F# in 20 Minutes - Part I

chris core

chris core — Sat, 03 May 2008 11:25:48 GMT

PingBack from http://www.bamor.net/chris-core-18791

re: F# in 20 Minutes - Part I

Sat, 03 May 2008 23:23:30 GMT

Interesting posting!

It seems to require the full version of VS, though. How would one go about using it with free Express editions?

Learning F#: Some great blog posts from the F# team

Don Syme's WebLog on F# and Other Research Projects — Sat, 03 May 2008 23:31:12 GMT

Brian and Chris have been writing some great blog posts of late on learning F#. I can highly recommend

re: F# in 20 Minutes - Part I

secretGeek — Sun, 04 May 2008 01:45:46 GMT

Great introduction Chris!

@R, re: "require the full version of VS..?"

If you don't have Visual Studio, then you can run F Sharp Interactive as a console application (instead of a visual studio window).

The compiler and tools don't require Visual Studio -- they just offer Visual Studio integration as a "value add" feature.

Visual Studio express prohibits add-ons, so that kind of kills its ability to benefit from FSI. (Don Syme wrote answered a question about that here: http://cs.hubfs.net/forums/thread/3434.aspx )

re: F# in 20 Minutes - Part I

Peter Gum — Sun, 04 May 2008 12:50:52 GMT

Terrific! Just the right level of explanation that I badly need. Keep it up.

re: F# in 20 Minutes - Part I

Jan Midtgaard — Sun, 04 May 2008 13:33:32 GMT

"Passing around functions as values is known as 'first order functions' and is a hallmark of functional programming."

No, it's called 'first-class functions' (because functions are first-class citizens, like, e.g., numbers) or 'higher-order functions' -- not 'first-order functions'. Languages with 'first-order functions' don't pass them around.

re: F# in 20 Minutes - Part I

James Hurst — Mon, 05 May 2008 01:57:32 GMT

I appreciate your article, and just want to point out a couple of items that could perhaps be improved. For beginner's tutorials, little syntax errors can be confusing.

1. Under "printfn "N^2 = %A" squares", do you mean to say "printf is a simpler.." or "printfn is a simpler"? These look like two distinct things to me (I'm a noob obviously).

2. Under "let numbers = [1..10]", how're we supposed to type these lines into the FSI window? Exactly as is? Nothing happens. Do you intend for us to use the ";;" syntax to cause the line to execute? That yields an error. You might want to be a bit more explicit yet at this point to avoid frustrating your readers.

3. Under the picture of the FSI window, where you instruct your user to type "[1 .. 2 .. 10]" - an explanation of what the heck that IS, would be helpful here!

4. I'm guessing that "FSI" means "F Sharp Interactive" - but perhaps you could say that explicitly?

5. Where you give instructions on how to use the Add-in Manager to add the FSI window, you could stand to be a bit more explicit (IMO). You leave us with that Add-in Manager still open! What about those other two checkboxes - "Startup" and "Command Line"? Do you intend for us to click "OK" here before continuing on to your next step?

6. You skip a few beats in instructing your readers in creating their first F# project. "Add a new F# Source File"? Don't you need to let them know that they first need to create a project? And what about all that code and comments that come pre-loaded within that new F# source file I created? What's up with that? Do we need to delete it all first?

re: F# in 20 Minutes - Part I

Jonathan Pryor — Mon, 05 May 2008 23:27:43 GMT

Minor quibble:

> Now notice what we just did - we passed a function as a parameter to another function. You simply can't do that in C#. (At least not easily.)

For C# we can make the distinction between definition and use. Definition is more verbose, due to the need for types, but it's not impossible either -- you simply use a Delegate type:

class List {

public static TOutput[] map<TInput, TOutput> (Func<Tnput, TOutput> f, TInput[] input) {...}

}

Usage would actually be similarly verbose to F#:

var squares = List.map (square, numbers);

List.map (x => x % 2 == 0, new int[]{1, /* ... */ 10});

The overhead here is the array creation, not the lambda expression.

