Back to Questions
1.3k
views5
comments
new kid on the block, first F# impressions
f#functional

Looked some time for this forum, as i expected it on the msdn-site, where also the main site points me to. but it doesn't seem to be vacuumed up onto that collecting all forum.

I am using the VS2008 ctp, so things might refer to the.

I noticed some on the fly compilation, simular feature on C# I had to deactivate, since it intefered far too much with my source-editing. Also it looks to me the 2nd comming of intellisense. The latter seems to be lacking a bit with F#, I reckon due to the version I'm using.

I trust though on the fly F# compilation can be deactivated too, currently I don't mind much since my programs are still small and expirimental.

This forums postings helped me to get me to define an array2d though confusing to see listing.

arr.(0).(0) (or simular) and stuff when arr.[i,j] is the F# array indexing, the jagged array which is favored in C# I'll look into shortly.

As said I'm ne to F#, and try to cut my way into it rapidly, as it looks that functional programming is what my hobby-project needs at the moment.

postings on this site where already helpfull, and I reckon some Q and A might get me through the problems I can't resolve on my own, as also the F#-help system ain't up to specs yet. (understandable given VS2008 ctp)

BTW, the postings prints awkward, which seems to be a problem with these kinds of forums I reckon.

further I noticed type Complex based on floats, though I don't reckon doing something with it shortly, I could imagin float not precise enough. In the absence of BigReal (I reckon must have some cutoff, as the not Rational numbers suffer from infinit digits(hi hi)) float is understandable. I can imagin Complex<BigRational> (aka Complex<BigNum>) might give possibilities beyond the current Complx<float>

.

dear friends

I thought I got the hang of things and reckoned I could iterate a few things shortly however while

1
2
3
4
5
6
7
8
9
10
11
12
13

let in3cube x = x / 16 <> 3 && x / 4 % 4 <> 3 && x % 4 <> 3
let rec check i j =
if j < i then 
if (knightpath.[j] = knightpath.) then
knightpath. <- -1
else
check i (j+1)
for i in 1..max-1 do
if (knightpath.[i - 1] = -1) then
knightpath. <- -1;
else
knightpath. <- knightreaches.[knightpath.[i - 1]].[0];
check i 0

gives me the expected result:

1
2
3
4
5
6
7

[|0; 6; 15; 43; 45; 63; 57; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1;

-1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1;

-1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1;

-1; -1; -1; -1|]

trying to iterate through more possible values stored in th array knightreaches

I got the following error messages:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

let rec knightpaths ii kk = [
let knightpath = Array.zeroCreate 10;// max;
knightpath.[0] <- 0
let rec check i j =
if j < i then 
if (knightpath.[j] = knightpath.) then
knightpath. <- -1
else
check i (j+1)
if (knightpath.[ii - 1] = -1) then
knightpath.[ii] <- -1;
else
knightpath.[ii] <- kk;
check ii 0
let in3cube x = x / 16 <> 3 && x / 4 % 4 <> 3 && x % 4 <> 3
if in3cube knightpath.[ii] then
yield knightpaths (ii+1) knightreaches.[knightpath.].[0]
// Error 1 Type mismatch. Expecting a 'a but given a 'a list. The resulting type would be infinite when unifying ''a' and ''a list' .... (omitted)
else
yield knightpath
]

or without recursive 'rec':

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

let knightpaths ii kk = [
let knightpath = Array.zeroCreate 10;// max;
knightpath.[0] <- 0
let rec check i j =
if j < i then 
if (knightpath.[j] = knightpath.) then
knightpath. <- -1
else
check i (j+1)
if (knightpath.[ii - 1] = -1) then
knightpath.[ii] <- -1;
else
knightpath.[ii] <- kk;
check ii 0
let in3cube x = x / 16 <> 3 && x / 4 % 4 <> 3 && x % 4 <> 3
if in3cube knightpath.[ii] then
yield knightpaths (ii+1) knightreaches.[knightpath.].[0]
//Error 1 The value or constructor 'knightpaths' is not defined. ....(omitted)
else
yield knightpath
]

