Thursday, October 25, 2007

Morning Coffee 120

• Doing these morning coffee posts is a lot tougher since I cut back my blog reading. Where I used to have no trouble finding 4-5 coffee-worthy items every day, these days I seem to only get 1-2, if that.
• After starting off 3-0 and 100% on the PK, the Caps dropped four in a row and have been miserable on special teams. The special teams woes continued last night against the Lightning, but they still won. Caps went 0-4 on the powerplay, and coughed up a short handed goal. But they also went 3-3 on the PK, so I guess it wasn't all bad. Maybe my mother will stop calling for Hanlon's job now. It's a long season and as Peerless Prognosticator points out, the rebuild isn't over.
• Jomo Fisher, who helped Scott Hanselman auto-merge assemblies, has been digging around in F# of late. As it turns out, he's joining the F# team so I'm thinking it's not a huge stretch for him. If you're a C# developer trying interested in getting a handle on this new F# thing, his blog is a good place to start.
• Speaking of F#, Don Syme posts about yet another new F# feature: Async Workflows. Workflow is a bad term here IMO since it can be easily confused with WF. Regardless of it's name, Async Workflows is about making .NET's Async Programming model a first class citizen in F#. Robert Pickering has a good post explaining how this new feature works.
• Microsoft sure has a lot of multi-threading / async-programming tools coming out. In addition to F# Async Workflows, there's the Concurrency and Coordination Runtime, Parallel LINQ and the Task Parallel Library. I would hope all this work eventually coalesces as a coherent product offering.
• Now that F# is being "producized", I wonder if the language evolution will slow down. Async workflows were introduced in F# 1.9.2.9. Other recent changes include Computation Expressions (v1.9.2), Use Bindings (v1.9.2) and Active Patterns (v1.9.1). F# seems to churn more in minor releases than C# does in major releases. Of course, that's because F# was a research project, not a "real" product. Now that it's going to be a product, will the rate of innovation slow?