(And in this case, using System.Array.Convert() would make more sense than defining a new List.map method.)

re: F# in 20 Minutes - Part I

ChrSmith — Tue, 06 May 2008 05:54:07 GMT

Thanks for the feedback, I'll update this post to (hopefully) clarify things more.

And if I need to remember to write out:

public static TOutput[] map<TInput, TOutput> (Func<Tnput, TOutput> f, TInput[] input) {...}

every time I pass declare method which takes a function as a parameter, I consider that hard to do. Even if the call-site code is clean ;)

Thanks,

-Chris

re: F# in 20 Minutes - Part I

IanS — Tue, 06 May 2008 15:54:05 GMT

@R, re: "require the full version of VS..?"

You can use F# with the free Visual Studio 2008 Shell. See the following for more info:

http://11011.net/archives/000721.html

http://fsharpnews.blogspot.com/2007/12/f-in-visual-studio-for-free.html

re: F# in 20 Minutes - Part I

Eric Eilebrecht — Tue, 06 May 2008 17:31:59 GMT

Great article; looking forward to the rest of the series. I've been meaning to look at F#, and this looks like a great quick intro.

On the comparison to C#, I'd like to echo the previous comment that it's actually quite easy to pass a function to a function. And, in fact, the System.Linq namespace already defines an equivalent of "map." Your full example looks something like this in C#:

using System;

using System.Linq;

class Program

{

static void Main()

{

Func<int, int> square = x => x * x;

int[] numbers = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 };

var squares = numbers.Select(square);

Console.WriteLine("N^2 = {0}", squares.Aggregate("", (s, i) => s + " " + i));

Console.ReadKey(true);

}

Note that the hard part is declaring the sequence (which of course would be easier if there was an "Enumerable.Sequence" method or somesuch), and formatting the output.

Another approach, more in the spirit of the latest wave of C# features, would be:

static IEnumerable<int> Sequence(int start, int end)

{

for (int i = start; i < end; i++)

yield return i;

}

static void Main()

{

var squares = from x in Sequence(1, 10) select x * x;

Console.WriteLine("N^2 = {0}", squares.Aggregate("", (s, i) => s + " " + i));

Console.ReadKey(true);

}

And it's not true that declaring a function that takes a function arg is necessarily as verbose as declaring "map" is; that example is complicated by the fact that "map" has to be generic. I wonder what the declaration of "map" looks like in F#? Don't you have to do something special to keep the compiler from assuming all the types are "int," like it did for your "square" function?

Love the interactive evaluation thing. That might make me pick up F# all by itself.

Thanks again for the article!

F# in 20 Minutes – Part II

Chris Smith's completely unique view — Sat, 10 May 2008 01:21:28 GMT

Now that we have covered the basics, in minutes 8 - 14 we will cover the foundational concepts and types

F# 20分钟快速上手（一）

Anders Cui — Sun, 03 Aug 2008 09:53:46 GMT

从Allen Lee的《从C# 3.0到F#》一文开始，感觉园子里F#正在升温。Chris Smith写了一个F#的小系列，这里翻译出来与大家分享。在本文从零开始编写我们的第一个F#程序。

re: F# in 20 Minutes - Part I

Ron Mitchell — Mon, 08 Sep 2008 02:18:09 GMT

Tried to install the latest F# CTP where the platform is "Vista Ultimate" and VS 2008...hmmm - install dies saying that it not an appropriate MSI file???

Strange..

Ron

re: F# in 20 Minutes - Part I

ChrSmith — Mon, 08 Sep 2008 02:24:38 GMT

Hey Ron, could you be more specific. What was the exact error message you recieved? Please send it to fsbugs@microsoft.com.

