Category Archives: 8950

Our cries have finally been answered then?

I for one am very happy that the next version of Visual Studio will finally embrace what some of us have been requesting for a long time now – UML adoption. A few months ago I mentioned some stuff that would be in the next version of VS and now that more has been going on I am thoroughly pleased.

For my part I’ve not really looked too much into what’s going on with the class diagrams etc and whether or not we will be graced with the standard l&f for UML diagrams. I for one hope that this does happen.

This is way overdue and both Eclipse and Netbeans have UML incorporated into them to a certain degree. Netbeans has pretty good UML support right now in fact. Of course the great thing about the addition of these UML features is that we will be able to get a more abstract view of our code and the way various components work with one another by reverse engineering it. Maybe this might not seem so helpful now, but when you revisit old code or encounter code that isn’t yours it will be.

I know that the VS team have had a tonne of stuff to do but why have we had to wait so long for tools which are pretty much essential? We’ve been palmed off to use things like Visio or fork out the big bucks for more comprehensive software.

Well, at least the wheels are finally in motion. Hip hip 😉

Before I go be sure to check out Channel9 this week for more insight on what’s coming up in VS2010. I briefly watched a video today just to get the gist of what has been added since I last looked and it was pretty decent.

Software pipelines

There is a pretty interesting architecture based on software pipelines that will at least allow your systems to utilize many-core processors in a far easier manner than other alternatives. This approach has a lot in common with fluid dynamics in that some principals are the same.

All service requests hit something called a pipeline distributor. The pipeline distributor routes the request to the appropriate pipeline. What is pretty nice about this approach is that the pipelines within a pool can delegate requests to other pipelines within the same pool  or even delegate requests to another pipeline distributor that maps requests to a different set of pipelines within a different pipeline pool.

I can’t do this approach justice myself via a simple blog post but upon further searching I came across a rough cuts book named Software Pipelines: The Key to Capitalizing on the Multi-core Revolution. I’ve quickly skimmed the book just now and it looks as though it covers the subject quite well.