Posts tagged fsharp

Are you tired of generating strongly typed classes with millions of lines of C# using crmsvcutil.exe that often crashes? Fed up with having to re-generate the classes every time something in the schema changes? Feeling restricted by the LINQ provider's limitations? Ever wonder why you should need to know what attribute joins to what in your relationships to perform your joins? F# to the rescue!

I am working on a F# type provider than aims to solve a lot of these pains, and more besides! This is just a sneak preview with what is to come, I have not released any of this yet. First, here is an example of a rather silly but fairly complex query using the new type provider and the F# query syntax :

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19

type XRM = Common.XRM.TypeProvider.XrmDataProvider<"http://something/Organization.svc">

let dc = XRM.GetDataContext()

type TestRecord = 
    { x : string; 
      y : int }

let test = { x = "John%"; y = 42 }

let q =
    query { for s in dc.new_squirrel do                            
            where (s.new_name <>% test.x || s.new_age = test.y)
            for f in s.``N:1 <- new_new_forest_new_squirrel`` do
            for o in f.``N:1 <- owner_new_forest`` do
            where (s.new_colour |=| [|"Pink";"Red"|])
            where (f.new_name = "Sherwood")
            where (o.name = "PEZI THE OWNER!")
            select (s, f.new_name, f, o) }

I currently do a lot of Office type automation work where I scan a bunch of email from exchange, download excel attachments, open and transform a bunch of data from them, reconcile these against databases using FLinq, produce graphs and charts with the results using FSharpChart, and so forth.

(p.s - as a side note, F# is awesomeat doing this kind of thing, I can knock all kinds of stuff out super fast. p.p.s - Active patterns with Excel = win)

As anyone who has done any office automation will know, cleaning up the COM objects is a right pain. Any object you bind to a value be it a workbook, sheet, cell, range or anything at all has to be explicitly freed with Marshal.FinalReleaseComObject. If you are not really careful with this you will end up with memory leaks, and EXCEL.EXE (or whatever) running forever in the background even once your program has shut down. The problem actually goes further than this, if you use the . . . notation to traverse the object models, all the intermediate objects are also bound and need freeing up. However, these ones as they were not explicitly bound you can get away with calling GC.Collect() andGC.WaitForPendingFinalizers().

You can see any number of articles on the interwebs about this issue and all solutions are ugly. F# to the rescue! Here is a simple computation expression I use to deal with this problem. It is not optimal but the interop is horribly slow anyway so it makes no difference.

It's been forever since I last posted, I worked quite a bit on PezHack and then stopped for a while. I'm back to it now. In this post I will describe a technique I used to greatly reduce the amount of code and abstract away some repetitive imperative code.

The Problem

PezHack is a turn based game. The screen does not re-render itself all the time like a real-time game, but only when something changes and at the end of a the player's turn. The agent-based approach I used to separate the various sub systems of the game allow me to provide isolation around the graphics processing. The graphics agent is totally responsible for knowing what it needs to draw and will draw it on demand when it receives the relevant message. It does not really know anything else about the game state at all except the visual data of the various tiles that are visible, some player data that allow it to draw the various data about the player, and any menus / other UI that it needs to draw. Other systems can send it messages to inform it of some new visual state it should care about.

Most actions that Pezi can perform require some form of additional input and conditional flow logic. For example, when you press the 'e' key to Eat, first a menu is displayed that shows the stuff in your inventory that is edible, if possible. If not it will display a message and not end the player's turn. Assuming the menu is displayed, the player then presses the key that relates to the item they wish to eat from the menu. If this key is invalid a relevant message is displayed, else the item in question is eaten. What then happens is dependent on the item, it might provide sustenance, or if it's a mushroom it could perform one of various kinds of effects, and then probably end the player's turn

This is a common pattern that appears everywhere, and at each stage the graphics need to be re-drawn to show the various messages, menus and so on. At each stage it might continue or it might not depending the player's input, and the returns differs at each stage (end the player turn, or not, for example). In an imperative language we would almost certainly model this control flow with a bunch of nested if/then/else statements which quickly gets ugly. Because I am using the agent based approach for the graphics, I would also need to post a request to the graphics agent each and every time I needed the changes to show on the screen, so suddenly the actions become a list of imperative statements peppered with common code to keep the screen updated.