Thanks!

re: F# in 20 Minutes - Part I

Matthew Doig — Tue, 09 Sep 2008 06:35:28 GMT

Good little intro to f#. I used this for my initial lesson when learning f#. If you're a VB programmer you may want to take a look at the series of blogs i've just started.

http://blog.matthewdoig.com/?p=25

re: F# in 20 Minutes - Part I - Don't do it

Alex — Tue, 09 Sep 2008 12:04:42 GMT

Unfortunately - as it seems, F# developers decided to use the way of the such aincient programming languages as Fortran - minimal type control, maximal possibility to make errors. If I am right, it's a mistake. The banner of this way could be expressed (using gross exaggeration in order not to say too much words) as follows - we will write the code as soon as possible (<all the rest will not be said to customers> but then we will look for the mistakes all the life remaining). Making mistakes is usual way for the middle-level programmer, F# will help him to do that?

re: F# in 20 Minutes - Part I

Matt Galligan — Fri, 12 Sep 2008 01:21:24 GMT

With F# being presumably distilled into the same CLI, does using F# confer any speed advantages to the same logically-identical function in C#/etc?

Basically, I'm wondering for intensive math calculations (such as stuff for http://projectueler.net), will F# confer a speed of execution advantage?

re: F# in 20 Minutes - Part I

Matt — Fri, 12 Sep 2008 01:23:00 GMT

whoops. I meant http://projecteuler.net

re: F# in 20 Minutes - Part I

Chris Smith — Fri, 12 Sep 2008 01:27:43 GMT

Since C# and F# both compile down to IL and get executed with the same JIT engine, there is no distinct speed 'advantage' for using one language over the other. F# should have the same performance profile as C#.

Any significant difference in benchmarks is usually due to different algorithms being used.

re: F# in 20 Minutes - Part I

maveric — Fri, 12 Sep 2008 13:50:08 GMT

Yes this is quite good . And what more advantanges with this lang.

re: F# in 20 Minutes - Part I

Rupert — Wed, 17 Sep 2008 20:25:21 GMT

Hi I'd like to know (as mentioned by somebody above) what List.map square [1 .. 2 .. 10];;

I can see from the result that it's outputting the squares of 1, 3, 5, 7 and 9 but I'm at a bit of a loss as to why!

Thanks, great article by the way.

re: F# in 20 Minutes - Part I

Rajesh Charagandla — Thu, 18 Sep 2008 08:58:50 GMT

Excellent.. I am very much impressed the way you explained about the functional programming language F#. I also keep the specification document for F# when i am reading this post. Twice i turned to spec document for :: and @ operators meaning.

Enjoyed writing F# code ... i am really become the fan of F# language after reading your blog. Thank you so much. Eagerly Waiting for second part. :)

re: F# in 20 Minutes - Part I

int19h — Thu, 18 Sep 2008 17:28:57 GMT

There may be some performance advantage in the fact that F# does tail call optimization where it can (C#, in contrast, never does it).

re: F# in 20 Minutes - Part I

Koistya `Navin — Fri, 19 Sep 2008 00:58:25 GMT

Great introduction to F# for beginners, thanks.

re: F# in 20 Minutes - Part I

anon — Tue, 23 Sep 2008 11:21:57 GMT

@Rupert

List.map square [1 .. 2 .. 10];;

2 must be increment by: 1+2=3+2=5+2=7+2=9

The Weekly Source Code 34 - The Rise of F#

ASPInsiders — Thu, 25 Sep 2008 12:14:34 GMT

First, let me remind you that in my new ongoing quest to read source code to be a better developer ,

The Weekly Source Code 34 - The Rise of F#

Readed By Wrocław NUG members — Thu, 25 Sep 2008 12:45:09 GMT

First, let me remind you that in my new ongoing quest to read source code to be a better developer ,

re: F# in 20 Minutes - Part I

echostorm — Thu, 25 Sep 2008 19:03:34 GMT

It can be even easier in C# than Eric's.

MessageBox.Show(Enumerable.Range(1, 10).Select(a => a * a).Aggregate("", (b, c) => b + " " + c));

re: F# in 20 Minutes - Part I

cbmeeks — Wed, 01 Oct 2008 05:27:04 GMT

Very nice tutorial. I hope to learn some F# soon and this really helped.

re: F# in 20 Minutes - Part I

Jason — Wed, 01 Oct 2008 18:19:28 GMT

Just learn Python and use IronPython and be done with it. Too many languages create vertical markets. VB dropped "let" 20 years ago, why bring it back?