so I'm back to:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

let knightpaths ii kk = [
let knightpath = Array.zeroCreate 10;// max;
knightpath.[0] <- 0
let rec check i j =
if j < i then 
if (knightpath.[j] = knightpath.) then
knightpath. <- -1
else
check i (j+1)
if (knightpath.[ii - 1] = -1) then
knightpath.[ii] <- -1;
else
knightpath.[ii] <- kk;
check ii 0
//let in3cube x = x / 16 <> 3 && x / 4 % 4 <> 3 && x % 4 <> 3
//if in3cube knightpath.[ii] then
// yield knightpaths (ii+1) knightreaches.[knightpath.].[0]
//else
yield knightpath
]

let knigtjumpies = [
for kk in knightreaches.[0] do
yield knightpaths 1 kk
]

printfn "" 
for kj in knigtjumpies do
printfn "%A" kj

which gives me only the first possibilities:

1
2
3
4
5
6
7
8
9
10
11

[[|0; 6; 0; 0; 0; 0; 0; 0; 0; 0|]]

[[|0; 9; 0; 0; 0; 0; 0; 0; 0; 0|]]

[[|0; 18; 0; 0; 0; 0; 0; 0; 0; 0|]]

[[|0; 24; 0; 0; 0; 0; 0; 0; 0; 0|]]

[[|0; 36; 0; 0; 0; 0; 0; 0; 0; 0|]]

[[|0; 33; 0; 0; 0; 0; 0; 0; 0; 0|]]

correctly

currently just a bit stumped perhaps due to my current state of mind and time.

just after the result shown near the top I reckoned I would get simular lines by now showing me all the possibilities uptill the condition (knightjump path getting outside the 3 by 3 by 3 cube)

sorry for loosing the indentation of the code for some reason "Formatted" returned to "paragraph" and now showing "normal"

i hope it is still understandable.

darn I see [ i ] stands for a light here, having to figure out how to paste code on every new forum?!

just a note I reckon to move the array knightpath outside the knightpaths function, and I need to formulate a copy constructor to yield the result.

other attempts I just got the last possibility mentioned several (6) times.

Further I hope to find when moving to 2010 (when released) to be able to set breakpoint on a code of line to enable debugging.

now it holds on an entire function and skips past the function in stead of allowing me to debug the function.

By on 6/10/2009 10:41 AM ()

If you want your code to appear correctly, use these tags:

{code language="F#"}
<ForumPost.Replace("{","[").Replace("}","]")>
{/code}

It's hard to see exactly what your code does, but I noticed something at the end of your knightpaths function.

The 

1

if in3cube knightpath.[ii]...

statement yields two different things at each branch. The first branch yields knightpaths and the second branch yields knightpath. The former is a list of arrays, but the second is an array. Outputting different kinds of data in different if/then/else branches is a no-no. The same goes for match statements, of course.

Instead of having the first branch spit out the entire list of arrays as a single entry in your list, try using 

1

yield! knightpaths (ii+1) ...

instead of 

1

yield knightpaths (ii+1) ...

to have the first branch spit out lists one-by-one.

P.S. Tip: try not to use library function names when naming your values. Naming a value max may lead to weird bugs later.

By on 6/11/2009 1:15 AM ()

{code language="F#"}
<ForumPost.Replace("{","[").Replace("}","]")>
{/code}

Sorry about this. I was being too cryptic for my own good. I meant this:

{code language="F#"}
.. insert code here ..
{/code}

replace { } by [ ]

By on 6/11/2009 4:11 AM ()

edited my former post thus, but still the font tags I didn't enter.

anyway the code now looks ok. I tried yesterday also code tags in square brackets but hadn't an idea of the language argument.

from the F#-specs I got the difference between yield! (for computation) and yield (for result)

thnx.

