Thursday, May 31, 2007

Czech Air Flight 0104

I am glad I did not take this flight on my way to the US last week.

Wednesday, May 30, 2007

European success at last (.fm)

Look here if you want to see a successful European startup...

Saturday, May 26, 2007

Lovely stuff

No more Skoda jokes please...

Thursday, May 24, 2007

Three reasons why HP should join NetBeans

1) Competition is good and with the upcoming 6.0 release it is clear that NetBeans is a viable competitor to Eclipse (maybe the only real long term competitor). Support from HP would give NetBeans project more credibility and this open source software would become less dependent on Sun Microsystems.

2) HP Software is focused on the lifecycle approach to automation of IT processes (a.k.a. BTO) and a Java IDE would make the application and service development and testing stages of the lifecycle more developer friendly. NetBeans extensions for quality, SOA, diagnostics and other parts of HP software would fit nicely into this picture.

3) It would make me happy to see the NetBeans and Systinet teams to work together. And I understand that this is not a real reason but it would feel really good...

Monday, May 21, 2007

Who is the Powersoft of the SOA era?

Dennis Byron asks "Who is the Powersoft of the SOA era?" here. Many companies claim that their tool is the PowerBuilder for SOA (BEA WebLogic Workshop, Above All Studio and others). And VCs would love to find and fund the next Powersoft. This means that Powersoft remains an iconic company 13 years after it disappeared. And since I worked with and for Powersoft since 1992 I feel I should write here my version of how Powersoft Corporation became the leader of client/server tools market:

Right time, right place

There were three megatrends happening at the same time in the early nineties:

• PC with MS Windows 3.1,
• Local Area Networks,
• SQL Databases

and Powersoft happened to be right in the middle of these three trends. Several other companies were building c/s development tools at the same time but only Powersoft got it right. It may sound strange from today’s perspective but while the competitive products supported MS Windows as only one of the client operating systems PowerBuilder supported Windows natively. And as the result it was visibly more user friendly.


The key component of PowerBuilder was an object called DataWindow. DataWindow allowed developers to generate screens and reports quickly without deep understanding of SQL. But the fact that the finished application had to be deployed on every single client machine quickly became a management nightmare - so called "DLL Hell". Powersoft tried to solve some of the problems with the server-side version of PowerBuilder but it was a proprietary solution and the EJB standard did not appear until 1997.

DataWindow was also able to search systables of SQL databases for simple metadata. It was fairly trivial but it worked really well in demo applications. The demo presenter was able to connect to the target database and to build a simple form in the matter of minutes. To the audience that was used to COBOL programming it was “indistinguishable from magic”.

Business model

The biggest weakness of Powersoft was its business model. The company was selling development licenses of PowerBuilder and there were not enough developers in the world to sustain the initial sales growth. And PowerBuilder based application did not generate any runtime revenues at all. Database vendors have much sounder business model - Oracle and Sybase can give away development tools as a loss leader and charge for the deployment of their database. And so it was only a matter of time before one of these companies would acquire Powersoft:

In a surprise announcement officially released at 7:30 a.m., EST, on November 14, 1994, Sybase revealed its plans to acquire Powersoft in an exchange-of-stock deal. With a cash equivalent valuation of approximately $945 million.


The initial focus on MS Windows that helped the company so much initially became a liability by 1995 when Netscape went public and application developers started to move to the web. There was an attempt to build a Netscape DataWindow plug-in but it was obvious that Powersoft days are over. Only 10 months after Sybase spent over $900M! Some of the people behind PowerBuilder tried to repeat the success at SilverStream. I happened to be one of the early investors in SilverStream but I did not like what I saw and I decided to build NetBeans instead.


I don’t believe there will be a Powersoft of SOA era - company worth a billion dollar based on $3,000 development tool. And the main reason may be the need for standardization that makes the differentiation much more difficult these days. But who wouldn't want to build the next Powersoft?

Sunday, May 20, 2007

Q&A: SOA and BPM

I will be speaking in London on Tuesday about SOA @ HP to a group of 20+ European journalists and I am getting ready for some potential questions. One of the questions in my Q&A document is this one: What is the relationship between SOA and BPM?