In my quest to learn the functional paradigm, one thing I have struggled with is game development. Assuming I mostly stick to the functional style of having little to no mutable state, how do you go about writing games? Games are pretty much ALL mutable state. Obviously, in a multi-paradigm language like F# you can have mutable state - and if used judiciously this can be very effective (not to mention offer some significant performance improvements). The blend of imperative and functional styles can indeed work well, I wrote a few small games with XNA and F# using this approach. However, I am still more interested for educational value to stick with a more pure functional approach. Along they way I have wrote a few small games such as a functional console based Tetris in about 300 lines, and a two player console based PONG clone (using the libtcod library as I will introduce in a bit) that uses the Windows Kinect as the input (was cool!) In these programs I would tend to have the game state formed with an F# record that is copied/modified on each game cycle, passing the new state back into the loop. This works well and both of these games used no mutable state at all. This approach soon falls down with something more ambitious though, you can't realistically propagate the entire game state through one loop.

The Roguelike

I decided to create something a lot more complex whilst attempting to stick with the functional guns. If you don't know what a roguelike is you can read all about them here and a great set of development resources with links to many on-going roguelike efforts here. I used to play these games a lot. Stemmed from D&D back in the days where the only computers were the terminal sort in colleges in universities (before I was alive!), a roguelike traditionally uses just ASCII characters as its graphics, leaving the rest to the player's imagination. If you haven't tried this before I highly recommend you try one, Nethack is probably the biggest most complicated one out there, but you can also play the original Rouge (where the roguelike genre name comes from) online here. A few things that make a roguelike;

• Random procedurally generated dungeons - every time you play there is a new dungeon
• Randomness in item drops - until things have been identified, you don't know what they are. The "clear potion" might be a healing potion in one game and a potion that causes severe hallucinations in the next
• Peramadeath and hard difficulty. These games are hard. Expect to die lots, and when you die you have to start again. Often the objective isn't to finish the game, but just see how long you can survive
• Roguelikes are usually turn-based affairs - although there is some variation with this
• Complexity - these games are amazingly complex and deep. The developer(s) don't have to worry about impressive graphics engines, so they can focus a lot more on interesting game mechanics. You will be amazed at some of the stuff you can do in a game like Nethack.

Here's a picture from Nethack to illustrate how a typical roguelike looks :

I’m sure every seasoned .NET developer has been in the situation at one stage or another, probably in testing code, where they need to access a non-public setter of a property (or maybe a private member), and it can’t be mocked.  We all know the (somewhat scary) reflection trick to get at the said setter method and invoke it.   I just hit this problem today trying to mock some response messages from a Microsoft Dynamics XRM 2011 OrganizationService.  Thankfully F# gives us cool things like Quotations and symbolic functions (custom operators) to make this process more succinct.  Instead of writing a method to reflect on a type, get the relevant method, then finally invoke the method and pass in both the original object instance and the new value being set, we can do the following:

1
2
3
4
5
6

let (<~) property value =
	match property with 
	| PropertyGet(Some(Value(x)),ri,_) -> 
	    ri.GetSetMethod(true).Invoke(fst x,[|value|]) |> ignore 
	| _ -> failwith "unsupported expression"

In this article we will discover how to emulate simple logic gates, and then build them up to form more complex circuits. By the end of this 2-part article (edit - not that I ever wrote part 2) we will have created:

• - Half-bit adder
• - Full adder
• - n-bit ripple carry adder formed from full adders
• - 4:1 line decoder
• - 4:1 multiplexer
• - 1 bit ALU that supports addition, subtraction, AND and OR operations
• - n-bit ALU formed from 1 bit ALUs