You’ve worked so hard on F# 3.0, it’s been an incredible pleasure and honour to work on F# 3.0 with you and the team!

Don

Is F# Async a killer app on the server-side?
I’ve seen a number of articles, and understand how it fits.
Has C++,# adopted Async?

What is the killer app for F# Async Rx?

Refolding Planar Polygons does computational geometry, but needs polygon input and display animation. F# is good for the CompGeom, not so much UX input (but I’m hoping to figure out Async, and Async Rx for UX), and WPF is good for the display animation. With all the new surfaces, cloud, cluster, client, etc. I’m having difficulty keeping up with the road maps.

I like the cute little demos. More please.

Best in ’12!
PLH OOE Art

I’ve had quite a problem tracking down the reason for frequent crashes of my VS2010SP1. I finally found this in the Activity.Log:

801
2011/09/24 16:55:13.943
Error
Editor or Editor Extension
System.ArgumentOutOfRangeException: Specified argument was out of the range of valid values. Parameter name: bufferPosition at Microsoft.VisualStudio.Text.Editor.Implementation.TextViewLineCollection.GetCharacterBounds(SnapshotPoint bufferPosition) at FSharpDepthColorizer.FullLineAdornmentManager.RefreshLine(ITextViewLine line) at <StartupCode$VSIXColorizer>.$MyAdornmentManager.-ctor@133-23.Invoke(TextViewLayoutChangedEventArgs e) at Microsoft.FSharp.Control.CommonExtensions.SubscribeToObservable@1777.OnNext(T value) at Microsoft.FSharp.Core.CompilerServices.RuntimeHelpers.h@744.Invoke(Object _arg1, TArgs args) at <StartupCode$VSIXColorizer>.$MyAdornmentManager.-ctor@133-22.Invoke(Object sender, TextViewLayoutChangedEventArgs e) at Microsoft.VisualStudio.Text.Utilities.GuardedOperations.RaiseEvent[TArgs](Object sender, EventHandler`1 eventHandlers, TArgs args)

Perhaps you can track it down? I hope it is not a freak colision with some other extension that I am using.

I must say you missed the chance to mention catamorphisms on mutually recursive types (I did not even find much _academic treatment_ on that), which would have been the case for the combination of Expr and Op, had you not delegated Op to the lower status of dumb data operated upon by the Expr stuff. You often state that generating catamorphisms is type-driven and can be automated, but any automatic processor of types would have to treat Op as a regular union type as well. To be fair, I’m not even sure how it could be done, but I’m going to try using catamorphisms upon a real AST comprised of tens of mutually recursive types, from a project of mine, and see how it goes.

I’d also like to leave a word of caution for future readers of the post series: you frequently state how using continuations makes the code tail-recursive, but, if one tries out the code, it appears that this does not happen in Debug mode, which is the usual mode one uses in order to play with – and step through – sample code.

for example

member X.deltaCompareArchive l1 l2 typenumber =
inconnection
match typenumber with
| 0 ->
Seq.filter(fun x -> x.ID_Line = l1 || x.ID_Line = l2) @>
| _ ->
Seq.filter(fun x -> (x.ID_Line = l1 || x.ID_Line = l2 ) && x.DataType = typenumber ) @>
|> fun pquery ->

|> query |> Seq.choose id
|> Seq.collect(fun a -> [fst a; snd a])
|> Seq.distinct
|> List.ofSeq
|> fun p ->
if p.Length = 0 then null
else
let STRackondeltas =

so my |> fun p -> breaks colorizer

> type MyVector(x : int, y , z) =
- let values = [| x; y; z |]
- member this.Values = values;;

> let v = MyVector(1,2,3);;

val v : MyVector

> v.Values.[0] printfn “%d” v.Values.[0];;
42
val it : unit = ()