I've heard this question many times before and every big SOA vendor has the same answer:

  1. SOA creates a set of services (ESB, network, fabric...) and these services are orchestrated by some clever BPM tool.
  2. BPM is for business people, SOA is for techies.

I absolutely disagree and this is why I was happy to read these two posts from Steve Jones yesterday. Here is a quote: "for most Services the concept of "Goals" will be more useful than the concept of "Process" ". It really brought back the memories of Radovan's controversial business process blogging in 2003 and 2004.

So I know how I want to answer the BPM question on Tuesday and I can move on to the next potential question: What is HP's acquisition strategy in the SOA area? Hmmm, I don't think I will be blogging about this one...

Friday, May 18, 2007

What Happened to Flexibility?

Peter Yared posted a very good article called "What Happened to LAMP?". He identifies the hidden cost of the new infrastructure software as one of the reasons why LAMP is not penetrating the enterprise.

I believe that it is the inability of current enterprise IT to support any change that makes any additions/replacements to infrastructure software so difficult. Last week I visited one of the largest global banks to talk about their plans for the new infrastructure software. And this bank has a policy in place that restricts the number of changes to their IT infrastructure - they can only deploy changes once a quarter and every such release breaks something and sometimes the damage is substantial. They actually showed me a report of emergency IT changes over last one year. It looked like this:

So every time they bring a new system in or change an old one the number of fixes goes dramatically up.

And until we solve the problem of sustainability of the current enterprise IT we can only dream about any new software infrastructure. And in the meantime we will be limited to four changes a year so that we have enough time to fix what broke...

Wednesday, May 16, 2007

Skype needs you

Let me be more precise: Skype needs you in Prague and only if you are a great software developer... The company plans to hire about 20 percent of the product-engineering workforce in Prague and I've learned about it from a huge Skype job advertisement that is posted next to the HP's sales office.

I am really happy to see companies to open their development centers in Prague. On one hand it creates more competition for the talent but on the other hand it produces more and more experienced people with global skills. So while Skype needs us we need even more companies like Skype...

Saturday, May 5, 2007

We love to print

A few weeks ago I had 100 of my Flickr pictures printed by Big MOO, Print Machine. Big MOO works for Moo Prints (more about the company here) and my pictures were printed on MiniCards - small cards approximately half the size of a normal business card. MiniCards are fun to look at, the website is well integrated with Flickr and the print service worked flawlessly. Little MOO, Print Robot did a really good job:

Hello Roman

I'm Little MOO - the bit of software that will be managing your order
with us. It will shortly be sent to Big MOO, our print machine who will
print it for you in the next few days. I'll let you know when it's done
and on its way to you.

In the meantime you can track and manage your order at:

Remember, I'm just a bit of software. So, if you have any questions
regarding your order please contact customer services (who are real
people) at:


Little MOO, Print Robot

MOO "We love to print"

Friday, May 4, 2007

South Park Mac vs. PC

This is what happens when you combine Get a Mac and the South Park:

The Economics of Abundance

Back in the nineties I worked for Powersoft and from time to time I had to go to Moscow for a business trip. And there you could have bought CDs full of software on a corner for a few bucks. I was very upset to see Borland, Inprise, Borland, CodeGear Delphi and not PowerBuilder on those CDs. It was obvious to me that once you take the price out of the price/performance equation it is only performance that matters. Nobody will copy an inferior product that is free and so the superior product can gain a big piece of a market share very quickly.

What I did not know then is that this one of the aspects of digital goods and it is called the economics of abundance. You can read more about it here and here.

Thursday, May 3, 2007

The New Czech National Library

One of my friends once told me that she prefers to stay in the Four Seasons Hotel in Prague. Apparently it is the only place in the city where the view of the castle is not spoiled by - the Four Seasons Hotel. In a few years she will have to read a book in the new library reading room (picture below) to enjoy an unspoiled view of Prague castle.

As you can see on the following pictures the new library building (in case it will be actually built) will be visible from everywhere else...

Wednesday, May 2, 2007

Does innovation matter?

Older and more experienced Radovan complains about enterprise IT that SUCKS. His point is that years of commoditization and vendor consolidation brought the costs down but it did not increase the business/IT alignment. Solution may come from a completely unexpected direction: startups. Read more about it in this Infoworld article.