New LSE management have done the right thing

Steven, a big THANK YOU for your fine work !

The new LSE management took a wise decision by abandoning the proprietary OS. Building such complex systems on proprietary software, that do not follow Open Standards (TM), good quality and vendor-neutral ones, is simply insane.

I wish to see more Free Software solutions implemented in public administrations, including at DESKTOP level. Unfortunately, it seems Microsoft is pushed back, day by day, from Western IT infrastructure, so it relies more and more on emerging markets in Eastern Europe - counting on the public level of IT ignorance and officials corruption here ! I wish to see the day when small stock exchanges, like the Bucharest Stock Exchange, follows the LSE's model !

And YES, NYSE is running on Red Hat Enterprise Linux - please don't spread lies about that:

http://customers.redhat.com/2008/05/12/nyse/

Exemption

This is a very exempt journalist, I googled the following words "Linux failure Windows" and the result was the following article:
http://blogs.computerworld.com/how_vistas_total_failure_hurt_linux

With the following phrase:

"So what can Linux do? Well, for one thing, we need to get the word out that desktop Linux is available and every bit as good, when it's not better, than Windows."

The "we" is very careless for a true journalist which is giving only the facts...

This is only a commercial war between Linux, Unix, Windows, Java, C#, .NET, Apache, IIS, etc. It's only for sell...

Believe it, the guys that screwed up the project would screw it up in Linux too or whichever platform they would do the work.

Wrong technology-mix choices

I will not comment here about how good I am, how many servers I've seen, what I've developed lately or anything similar (even if I have some 60 completed projects on SQL+SSAS+SSIS+ASP.NET+IIS, both on 2K and 2K5 versions, relying on Windows 2003). I will state only two issues that were wrongly chosen in this project :
a) SQL 2000 -> even if it may be pretty good for its age (~10 y), it is not built for such stuff ( .NET is a technology that has nothing to do with SQL 2000. It is from different era. Even if you can get decent communication between the two, it's not logical to combine them. SQL 2K was designed to work with VB 6 and stuff like that. So SQL 2005 was the only logical choice. I see some say it works shitty. I will contradict them. Framework.NET works great with SQL 2005. And this SQL is way better than its predecessor.
Of course, everybody can say Linux+Apache+PHP+MySQL or anything else in there (Oracle,DB2,Java) just for being against MS (a.k.a. M$). But I say projects should be done considering only their requirements, ignoring (and sometimes) despite any marketing-based choices.

Generalists vs. Specialists

I think the basic issue is having a general IT integrator develop very technically demanding software (HA + Low Latency + High Load.) This should have been contracted out to a specialty team with a proven track record making very similar software.

There are companies that specialize in this. I worked at one several years ago that developed Windows-based low latency transaction routing software. But, it was not .NET, and it relied heavily on proprietary C libraries that had been developed and stabilized in financial transaction processing for over a decade.

I speculate that for this project someone at the overall project level looked at build or buy, and concluded they could build an acceptable solution inhouse for much less money than outsourcing it. Another painful lesson on the real value of a proven track record!

It is too simplistic to blame this outage entirely on OS platform selection, though it probably resonates well with people's black and white views of OS environments.