Matevž Gačnik's Weblog

I'm currenlty involved in designing a client application for a big hotel chain. The application is able to show rich media (MPEG II digital TV, MPEG II video on demand, Windows Media video on demand, WMV-HD, music, live webcasts, you name it...).

The problem we faced on friday was connected to the way WM services allow a client (WM Player 9) to cache streamed contents. It was actualy not a functional but rather a performance problem, because the application architecture is designed in a way that allows client computers to reflash their HD partition at every reboot. Since a lot of video content was cached (hundreds of megs), reflashing time increased.

Two ways to get rid of client side caching:

1. Allow client side caching property in Windows Media Services / Windows Server 2003
2. URL encoded parameter in query string (ie. http://wmedia/mymedia.wmv?wmcache=0)

This got my attention today:

Ultimately, I think the question is really a distraction. One of the great strengths of XML is that the instances exist independently not only from individual schema definitions, but also independently from the schema language of the day. [From: Don Box]

True indeed. In case of industry shifting to RelaxNG (very unlikely), instance documents would survive nicely. As long as there is a schema, that describes an instance document, everything is fine. When the connection is lost somehow, we can't talk about instances any more.

XML without a defined schema is no better than CSV. It's not even easier to parse.

I just bought a new George Michael single - Amazing, which is distributed as a Windows Media 9 audio using DRM protection. The process works flawlessly:

1. Enter credit card number, pay $3
2. Get PIN number
3. Download the single
4. Open the single using WMP9
5. Go through WMP personalization process, get license
6. Enter PIN
7. Store license
8. Play

This gives me more control of what I own. Bye bye ITunes...

We'll see how this works out. I have been at the PDC, seen Doug's talk and I agree that this allows schema to be versioned over time. What is bothering me is the structural extension of the schema itself, just to support versioning.

And yes, I know this is the only way, since W3C didn't pay attention to versioning in the first place. It still bothers me, since I like my content models clean.

What happened with Clemens?

It was a nice show (I watched via live Webcast), but someone was missing. :)

Well, until today (and I don't know why) I had a concept in my head, that WSDL operation names need to be unique.

I'm really glad that guys from a big local mobile operator helped me clear things up (actually, they helped me lead myself into a no-exit alley). It happened today, during my web services day, a part of a bigger .NET training course.

I reread the spec. It says:

Since operation names are not required to be unique (for example, in the case of overloading of method names), the name attribute in the operation binding element might not be enough to uniquely identify an operation. In that case, the correct operation should be identified by providing the name attributes of the corresponding wsdl:input and wsdl:output elements.

Polymorphic methods are my friends again. I was mislead about the WSDL serialization details in that scenario.

Enough said.

Christian Weyer admits he's not the only one thinking that SoapScope is that great.

My priorities go like this:

1. Excellent Testing Tool
2. BP 1.0 Complience Verifier
3. Great VS Integration

Keep it up.

It happened. Mindreef released a new version of its excellent tool - SoapScope 3.0.

Web services diagnostics has never been so efficient.