In this post, I will be recounting my experience with Ubuntu server 6.10, WebLogic Workshop 9.2, and REST. Note: This in no way implies support of any kind from me or BEA. This is just the personal experiences of someone who can not keep his hands off the keyboard. This also gave me an opportunity to code up some REST samples, since many customer I come in contact with ask about it.

That said: I had a chance to play with the new Ubuntu server over the long weekend. I set up a server in my garage, and loaded the latest linux bits from WebLogic Platform. (Its actually labeled Portal 9.2, but Integration should be online shortly.) The first thing I noticed is that the installation did not complain at all about installing onto a non-supported platform. The second thing I noticed is that it just worked. I was able to start up a domain and get the samples server running almost immediately.

I have been working on Page Flows and XML Beans lately, trying to come up with some REST samples, since customers keep asking about it. Since I do most development from a laptop, I wanted to try setting up a remote server and see what it was like. I knew it was possible, since I read Chris Hogue's blog on remote debugging.  Chris has warned me that it would be slow, and he was right, but this is to be expected across a network. Stepping through individual statements was painful, but is was mind-rendingly cool to see.

I was basically able to create a web service and a page flow for the Yahoo! GeoCode service over REST. 

(Time for a momentary digression)

REST stands for "representational state transfer" and is "PhD-ease" for a network of pages, which represent state, and a client that uses data from that state to make decisions about where to go to next. Its an interesting take on the traditional finite state automata, or at the very least, an interesting application of it. (See the Wikipedia link above for more info)

(Back to the article)

I was able to easily catenate strings to create the request, and read the result, which I passed back as a String. From there, I used XML Beans' parsing mechanism to get the results into a workable form, based on Yahoo's GeoCode XSD, and then just returned it in the XML Beans generated type.

All of this was done from a Custom Control, which made it easy to then create a web service, (as shown in Chris Hogue's Web Services blog)  and then add the same logic to a page flow. Since Workshop understands the Schema, it was easy to create forms to search and present the data.

Not bad way to waste a weekend. If you would like the code, drop me a note.