[;)]STRANGE code tag in square brackets generates font tags in the following code ???[:O]

figured out the following:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

let knecht ii kk jj: array<int> =
   let knightpath = Array.zeroCreate 10
   let rec knechtp ii kk =
      knightpath.[ii] <- kk
      check ii 0
      if (ii < 9 && knightpath.[ii] <> -1 && in3cube knightpath.[ii]) then
         knechtp (ii+1) knightreaches.[kk].[jj]
      else
         knightpath.[ii + 1] <- -1
   knightpath.[ii] <- 0;
   knechtp (ii+1) kk
   let mutable len = 1
   while len < 10 && knightpath.[len] <> -1 do
      len <- len + 1; 
   let knightres = Array.zeroCreate len
   array.Copy(knightpath,knightres,len)
   knightres

let knigtjumpies = [
   for kk in knightreaches.[0] do
      yield knecht 0 kk 0
      yield knecht 0 kk 1
   ]

to give me the knightjumpies:

1
2
3
4
5
6
7
8
9
10
11
12

[|0; 6; 15|]
[|0; 6; 30|]
[|0; 9; 15|]
[|0; 9; 27|]
[|0; 18; 27|]
[|0; 18; 42; 35|]
[|0; 24; 30|]
[|0; 24; 42; 35|]
[|0; 36; 42; 44|]
[|0; 36; 45|]
[|0; 33; 39|]
[|0; 33; 42; 35|]

I need but to figure out how to iterate through all possible values.

A bit strange though as I tried to change the constant '10' to the length of the array

i found knightpath.GetLength(int) I'll have no clue what the int parameter is supposed

to do there! Expected GetLength to give me 10 back! ???

By on 6/11/2009 4:58 AM ()

thanks cfern

What I intend to do is to find KnightJump Path's through an order 4 cube within the order 3 subcube, test the found values whether a semi-magic cube can be obtained, etc. which goes a bit too far here. a knightjump path looks like 

1
2
3
4
5
6
7

[|0; 6; 15; 43; 45; 63; 57; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1;

-1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1;

-1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1; -1;

-1; -1; -1; -1|]

where positions are coded as p = z*16+y*4+x for convenience.

Thanks for the rule of thumb I could have guessed that is was a no-no to output different types. I'm new to the language (in fact new to functional programming, save some C# linq statments) so I'll need to dig into the difference between yield and yield!. Also familiarize me with library function names like max.

This is just a puny test project, as functional programming looks still like a good idea for the problem I encountered. And the data to knightjump I obtained in but a couple of days.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50

let TestingArray() =
    let knightjumps = [
      yield (0,1,2)
      yield (0,2,1)
      yield (1,0,2)
      yield (1,2,0)
      yield (2,1,0)
      yield (2,0,1)
      yield (0,-1,2)
      yield (0,2,-1)
      yield (-1,0,2)
      yield (-1,2,0)
      yield (2,-1,0)
      yield (2,0,-1)
      yield (0,1,-2)
      yield (0,-2,1)
      yield (1,0,-2)
      yield (1,-2,0)
      yield (-2,1,0)
      yield (-2,0,1)
      yield (0,-1,-2)
      yield (0,-2,-1)
      yield (-1,0,-2)
      yield (-1,-2,0)
      yield (-2,-1,0)
      yield (-2,0,-1)
      ]
   let add (x,y,z)
 (vx,vy,vz) = (x+vx,y+vy,z+vz)
   let order = 4
   let max = order * order * order
   let square = Array3D.create order order order 0;
   let nums = [0 .. order - 1]
   for i in nums do
      for j in nums do
         for k in nums do
            square.[i,j,k] <- i * order * order + j * order + k;
   let ok (x,y,z) = x >= 0 && x < order && y >= 0 && y < order && z >= 0 && z < order
   let posnum (x,y,z) = x * order * order + y * order + z
   let knightreach (x,y,z) = [
         for vec in knightjumps do
         let pos = add (x,y,z) vec
            if (ok pos) then
               yield posnum pos
      ]
   let knightreaches = Array.zeroCreate max;
   for i in 0..max-1 do
      knightreaches.[ i ] <- knightreach (i / (order * order) , i / order % order, i % order)
   for i in 0..max-1 do
      printfn "%02d %A" i knightreaches.[ i ]

without 'posnum' positions are ouputted, which would be ok also, and probably with F# matching could be used just as simple as the coded positions.

[:O]hope it is formatted, yet pasted it looks strange (manually corrected)

[:@]why does it looks like I am changing font!

By on 6/11/2009 3:48 AM ()

    Topic tags

    IntelliFactory Offices Copyright (c) 2011-2012 IntelliFactory. All rights reserved.
    Home | Products | Consulting | Trainings | Blogs | Jobs | Contact Us | Terms of Use | Privacy Policy | Cookie Policy     		Built with WebSharper