Wednesday, October 26, 2005
CNN.com - Web-based software challenges Windows - Oct 25, 2005
More on AJAX and its sudden rise in popularity. All this thanks to Google according to the article. I can certainly think of some potential apps for AJAX development. webMethods administration is all browser based. There is a lot of potential there.
Tuesday, October 25, 2005
Computerworld | 'Nightmare' drove desperate user to open source
Well this is one way to do it. I can only imagine trying this at my company. I don't know if I would describe Windows servers as a nightmare, but they certainly can be expensive and difficult to maintain. Of course so can any OS in the wrong hands.
Tuesday, October 18, 2005
New hopes from Sun's idea factory | CNET News.com
Originally saw a reference to this on Slashdot. Sun seems to be slowly turning the ship around. I think the Google announcement, although lacking in substance, really helped with some image issues.
Monday, October 10, 2005
David Kirkpatrick - Sun and Google: Publicity Ploy or Power Grab? - FORTUNE
Interesting, very interesting.
Friday, October 07, 2005
What is Proprietary?
I have been thinking a lot about this subject lately (a lot is all relative). Two subjects that are getting a lot of air time these days are Open-Source software and SOA's. One word that seems to come up a lot in any conversation about SOA's and Open-Source software is "Proprietary". So I asked myself, what is proprietary and should I care? I am just going to talk about SOA's for now. I will address Open-Source at a later time. You can hardly wait, I can tell.
Some of the new upstart SOA proponents warn against using proprietary technology in your SOA. The finger is quick to point to the evil expensive proprietary traditional EAI vendors as what is wrong in the world. Standards based solutions are the way to go, otherwise you get locked in and interoperability becomes an issue. These vendors typically also point out that there solution is 100% standards based and they do not use any proprietary technology.
So what's the truth to that claim? Are you going to get locked in with a traditional EAI vendor? Are interoperability issues going to plague your until retirement and be passed on to your children? The answer is yes and no. So let's add another concept, it is called the provider. Let's use Java's JMS (Java Messaging Service) as an example. JMS is really nothing more than a series of interfaces and api's with no real meat. It is up to a provider (SonicMQ, webMethods, Tibco, MQ Series, etc) to provide the meat for JMS. The interfaces and api's are contracts that more or less bind the provider into providing a set of services that are standard across the board.
Just like in life however, contracts are open to interpretation. Each vendor selects and interprets which set of api's and interfaces they are going to implement and support. So your nice JMS client over here may not be your nice JMS client with a different provider over there. So while the idea of standards sounds nice, the reality is a little more complicated. It is really up to the architect and the developers then to know the standards very well and know when the provider is deviating from them. Otherwise your highly touted portable service becomes something more akin to an anchor.
This brings me to my point (you thought I would never get there). Underneath the covers there is always a provider. And swapping out providers is almost never an easy thing to do. Spending more time on the principles of loose coupling, abstraction and common data models regardless of the technology provider will pay off in a much bigger way than spending time chasing ever moving standards and the vendors that are touting them. There are no bad standards, just bad providers. Okay I was trying to use the dog analogy but it doesn't really work because there are some bad standards. The point is, standards are a good thing but they are just a part of the total solution.
Some of the new upstart SOA proponents warn against using proprietary technology in your SOA. The finger is quick to point to the evil expensive proprietary traditional EAI vendors as what is wrong in the world. Standards based solutions are the way to go, otherwise you get locked in and interoperability becomes an issue. These vendors typically also point out that there solution is 100% standards based and they do not use any proprietary technology.
So what's the truth to that claim? Are you going to get locked in with a traditional EAI vendor? Are interoperability issues going to plague your until retirement and be passed on to your children? The answer is yes and no. So let's add another concept, it is called the provider. Let's use Java's JMS (Java Messaging Service) as an example. JMS is really nothing more than a series of interfaces and api's with no real meat. It is up to a provider (SonicMQ, webMethods, Tibco, MQ Series, etc) to provide the meat for JMS. The interfaces and api's are contracts that more or less bind the provider into providing a set of services that are standard across the board.
Just like in life however, contracts are open to interpretation. Each vendor selects and interprets which set of api's and interfaces they are going to implement and support. So your nice JMS client over here may not be your nice JMS client with a different provider over there. So while the idea of standards sounds nice, the reality is a little more complicated. It is really up to the architect and the developers then to know the standards very well and know when the provider is deviating from them. Otherwise your highly touted portable service becomes something more akin to an anchor.
This brings me to my point (you thought I would never get there). Underneath the covers there is always a provider. And swapping out providers is almost never an easy thing to do. Spending more time on the principles of loose coupling, abstraction and common data models regardless of the technology provider will pay off in a much bigger way than spending time chasing ever moving standards and the vendors that are touting them. There are no bad standards, just bad providers. Okay I was trying to use the dog analogy but it doesn't really work because there are some bad standards. The point is, standards are a good thing but they are just a part of the total solution.
Tuesday, October 04, 2005
Sun Microsystems announces a partnership with Google
I think this is good news for Sun. I not really sure how earth shattering the announcement is though. Downloading the Google toolbar with Java doesn't seem to make a whole lot of sense. Just a forecast of things to come maybe.
Saturday, October 01, 2005
LinuxPlanet - Reviews - Opening Solaris - The Solaris 10 Pitch
Just read this review of Solaris 10. I have not had a chance to get my hands on 10 yet. The review points out some pretty cool new features. The one I like the best is Dtrace. If you have ever used Solaris you know that the truss command can be pretty useful in debugging programs. But it had some limits. Dtrace takes this concept much further. Being able to look inside your program while it is running can save a lot of time when it is misbehaving. Things like why aren't my libraries loading, what's it looking for etc.
The reviews goes on to mention things like self-healing(new Solaris service manager), virtualization(containers), and very quick boot times ( although I never really saw the need for this, my servers rarely have to be booted except for patches)
If you are looking for a good server operating system, Solaris continues to be at the top of the heap. I like Linux ( the blog is hosted on a Linux server at my house) but I don't think it is as far along as Solaris. No doubt though that Linux's capability will continue to rise.
The reviews goes on to mention things like self-healing(new Solaris service manager), virtualization(containers), and very quick boot times ( although I never really saw the need for this, my servers rarely have to be booted except for patches)
If you are looking for a good server operating system, Solaris continues to be at the top of the heap. I like Linux ( the blog is hosted on a Linux server at my house) but I don't think it is as far along as Solaris. No doubt though that Linux's capability will continue to rise.
Subscribe to:
Posts (Atom)