TweetFollow Us on Twitter

Sep 98 Prog Challenge

Volume Number: 14 (1998)
Issue Number: 9
Column Tag: Programmer's Challenge

Sep 98 Programmer's Challenge

by Bob Boonstra, Westford, MA

Big Baby

Fifty years ago this past June, the Manchester Mark I prototype computer, also known as "Baby", became operational. Baby was the first computer to store a program electronically, and was also the first computer to store instructions and data in the same memory. Because vacuum tube technology was too immature to store memory reliably, Baby was designed to test memory based on a cathode ray tube. Not much memory, mind you. Baby boasted a full 1K bits of memory, organized as 32 words (or lines) of 32 bits each.

In celebration of the birth of the first stored program computer on June 21, 1948, the Department of Computer Science at the University of Manchester recently reconstructed Baby and ran a programming contest to write the most imaginative program for Baby. Inspired by that contest, your Challenge is to write an assembler and an emulator for an extended ("Big") version of Baby. The prototype for the code you should write is:

#if defined(__cplusplus)
pragma extern "C" {
#endif

#define kMaxInstructions 32

typedef UInt32 CRT_memory[kMaxInstructions];

pascal void AssembleBabyProgram(
   char *program,
   CRT_memory memory,
   UInt32 address_bits
);

pascal void ExecuteBabyProgram(
   CRT_memory memory,
   UInt32 address_bits
);

#if defined(__cplusplus)
}
#endif

Baby has a single general-purpose register, called the Accumulator. The program counter is called the Control Instruction, or CI. The CI is incremented just before the next instruction is fetched, which means that a jump instruction, for example, is coded with a value one less than the actual target address. Baby also has a red light that indicates the program has halted. One interesting thing about Baby is that it lacks an addition instruction - addition is done by subtraction.

Baby's instruction repertoire is listed below. The function bits (or opcode) associated with each instruction is listed in parentheses after the mnemonic.

STO (110)
Store the contents of the Accumulator in the store line.
SUB (001 or 101)
Subtract the contents of the store line from the Accumulator. There is no ADD instruction; addition is done indirectly by combining the SUB and the LDN instruction.
LDN (010)
Copy the contents of the store line, negated, to the accumulator.
JMP (000)
Copy the contents of the store line to the CI (so the store line holds the number of the line one before we want to jump to). In modern terms, this an indirect jump, which uses up an extra store line compared to a direct jump.
JRP (100)
Add the contents of the store line to the CI. This looks forward to larger machines, where it would be important to be able to load the same code in different places, and hence would need relative jumps.
CMP (011)
Skip the next instruction if the contents of the Accumulator are negative, i.e. a conditional branch.
STOP (111)
Stop the machine and turn the red light on
NUM (N/A)
An assembler mnemonic to initialize a store line to a data value.

For example, the following program computes the greatest common divisor of the number in locations 30 and 31:

22
0000 NUM 0
0001 LDN 30
0002 STO 29
0003 LDN 31
0004 STO 31
0005 LDN 31
0006 STO 30
0007 LDN 29
0008 SUB 30
0009 CMP
0010 JRP 27
0011 SUB 31
0012 STO 31
0013 SUB 28
0014 CMP
0015 JMP 00
0016 STP
0027 NUM -3
0028 NUM 2
0029 NUM 0
0030 NUM 3141593
0031 NUM 5214

Baby's instructions are assembled into a 32 bit word by placing the function code associated with the mnemonic into bits 13-15 (numbered with bit 0 as the most significant bit). In the original Baby, the store line associated with the instruction is placed in bits 0-4. Bits 5-12 and 16-31 are not used as part of the instruction, although they can be used as data. The program listed above assembles to the following:

