RafaelF's blog - All Comments

re: Various implementations of Rx

Christopher Harris — Fri, 22 Mar 2013 20:00:22 GMT

I'm not sure we can put Pex and Moles in the same category as Rx. Last I checked, Rx is pure .NET.

re: Various implementations of Rx

Crispin Wright — Tue, 19 Mar 2013 23:06:21 GMT

"How far away are we from an Rx Summit and a committee for standardization?"

Probably 1 week further away than we are from the VS product team sucking in the work and making it available in the VS20xx "Ultra Composable Premium Edition" only......Just like they did with Pex and Moles.

re: How to express common windowing patterns using Rx

Rafael.F — Tue, 22 Jan 2013 00:35:36 GMT

Not sure what you mean by "optimisations available through Rx". If you're asking whether Rx is implemented to be as fast as possible over observable collections, you can look at the code here: http://rx.codeplex.com.

For things like variance,I usually compute them over the output of a Buffer, which is an IList<T>. Therefore, a simple extension method works :)

re: How to express common windowing patterns using Rx

Uzair — Mon, 21 Jan 2013 23:56:39 GMT

When calculating a average or a standard deviation over a sliding window, for example, one typically implements a circular buffer so that even if the calculation isn't online, data manipulation is kept to a minimum for speed purposes. That is, no new memory needs to be allocated and only a couple of operations are needed (insert the new element and update the index that marks the start of the buffer). Are any such optimisations available through Rx?