SOA just vendor talk?

"‘SOA’ may have meant something once but it’s just vendor bullshit now."

Very interesting yet highly questionnable statement by Tim Bray. Without giving any reasonable argumentation for this bold statement, he dismisses SOA as something that does not matter anymore, and he says we should rather focus on what he calls Web style. Of course, the highly fashionable term Enterprisey is used in his post, and increasingly people are using this term to dismiss SOA without giving any argumentation for dismissing it, other than "SOA is Enterprisey". It is kind of like what open source zealots throw at you when you talk about MSFT: "oh, that's FUD!".

I do agree with Tim that simplicity is a virtue when working with the web. After all, the Internet has been such a big hit because of its simplicity. However, Tim misses the fact that the services he talks about when discussing the Web Style thing, are a completely different thing than the services we talk about when discussing SOA. I wrote before that there is something fundamentally difficult when talking about "services". There are different kind of services, and probably the name "web service" has not been the best choice. The Web Style kind of services Tim speaks about are fundamentally different from services you will usually use within an SOA. I wrote about this before too: WS-* services and REST services are not competing, they are complementary. Web Style serves another purpose than SOA. Consumer-facing services have other QoS requirements than high-volume, cross-platform A2A transaction services. One service needs to be simple and flexible, while another service should seamlessly integrate with different platforms. Different requirements ask for different services, which in turn ask for different implementations. That is the reality of today's IT world.

Once again: it is all a matter of perspective.

SOA just vendor talk?

"‘SOA’ may have meant something once but it’s just vendor bullshit now."

Very interesting yet highly questionnable statement by Tim Bray. Without giving any reasonable argumentation for this bold statement, he dismisses SOA as something that does not matter anymore, and he says we should rather focus on what he calls Web style. Of course, the highly fashionable term Enterprisey is used in his post, and increasingly people are using this term to dismiss SOA without giving any argumentation for dismissing it, other than "SOA is Enterprisey". It is kind of like what open source zealots throw at you when you talk about MSFT: "oh, that's FUD!".

I do agree with Tim that simplicity is a virtue when working with the web. After all, the Internet has been such a big hit because of its simplicity. However, Tim misses the fact that the services he talks about when discussing the Web Style thing, are a completely different thing than the services we talk about when discussing SOA. I wrote before that there is something fundamentally difficult when talking about "services". There are different kind of services, and probably the name "web service" has not been the best choice. The Web Style kind of services Tim speaks about are fundamentally different from services you will usually use within an SOA. I wrote about this before too: WS-* services and REST services are not competing, they are complementary. Web Style serves another purpose than SOA. Consumer-facing services have other QoS requirements than high-volume, cross-platform A2A transaction services. One service needs to be simple and flexible, while another service should seamlessly integrate with different platforms. Different requirements ask for different services, which in turn ask for different implementations. That is the reality of today's IT world.

Once again: it is all a matter of perspective.

The .NET boomerang effect

Every once in a while somebody posts a blog or article on how the monopoly of Microsoft / Windows on the desktop can be ended, or at least be threatened. Of course I am also guilty of this, although I have said before that it will be quite hard to threaten Windows on the desktop. Keith Harrison-Broninski recently had a quite interesting discussion on his web log, and although I do disagree with him when he states that Eclipse will end the Microsoft monopoly on the desktop, his viewpoint sure made me think about other possibilities, stemming from the open source (.NET) community: Mono.

More and more Linux distributions have the open source .NET framework Mono incorporated, which means that these Linux distributions can run .NET applications that comply with the ECMA standard for .NET. Although key distributions like RedHat for instance do not include Mono (yet?), there are still signs that the Mono framework is becoming an increasingly important factor for Linux development.

I said before that one of the key barriers for a switch to a Linux desktop are the tons of applications that will only run on Windows. I also said before that a Windows emulator (such as Wine) could be helpful to run those apps, but the number of supported apps is still very limited.
As it is in Microsoft's best interest to have as much applications running on the .NET framework as soon as possible (to acceleratie the expansion of the .NET framework and runtime), many organizations will be forced to migrate their applications to this .NET framework. Which in turn makes it possible to run those applications on a Linux distribution with Mono included. So it could turn out that the rapid expansion of .NET that Microsoft is pursuing, could work as a boomerang as it opens up the possibilities of Linux (with Mono) ending Windows' desktop monopoly. And just like it is the case with Ajax, it could turn out that Microsoft started this themselves, by submitting their .NET stuff to the ECMA. Ironic again.

JBoss not Fusion-ized

Despite Larry Ellison's attempts in the past, it is now safe to say that JBoss will not be Fusion-ized: today the news was announced that RedHat has acquired the open source application server producer for 350 million dollar.