I was hoping to see something tuly nice, and readable:

numbers = range(1,10)

print "N^2="+str([x*x for x in numbers])

Oh wait, that is python. If you feel the need to use a map construct:

def square(n):

return n*n

print "N^2="+str(map(square, numbers))

What is more readable?

re: F# in 20 Minutes - Part I

Ian Marteens — Thu, 02 Oct 2008 22:42:43 GMT

F# is more readable and powerful, of course.

There's no obligation about learning functional programming, but it's an important part from Computer Sciences, and, if you want to compare two different languages as Python and F#, you should know the basic concepts behind FPLs for making a sensible evaluation.

re: F# in 20 Minutes - Part I

Srinivasan S Saripalli — Tue, 07 Oct 2008 05:30:53 GMT

I have F#ver1.9.6.2 installed on my machine.Since I dont have VS2008 installed,I run F#I Console.I have executed the below two lines of code

let vowels = ['e','i','o','u'];;

let attacha = 'a' :: vowels;;

however i get the below errors

let attacha = 'a' :: vowels;;

---------------------^^^^^^^

stdin(15,22): error FS0001: Type mismatch. Expecting a char list but given a ( char * char * char * char) list

the type char does not match the type 'char * char * char * char'.

Can anybody please explain why I get the above error.

I certainly feel that to learn F# C# is not a criteria,because I dont know C# but I am primarily a VC++ guy.

Regards,

Srinivasan S Saripalli

Bangalore,India

re: F# in 20 Minutes - Part I

ChrSmith — Tue, 07 Oct 2008 19:11:07 GMT

The problem is the way you declared the list:

['e';'i';'o';'u']

is a list of characters (semicolon delimited)

['e','i','o','u']

is a single-element list, which is a tuple of 4 characters. (Duples are comma delimited.)

Make sense?

re: F# in 20 Minutes - Part I

anon — Thu, 09 Oct 2008 18:24:09 GMT

The reason it didn't work Srinivasan is because you separated the list elements with commas instead of semicolons. Comma separation creates a tuple, or compound type, so what you actually made there was a list with one element, a tuple of four chars, instead of a list with four char elements.

This article should explain it better than I can:

http://diditwith.net/2008/01/18/WhyILoveFTuples.aspx

re: F# in 20 Minutes - Part I

anon — Thu, 09 Oct 2008 18:50:03 GMT

Whoops, sorry Chris, for some reason your answer wasn't showing up on my browser until after I posted mine.

F# 20分钟快速上手（一）

心界使者 — Thu, 16 Oct 2008 17:35:37 GMT

原文链接：http://www.cnblogs.com/anderslly/archive/2008/08/03/fs-in-20-minutes.html

Chris Smith's completely unique view : F# in 20 Minutes - Part I

DotNetKicks.com — Sat, 18 Oct 2008 05:50:38 GMT

You've been kicked (a good thing) - Trackback from DotNetKicks.com

re: F# in 20 Minutes - Part I

Frigyes Reisch — Mon, 24 Nov 2008 21:14:46 GMT

Hello,

The programs below are well received. However an expensive source code (MATLAB) is used. Several potential users would like to use an open source code like F#. Is it suitabel to use F# to solve differential equations?

Research reactor

http://www.euronuclear.org/e-news/e-news-20/neutron-flux.htm

Chernobyl avalanche

http://www.euronuclear.org/e-news/e-news-20/chernobyl-avalanche.htm

re: F# in 20 Minutes - Part I

ChrSmith — Tue, 25 Nov 2008 22:14:34 GMT

Re: "Is it suitabel to use F# to solve differential equations?"

Certainly. F# seems to be very productive in highly algorithmic domains - e.g. the math and number crunching. As for the MatLab piece, I assume it is just a matter of finding the right math libraries and graphing libraries.

