The concept of excellence

Moved this post to a blog of my own. People wanted to comment and I could not change these settings...

http://fabioeidi.blogspot.com/2010/02/concept-of-excellence.html

Lots has happened

The topic is really the Rich Internet Applications (RIA) which can now receive data asynchronously using XMLHttp objects in JavaScript (AJAX).

There are a few contenders:
1. Google Web Toolkit (GWT). Using Java on the server side and for design time.
2. Adobe FLEX which requires the Macromedia flash player. Here is a really cool application, from Google of all players, using FLEX. Looks like a neat trender.
3. Microsoft ATLAS which uses ASP.NET on the server and AJAX. So far all the controls I have seen make way to many trips to the server.
4. Microsfot Silverlight. This requires a plugin which loads a mini version of .NET on your browser. It run in Mac OSX and Linux and works with most browsers not, just IE. Maybe this why Apple is now offering Safari on Windows. Silverlight puts Vista-like presentation into the browsers.
5. Various Javascript frameworks such as Prototype. Most of them make use of JSON which encodes the data in Javascript objects versus XML. It is faster.

All this to be real-time and scalable requires that the data be published using techniques such as publisher/subscriber. HTTP by default is not that way. So enter the COMET technique. I am working on getting this working in .NET using HTTPHandler (.ashx) pages. More to come later.

Process plant of the future

Here's a stream-of-consciousness post from a guy who is hardly a process industry person. But one who's been around more than a few refineries and 'seen' things.

First, some observations.

- The oil and gas industry has money coming out of their ears (and noses and other orifices) right now. You'd never know it by talking to them. They are getting a lot smarter about what they do with these massive windfalls of money when they come, and they certainly are not using it to pump-up employee morale or benefits. They are using it to buy their competitors, consolidate and streamline. Streamline is a way of saying, less people, more automation.

- Ultimately, it does not matter if there are 20 years or 200 years of petroleum left in the ground, the writing is on the wall. Not only are carbon-based resources becoming politically-incorrect (all the while we become ever more dependent on them) it has become apparent that the parts of the world that buy the oil from the people who own it are enriching tyrants and terrorists alike. No one likes to voluntarily pay for their own contract to be assassinated.

- Although the o&g industry isn't in any mood to hire more warm bodies that they have to pay for, they are very interested in what the process plant of the future will look like. They have the money to spend on it -- whatever it will be. They will even allow some lattitude in entertaining rather grandiose schemes and fragile technologies, all in the name of automation research. In the end, they will pick some new companies/technologies and some old companies with new technologies and go forward. This will happen, I predict, in the next 4 to 6 years, with real results showing up not long after.

- No one knows--yet--what the process plant of the future will look like. The oil refineries being built today (most conspicuously in China) are being built pretty much like the ones of yesterday. Don't look to them for cutting-edge technological changes. But there are a lot of people starting to pay attention to the next 'big thing'. Money has a way of focusing your attention.

I believe that the process plant of the future, among other things, will reverse the process information paradigm. The information that is currently being collected and stuffed into data silos will take on a life where it flows into smart systems. Interesting tidbits will be harvested immediately and the systems will analyze them and choose one or more intersted humansd to inform. After the humans are informed of the interesting information, they will be queried in such a way to find out, "was that helpful? If it happens again, do you want to know about it--should I inform someone else--or should I just not bother?" And in the process, the system will learn from the humans. This ain't science fiction -- it's science fact. The newest jumbo jets (and to a large extent the old ones) are not dependent upon human pilots to acknowledge every instrument reading for the duration of the flight. The systems are designed to inform the pilots of the most "interesting" readings so that pilots can choose to act or not. As the planes become bigger and more complex, they also will incorporate 'learning systems' where based on information that has been qualified, only a portion of the events taking place in the plane (or outside of it) will "warrant" the intervention of a human. This is exactly where process systems will go. They have no other choice because of the stated goal of process companies to reduce the need for humans. Manufacturing companies have started down this road a long time ago. Once you build a robot that performs a series of actions repetitively, the next step in its evolution is to have it start learning and doing its own thinking. There are a *lot* of robots in the world right now, only they are in factories and plants. We don't see them performing--tirelessly, repetitively, without a care or bother in the world. Soon, we will start seeing them. But that's for another blog at another time.

There's lots of money to be made in the process space right now, if you are oriented towards the next generation of process information systems. And can provides answers to the riddle: How do we reduce the need for the fleshy types?

What about Process Information Systems?

Since I live and breathe the Process Industry, I have been pondering what can the future look like? Is there a new breed of systems in the future?

I think there are and I would like to see everyone's collective ideas.

What I see:
1. Process Historians are old
2. The do not handle the entire gamma of information
3. They certainly cannot deal with relational data well and are missing the boat on unstructured information.
4. They fall short on delivering a model that is not Tag based.

Where I really come from

volcanbaru
Originally uploaded by mario brenes.

A picture of the Volcan Baru in Chirqui, Panama

Where I come from

Spen the day figuring out that there is no browser-agnostic way of doing graphics in JavaScript. the best I have seen is "Canvas" which Apple and now Mozilla support. It is called "Canvas". Apple Canvas Technology
Why you may ask, because it is an alternative to Applets, Flash and SVG, which all require browser plugins.