22
0000:00000000000000000000000000000000
0001:01111000000000100000000000000000
0002:10111000000001100000000000000000
0003:11111000000000100000000000000000
0004:11111000000001100000000000000000
0005:11111000000000100000000000000000
0006:01111000000001100000000000000000
0007:10111000000000100000000000000000
0008:01111000000000010000000000000000
0009:00000000000000110000000000000000
0010:11011000000001000000000000000000
0011:11111000000000010000000000000000
0012:11111000000001100000000000000000
0013:00111000000000010000000000000000
0014:00000000000000110000000000000000
0015:00000000000000000000000000000000
0016:00000000000001110000000000000000
0027:10111111111111111111111111111111
0028:01000000000000000000000000000000
0029:00000000000000000000000000000000
0030:10011011111101111111010000000000
0031:01111010001010000000000000000000

Our contest will make one change to the original Baby: in our extended, Big Baby, machine, the store line is extended from 5 bits (0-4) to address_bits bits (0 - address_bits-1). This allows more than 32 words of memory and therefore larger programs.

Your AssembleBabyProgram routine should accept the mnemonic input listed above, pointed to by the program parameter, and assemble them into 32-bit Baby instructions in memory. Your ExecuteBabyProgram routine will be called to execute the program one or more times. Both of your routines will be provided an address_bits parameter that describes the size of memory. You will be asked to assemble more than one program, your assembled programs may be executed more than one time each, and you may be asked to execute a program that has been hand-assembled.

More information about the University of Manchester Baby programming contest can be found at http://www.cs.man.ac.uk/prog98/. Programming reference documentation for Baby can be found at http://www.cs.man.ac.uk/prog98/ssemref.html and at ftp://ftp.cs.man.ac.uk/pub/CCS-Archive/misc/progref1.doc.

The winner will be the solution that assembles and executes a set of test programs in the minimum amount of time.

This will be a native PowerPC Challenge, using the latest CodeWarrior environment. Solutions may be coded in C, C++, Pascal or, as is our tradition in the month of September, in assembly language. Thanks to Eric Shapiro for suggesting this Challenge.

Three Months Ago Winner

Congratulations to Tom Saxton for writing the most successful simulated gambler at the blackjack table of our June Programmer's Challenge Casino. Tom beat out four other entries and was one of only two entries to actually come out ahead at the blackjack table.

Tom precomputed the expected winnings for each situation and created tables with the action that led to the best result. He uses the Hi-Lo card counting method to determine whether the remaining cards contain a disproportionate number of high-valued cards, and then uses that estimate to adjust his wager. Tom's solution is also not too greedy; it contains heuristics to quit when it has won a reasonable amount or played long enough, ensuring that it has wagered enough credits to avoid the "freeloader" penalty imposed by the problem.

A few words about our other gamblers are in order. The second-place solution, by Kevin Hewitt, also used precomputed tables, but his were based only on the initial pair of cards dealt. Kevin also spent more time at the table, quitting only when winnings or losses exceeded a threshold. JG Heithcock's solution spent the least amount of time at the table. He quit soon after the minimum total bet criterion was met. Ken Slezak kept playing until he lost 75% of his bankroll (or quadrupled his money) and Randy Boring played until he ran out of money or, as it turned out, until the house threw him out of the casino. Both of those players left with not much more than the shirts on their backs.

Here are the statistics for the entries to the Blackjack Challenge. Each player played a series of five games where the house varied the number of decks of cards used. Players were given the same number of credits at the start of each game, totaling 21000 credits for all of the games. The table below lists the total number of credits wagered by the player, the number of credits left when the player decided to quit, the number of hands played, total execution time, and the overall player score. Also listed are the code and data sizes for the entries, along with the programming language used. As usual, the number in parentheses after the entrant's name is the total number of Challenge points earned in all Challenges to date prior to this one.

Name Credits Wagered Credits Left Hands Played Exec. Time Score Code Size Data Size Lang
Tom Saxton (19) 47451 25199 327 7169 25194 1496 1924 C
Kevin Hewitt 438700 23800 1833 37923 23766 996 2156 C
JG Heithcock (20) 22616 20484 769 17950 20470 1304 232 C
Ken Slezak (20) 91760 9140 1701 36911 9106 1240 172 C
Randy Boring (81) 437230 8670 15425 460099 8213 4920 353 C

Top Contestants

Here are the Top Contestants for the Programmer's Challenge, including everyone who has accumulated more than 20 points during the past two years. The numbers below include points awarded over the 24 most recent contests, including points earned by this month's entrants.