Unfortunately I don't have any reccomendations for mathematical .NET libraries.

re: F# differential equations

Frigyes Reisch — Wed, 26 Nov 2008 11:59:32 GMT

Thanks for Your answer. Could You recomend somebody I could turn to for answer?

re: F# in 20 Minutes - Part I

Tim — Wed, 26 Nov 2008 21:24:19 GMT

I don't understand Alex's complaint about minimal type checking. As far as I can see, F# has quite strict type checking at compile time (much stricter than the C family of languages, for example). Maybe it's not obvious from the tutorial, because you don't have to /tell/ F# the type of an identifier if F# can unambiguously infer it.

re: F# in 20 Minutes - Part I

Tim — Wed, 26 Nov 2008 21:35:32 GMT

Jason, you do realise that "let" in F# is completely unrelated to "let" in BASIC, don't you? The F# equivalent to BASIC's

LET A = 5

a <- 5

No sign of a "let". In F#, "let" means something quite different, that I don't think has a direct equivalent in non-functional languages (I suppose it has /some/ similarity to "&" in C++, because it's defining a sort of alias).

re: F# in 20 Minutes - Part I

Dustin Boyd — Tue, 02 Dec 2008 22:21:42 GMT

For those that still don't understand the [1 .. 2 .. 10] bit, it reminds me a bit of the FOR statement when used with the /L option:

FOR /L (start, step, end) DO action

Although I don't have much knowledge of the .NET library, it is reasonable to infer that "List.map square [1 .. 2 .. 10]" creates a list by mapping its elements to the result of calling square with arguments 1, 3, 5, 7 and 9 (because the step value is 2, and the constraints specified 1 to 10 inclusive). The end result is that we end up with with the squares of the numbers above. It also reminds me of list comprehensions in Python:

>>> [square(i) for i in range(1, 11, 2)]

[1, 9, 25, 49, 81]

Of course, in Python, the range function excludes the upper limit, so 11 needs to be used rather than 10.

I love these types of languages. They can get so much done with such little code!

re: F# in 20 Minutes - Part I

Selami Ozlu selamiozlu — Tue, 09 Dec 2008 19:40:22 GMT

F# you can even write XBox games using XNA.

I tried a small demo, it works very nicely.

Selami Ozlu

re: F# in 20 Minutes - Part I

wised — Fri, 09 Jan 2009 13:37:19 GMT

Stupid Microsoft, this is just another VBscript.

Stupid VBstscript, certainly stuoid F# again.

re: F# in 20 Minutes - Part I

Monty — Fri, 16 Jan 2009 19:36:22 GMT

Just Learn Python instead. Get all the Dynamic stuff without the .net Framework. Why would you try to do dynamic stuff in the CLR?

I tried doing a recursive lambda once in c#. nightmare and it doesn't work.

The new stuff on DLR looks interesting though, although I am thinking of jumping ship to Python as it does all that I want and then some. Who needs stupid acronyms like LINQ when you got list comprehensions.

If you really want the functional style then go Haskell, the only one that doesn't cause a stack overflow as far as I can tell.

re: F# in 20 Minutes - Part I

Craig — Wed, 21 Jan 2009 01:17:13 GMT

It would be nice to see a better example of the advantages of F# over C#. As a C# developer, I agree with some others that it just looks like some awkward syntax and easy way to code in more errors so far.

Can you give an example of something F# can do that C# can't?

re: F# in 20 Minutes - Part I

Marcin Kasperski — Thu, 29 Jan 2009 03:26:09 GMT

Why Caml (instead of - say - Haskell, Standard ML, or even Erlang)?

Just curious - I don't consider this decision to be bad, after all...

re: F# in 20 Minutes - Part I

renu — Sat, 31 Jan 2009 21:36:11 GMT

@MarcinK - Haskell is pure FP, no OO construct. Ocaml has the OO stuff that F# also supports.

F is for… F#

Jim O'Neil's Blog — Mon, 09 Feb 2009 08:31:37 GMT

F# is the new kid on the block in terms of .NET languages.  Currently in CTP form , it will be one

re: F# in 20 Minutes - Part I

John — Thu, 26 Feb 2009 12:58:28 GMT

F# has its roots in a programming language called OCaml, C# has its roots in a programming language called Java, Garbage Collection has its roots in Lisp, etc.

re: F# in 20 Minutes - Part I

John II — Thu, 26 Feb 2009 13:03:08 GMT

So why all this efforts to re-invent the wheel?

re: F# in 20 Minutes - Part I

Surendra Sharma — Wed, 04 Mar 2009 08:04:15 GMT

Good Article. But why I used F#, if I already have C#? If possible make a comparison among F#, C# and Java. Only for mathematical calculation, I will never usr F#.

re: F# in 20 Minutes - Part I

Terry — Thu, 05 Mar 2009 18:45:52 GMT

Why would you use F# ........ ?