When Ellison attempted to buy JBoss earlier this year, they boldly stated they were not for sale. However, RedHat succeeded in acquiring the open source company.
I think this news will create some disruption in the open source world, as RedHat competitor Novell uses JBoss database software in its Linux distribution Suse, a key competitor for the RedHat Linux distribution. However, as JBoss is open source of course Novell will still be able to use it, but the open source model sure can create some peculiar situations, especially as at some points a product / project is acquired by a vendor.

So it is now up to RedHat to make a tasty fusion of RedHat and JBoss capabilities. And Oracle will probably target Zend even more after failing to acquire JBoss.

Taylorism in IT

Recently Gartner started a blog named Unconventional Thinking, and I was particularly triggered by a posting by Richard Hunter. He expects that although the concept of scientific management dates from the early 20th century, there is great potential in it for IT to change how executives run the show. He thinks the next big wave is automation that changes the way managers, not frontline employees, work.

Personally, I do not think this is going to happen because, in fact, it already is happening. Like with so many changes that involve IT, our IT organizations show the first signs of such a type of Taylorism. The most eminent example of the practical application of scientific management or Taylorism has always been the assembly line. And what have we done the past years in IT departments? Exactly, we have created assembly lines.
What do we want when managing IT departments? Speed and predictability. Exactly the same keywords that most are used throughout scientific management.

This of course applies particularly to the operational processes of IT (some might argue that this applies for the people that do the real work). However, at the tactical and strategic level there has also been an increase in Tayloristic management principles. Corporate Performance Management (CPM) and increased data mining capabilities fuel the more quantitative way in which departments are run. Managers and their superiors agree upon key performance indicators (KPIs), and are increasingly evaluated against those.

I think technology areas such as business intelligence and service-oriented applications will benefit greatly from this evolvement, as it relies heavily on having the right information at the right time.
The increased transparency in organizations and processes allow an increased emphasis on delivering results at the highest possible level of efficiency. This will not only be for the normal workers, but also for management.

However, one of the founding fathers of this approach (actually the synonym was named after him) Frederick Taylor does not have the best of perceptions of mankind:

Now one of the very first requirements for a man who is fit to handle pig iron as a regular occupation is that he shall be so stupid and so phlegmatic that he more nearly resembles in his mental make-up the ox than any other type.

All I can say is that for some reason I feel that this perception of people does not fit very well with the current culture in most organizations. Also, transparency has an undeniable other effect on organizations: there is no longer an information monopoly for top management, in fact there are numerous examples where the people on the work floor outsmart the executives, because they are better informed. So my prediction is that scientific management will certainly become an increasingly popular concept in IT, but the technology it needs for further expansion, will also limit its impact, as the mirror has two faces.

ajaxWrite, LimeWire and the digital divide

First there was the news last week that the much talked about web-based word editor ajaxWrite has nothing to do with Ajax as it is fully written in Firefox-proprietary language XUL (the fact that it only runs in Firefox should have raised some suspicion...). Typical case of randomly selecting a hyped term and incorporating it in the name of your product. The web service tooling arena has suffered from this before, when every small vendor incorporated the term SOA in the name of its products, and now it is time for Ajax to suffer from these effects.

Then this morning when I was stuck for 2 hours in a traffic jam, I heard the news on the radio that a Dutch newspaper took the test, and searched on the Internet for confidential information people by accident had put on the web through file sharing programs such as LimeWire. It was shocking what they found within the hour: passwords, tax declarations, application letters, resumes, passport scans and even proposals for the Dutch national security agency and the Ministry of Justice and Defense. All this by just searching for the term "confidential" (but then the Dutch word for it of course).

These two news items, that do not seem to have much in common, show that a lot of people have no clue what they are doing or what they are dealing with. As with the Ajax case, a lot of (techie) people will think they have a cool Ajax app running, while in fact it is just as proprietary as MSFTs Word for instance. The results of the ignorance of the file sharing capabilities are worse however. I am sure that all those persons put their confidential data on the web, just thought that it would be cool too if they could download the latest music or movies just like their neighbor kid, and they never intended to throw half of their private life on the net.

Everytime when I hear such news, I get the feeling that maybe in some aspects we are too far ahead in IT with our newest technology stuff. A lot of people are having trouble catching up with the latest evolvements, and because sometimes they are not fully aware of the possibilities and power of technology (as with the file sharing item), the results can be quite bad. I wrote about this before, and not only within IT we can find the problem of a divide between the front runners and a large group that has trouble catching up, but this is also the case on a macro level, in society. This is often called the digital divide, and with the increasing consumerization of IT, which will be driven by a young group of very tech-savvy consumers, I think this divide will become even greater. The results will be likewise.