  1. Munter, Ernst 190
  2. Boring, Randy 76
  3. Cooper, Greg 54
  4. Mallett, Jeff 50
  5. Rieken, Willeke 47
  6. Nicolle, Ludovic 34
  7. Lewis, Peter 31
  8. Maurer, Sebastian 30
  9. Saxton, Tom 29
  10. Heithcock, JG 27
  11. Gregg, Xan 24
  12. Murphy, ACC 24
  13. Hart, Alan 21
  14. Antoniewicz, Andy 20
  15. Day, Mark 20
  16. Higgins, Charles 20
  17. Hostetter, Mat 20
  18. Studer, Thomas 20

There are three ways to earn points: (1) scoring in the top 5 of any Challenge, (2) being the first person to find a bug in a published winning solution or, (3) being the first person to suggest a Challenge that I use. The points you can win are:

  • 1st place 20 points
  • 2nd place 10 points
  • 3rd place 7 points
  • 4th place 4 points
  • 5th place 2 points
  • Finding bug 2 points
  • Suggesting Challenge 2 points

Here is Tom's winning solution:

Player.c
Copyright © 1998 Tom Saxton

#include "BlackJack.h"

// Naming Conventions:
//
// Without getting into the gory details of the Hungarian naming convention,
// here are some common prefixes and their meanings:
//
//      a       array
//      p       pointer
//      c       count
//      mp      map (one data type to another)
//      i       index
//
// The prefixes modify a base type. So, if FOO is a base type (like a struct,
// or an enum), the following declarations illustrate the above prefixes:
//
//      FOO     foo;
//      FOO *   pfoo;
//      FOO     afoo[10];
//      int     ifoo; // an index into an array of FOOs
//      int     cfoo; // a count of FOOs.
//
//      for (ifoo = 0; ifoo < cfoo; ++ifoo)
//         pfoo = &afoo[ifoo];
//

enum { fFalse = 0, fTrue = 1 };
#define DIM(a) (sizeof(a)/sizeof((a)[0]))

// Be sure to enable this define to pick up a couple of post-deadline bug fixes.
//
// #define BUGFIX

// disable debug code
#define Assert(f)

// The following tables determine the actions for all possible hands,
// divided into three groups: pairs, soft hands and hard hands, considered
// in that order. (A pair of aces is treated as a pair, not as a soft hand.)
//
// The tables were computed by taking the Dealer's up card and assuming
// a huge shoe with an even card distribution finding the probability
// for each of the possible final dealers scores (bust, 17, 18, 19, 20 and 21).
//
// Then, given that table, I computed the expected earnings (win, lose or
// push) for each of the possible actions, and recorded the action with
// the best result.
//
// I found the book "Best Blackjack" by Frank Scoblete (c) 1996 to be
// helpful, and my tables are close to his multi-deck tables.
// I modeled an infinite, evenly distributed shoe, he may have modeled
// a fixed number of decks.

// macros to make these tables manageable...
#define H kHitMe
#define D kDoubleDownAndHitMe
#define S kStandPat
#define X kSplitAndHitMe
#define B kClaimBlackjack

// For "hard" hands (no Aces scored as 11), plug in the dealer's up card
// (minus 1) and the hand's score to find the next action. If this isn't the
// first action of the hand, treat kDoubleDownAndHitMe as kHitMe.
Action mp_spot_score_actionHard[10][22] = 
{
//      0   - 21
   { H,H,H,H,H,H,H,H,H,H,H,H,H,H,H,H,H,S,S,S,S,S },   /*  A */
   { H,H,H,H,H,H,H,H,H,D,D,D,H,S,S,S,S,S,S,S,S,S },   /*  2 */
   { H,H,H,H,H,H,H,H,H,D,D,D,H,S,S,S,S,S,S,S,S,S },   /*  3 */
   { H,H,H,H,H,H,H,H,H,D,D,D,S,S,S,S,S,S,S,S,S,S },   /*  4 */
   { H,H,H,H,H,H,H,H,H,D,D,D,S,S,S,S,S,S,S,S,S,S },   /*  5 */
   { H,H,H,H,H,H,H,H,D,D,D,D,S,S,S,S,S,S,S,S,S,S },   /*  6 */
   { H,H,H,H,H,H,H,H,H,D,D,D,H,H,H,H,H,S,S,S,S,S },   /*  7 */
   { H,H,H,H,H,H,H,H,H,H,D,D,H,H,H,H,H,S,S,S,S,S },   /*  8 */
   { H,H,H,H,H,H,H,H,H,H,D,D,H,H,H,H,H,S,S,S,S,S },   /*  9 */
   { H,H,H,H,H,H,H,H,H,H,H,D,H,H,H,H,H,S,S,S,S,S }    /*  10 */
};