Anyone who thinks that this is a) like Fortran (trust me it isnt) or b) just for calculation so wont use it either hasnt gone far enough yet or has their head wedged where the sun dont shine :-)

May I point you to the excellent PDC demo (about 78 minutes) go to the F# research page click on getting started and about halfway down you will see a thumbnail marked demo - run that demo and be amazed.

re: F# in 20 Minutes - Part I

balboos — Fri, 13 Mar 2009 16:59:12 GMT

It's nice to have had someone write a tutorial, but (IMHO) this one just leaves out too many concepts for a part-one. It's borders on a turn-off, for what is basically a lot of clutter.

It's not that I'm particularly dim-witted, having followed a language use path of FORTRAN->assembly->C->C++ , and C# when necessary (because MS didn't see fit to include the functionality in C++). The [1..2..10] was (to me), an obvious syntax, given the results.

I just think that a tutorial should <i>teach more and extol less</i>. Again, this is an IMHO.

re: F# in 20 Minutes - Part I

Rod Wue — Tue, 24 Mar 2009 05:12:38 GMT

I read these comments and I wonder about if some of you even read anything about F# before you commented. I'm certainly not a F# evangelist as I only really started looking at it more closely in the past few weeks.

It's interesting how many of the commenters don't understand "side effects" They keep asking what good is F#? WHy would I use this.. Etc. Etc. Perhaps you should watch the Anders interview about C# 4.0 they go off on a tangent that might help you understand.

re: F# in 20 Minutes - Part I

Abhijeet Patel — Sun, 19 Apr 2009 19:57:57 GMT

@Eric Eilebrecht

"Note that the hard part is declaring the sequence (which of course would be easier if there was an "Enumerable.Sequence" method or somesuch)"

You could use Enumerable.Range

re: F# in 20 Minutes - Part I

BAIJUMAX — Fri, 24 Apr 2009 11:14:47 GMT

Nice Intro .

I'm new to F#, enjoyed reading this article.

F# looks similar to Matlab programming (.M code) .

Talk: If you think F# has too limited applicability

Sergio Pereira — Fri, 01 May 2009 07:19:56 GMT

Next Chicago ALT.NET meeting will bring a practical look at F#, showing that it does not need to be seen

re: F# in 20 Minutes - Part I

Bui Viet Duc — Thu, 07 May 2009 05:04:18 GMT

I did not try F#

But based on this introduction, I think F# borrows the idea of Prolog

re: F# in 20 Minutes - Part I

person-b — Tue, 26 May 2009 19:55:24 GMT

@Bui Viet Duc:

F# is part of a family of so-called "functional languages". Examples of these languages are Haskell, LISP, Scheme. They also borrow some concepts from languages such as Prolog. They don't, however, copy their core ideals and syntax from Prolog.

INTRODUCING F#

Web开发技术 — Wed, 27 May 2009 10:33:23 GMT

第一篇，从零开始编写我们的第一个F#程序。什么是F#，我为何要学它？ F#是一种.NET平台上的函数式编程语言。就像C#和VB.NET，F#可以利用.NET的核心类库，如 WPF ， WCF ，

re: F# in 20 Minutes - Part I

raghu — Tue, 02 Jun 2009 20:09:52 GMT

Great article. Will definitely be reading up more on F#.

Just wish that MS would offer lisp as well.

re: F# in 20 Minutes - Part I

ming — Fri, 12 Jun 2009 00:21:15 GMT

I suspect for a beginner this might not be the most friendy introduction...

re: F# in 20 Minutes - Part I

Jeba Singh — Fri, 12 Jun 2009 07:43:34 GMT

It's a great article to know about F#.