// For "soft" hands (at least one Ace used as an 11), plug in the dealer's up card
// (minus 1) and the hand's "other" card (or combined score without the ace)
// to find the next action.
Action mp_spot_spot_actionSoft[10][10] = 
{
   { H, H, H, H, H, H, S, S, S, B },   /*  A */
   { H, H, H, H, H, D, S, S, S, B },   /*  2 */
   { H, H, H, H, H, D, D, S, S, B },   /*  3 */
   { H, H, H, H, D, D, D, S, S, B },   /*  4 */
   { H, H, H, D, D, D, D, S, S, B },   /*  5 */
   { H, H, H, D, D, D, D, S, S, B },   /*  6 */
   { H, H, H, H, H, D, S, S, S, B },   /*  7 */
   { H, H, H, H, H, H, S, S, S, B },   /*  8 */
   { H, H, H, H, H, H, H, S, S, B },   /*  9 */
   { H, H, H, H, H, H, H, S, S, B }    /* 10 */
};

// If dealt a pair, plug in the dealer's up card (minus 1) and the spot
// value of the pair (minus 1) to find the suggested action.
Action mp_spot_spot_actionPair[10][10] = 
{
   { X, H, H, H, H, H, H, H, S, S },   /*  A */
   { X, X, X, H, D, H, X, X, X, S },   /*  2 */
   { X, X, X, H, D, H, X, X, X, S },   /*  3 */
   { X, X, X, H, D, X, X, X, X, S },   /*  4 */
   { X, X, X, H, D, X, X, X, X, S },   /*  5 */
   { X, X, X, D, D, X, X, X, X, S },   /*  6 */
   { X, X, X, H, D, H, X, X, S, S },   /*  7 */
   { X, X, X, H, D, H, H, X, X, S },   /*  8 */
   { X, X, H, H, D, H, H, X, X, S },   /*  9 */
   { X, H, H, H, H, H, H, H, S, S }    /*  10 */
};

// undefine shortcuts used in the above tables
#undef H
#undef D
#undef S
#undef X
#undef B

// Score bust as zero.
#define scoreBust            0
// The Player's hand is limited to five cards.
#define ccardMaxPlayer       5
// The dealer can never draw more than nine cards
// (nine 2's for example).
#define ccardMaxDealer      10

// Entry in the SPOT table, used for scoring and printing card values
typedef struct ESPOT
{
   char   score;
   char   sz[3];
} ESPOT;

static const ESPOT s_dnspot[kKing+1] = 
{
   {  0, "?" },
   {  1, "A" },
   {  2, "2" },
   {  3, "3" },
   {  4, "4" },
   {  5, "5" },
   {  6, "6" },
   {  7, "7" },
   {  8, "8" },
   {  9, "9" },
   { 10, "10" },
   { 10, "J" },
   { 10, "Q" },
   { 10, "K" },
};

// game statistics...
static int s_cdeck;
static int s_ccreditStart;
static int s_ccreditMinBet;
static int s_ccreditMidBet;
static int s_ccreditMaxBet;
static int s_ccreditBalance;
static int s_ccreditTotalBet;

// callback functions
static BetProc *s_pfnMakeABet;
static HitProc *s_pfnHitMe;

// function to score a hand
static int _ScoreHand(const Card acard[], int ccard,
                              int *pcAce);

// struct for counting cards...
typedef struct DECK
{
   int acspot[10+1];
   int cspotStart;
   int cspotRemain;
   int dcount;
   int fInfinite;
} DECK;
static DECK s_deck;
// calls to reset card counters and count the cards in a hand
void _InitDeck(int cdeck, int fInfinite);
void _CountCards(const Card acard[], int ccard);
// call to compute the proper action given the player's hand and
// the dealer's up card, and whether or not this is the first action
static Action _ActionLookupHand(Spot spotDealer, Card acard[], int ccard, int fFirst);

InitBlackjack
// Call to start a game
void InitBlackjack(
   int numDecks,       /* number of decks used by the dealer, 2..10*/
   int yourBankroll,   /* number of credits you have to start */
   int minBet,         /* minimum bet for each hand */
   int maxBet,         /* maximum bet for each hand */
   BetProc pfnMakeABet,/* callback to place a wager */
   HitProc pfnHitMe    /* callback to get a card */
)
{
   s_cdeck             = numDecks;
   s_ccreditStart      = yourBankroll;
   s_ccreditMinBet     = minBet;
   s_ccreditMaxBet     = maxBet;
   s_pfnMakeABet       = pfnMakeABet;
   s_pfnHitMe          = pfnHitMe;

   s_ccreditTotalBet   = 0;
   s_ccreditBalance    = s_ccreditStart;
}

Blackjack
// Call to play a hand
Boolean Blackjack(Boolean fNewDeck)
{
   int      ccreditBet;
   Card     acardPlayer[ccardMaxPlayer], acardDealer[ccardMaxDealer];
   int      ccardPlayer, ccardDealer;
   Action   actionFirst;
   int      ccreditWin;
   Spot     spotDealer;
   int      ihand, chand;
   int      count;
   Result   result;
   
   if (fNewDeck)
      _InitDeck(s_cdeck, fFalse /*fInfinite*/);
   // normalize the card count. A positive count means that the shoe is
   // heavy in large cards, which makes it more likely for the dealer to
   // bust. A negative count means that the shoe is heavy in small cards,
   // which makes it less likely that the dealer with bust.
   count = (s_deck.dcount*52)/s_deck.cspotRemain;
   // Make a bet that is proportional to our current balance, so that losing
   // streaks don't clean us out, and winning streaks rake in extra chips.
   // Bet more when the count is high, less when it's low, but stay within
   // the stated betting limits.
   ccreditBet = (count+2)*s_ccreditBalance/50;
   if (ccreditBet < s_ccreditMinBet)
      ccreditBet = s_ccreditMinBet;
   else if (ccreditBet > s_ccreditMaxBet)
      ccreditBet = s_ccreditMaxBet;
   // Place bet, get some cards
   (*s_pfnMakeABet)(ccreditBet, acardPlayer, acardDealer);
   ccardPlayer = ccardDealer = 2;
   // store and normalize the dealer's up card
   spotDealer = acardDealer[1].spot;
   if (spotDealer > k10)
      spotDealer = k10;
   // get the first action for the hand
   actionFirst = 
            _ActionLookupHand(spotDealer, acardPlayer, 2, fTrue);
   if (actionFirst == kDoubleDownAndHitMe)
      ccreditBet *= 2;
   chand = (actionFirst == kSplitAndHitMe) ? 2 : 1;
   // play out the hand(s) (there are two hands if we kSplitAndHitMe)
   for (ihand = 0; ihand < chand; ++ihand)
   {
      Boolean   fInsurance;
      Action action = actionFirst;
      // take the "insurance" side bet when the dealer shows an Ace and there is
      // a better than one third chance of the dealer having a 10 for the other card.
      fInsurance = (spotDealer == kAce) && 
                        (3*s_deck.acspot[k10] > s_deck.cspotRemain);
      for(;;)
      {
         // play out an action
         result = 
               (*s_pfnHitMe)(action, fInsurance, acardPlayer, 
                           &ccardPlayer, acardDealer, &ccardDealer, 
                                 &ccreditWin);
         if (result != kNoResult)
            break;
         
         // If we didn't kStandPat or bust, calculate the next action
         action = _
         ActionLookupHand(spotDealer, acardPlayer, ccardPlayer, 
                                                fFalse);
      }
      
      // count the cards shown in this hand
      _CountCards(acardPlayer, ccardPlayer);
      if (ihand == chand-1)
         _CountCards(acardDealer, ccardDealer);
      
      // tally our win/loss
      s_ccreditBalance += ccreditWin;
      s_ccreditTotalBet += ccreditBet;
   }

   // If we have lost most of our money, quit
   if (s_ccreditBalance < s_ccreditStart/3)
   {
      return fFalse;
   }
   // If we have won a lot, and will avoid the freeloader penalty, quit
   if (s_ccreditBalance > 7*s_ccreditStart/4 && 
            s_ccreditTotalBet > s_ccreditStart)
   {
      return fFalse;
   }
   // If we have won some, and played for twice the freeloader requirement, quit
   if (s_ccreditBalance > 5*s_ccreditStart/4 && 
            s_ccreditTotalBet > 2*s_ccreditStart)
   {
      return fFalse;
   }
   // If we haven't lost, and played for five times the freeloader requirement, quit
   if (s_ccreditBalance > s_ccreditStart && 
            s_ccreditTotalBet > 5*s_ccreditStart)
   {
      return fFalse;
   }
   // If we've played 10 times the freeloader penalty, quit before the time penalty
   // takes it all away...
   if (s_ccreditTotalBet > 10*s_ccreditStart)
   {
      return fFalse;
   }
   return fTrue;
}

_ActionLookupHand
// Call to get the next action for this hand
static Action _ActionLookupHand(Spot spotDealer, Card acard[], int ccard, int fFirst)
{
   int score, cAce;
   Spot spot;
   Action action;
   
   // get the hand's score, and the count of Aces scored as 11
   score = _ScoreHand(acard, ccard, &cace);
   Assert(kAce <= spotDealer && spotDealer <= k10);
   if (fFirst && ccard == 2 && acard[0].spot == acard[1].spot)
   {
      // first action on a pair, check the pair's table
      if ((spot = acard[0].spot) > k10)
         spot = k10;
      action = 
         (Action)mp_spot_spot_actionPair[spotDealer-1][spot-1];
   }
   else if (cAce > 0)
   {
      // "soft" hand, check the soft table
      spot = (Spot)(score - 11);
      Assert(kAce <= spot && spot <= k10);
#ifdef DEBUG
      if (ccard == 2)
      {
         int icard = acard[0].spot == kAce ? 1 : 0;
         Assert(spot == acard[icard].spot || 
                        (spot == k10 && acard[icard].spot > k10));
      }
#endif
      action = 
         (Action)mp_spot_spot_actionSoft[spotDealer-1][spot-1];
   }
   else
   {
      // "hard" hand, chech the hard table
      action = 
         (Action)mp_spot_score_actionHard[spotDealer-1][score];
   }
   
   // If it's not the first play of the hand, we can only kStandPat or kHitMe
#ifdef BUGFIX
   // Another Bug Fix: be careful trying to catch illegal actions...
   if (action == kClaimBlackjack && ccard != 2)
      action = kStandPat;
   if (!fFirst && (action == kDoubleDownAndHitMe || 
                                    action == kSplitAndHitMe))
      action = kHitMe;
#else
   // This code is wrong, it incorrectly returns kHitMe in two cases:
   //  1. If a pair of 10s or Aces is split, then one of them turned into a blackjack
   //  2. A score of 21 is reached with an Ace and two or more cards.
   if (!fFirst && action != kStandPat)
      action = kHitMe;
#endif
   return action;
}

_InitDeck
// reset the counts for a fresh set of decks
static void _InitDeck(int cdeck, int fInfinite)
{
   Spot spot;
   int cspot = 4*cdeck;
   for (spot = kAce; spot < k10; ++spot)
      s_deck.acspot[spot] = cspot;
   Assert(spot == k10);
   s_deck.acspot[k10] = 4*cspot;
   s_deck.cspotRemain = s_deck.cspotStart = 52*cdeck;
   s_deck.dcount = 0;
   s_deck.fInfinite = fInfinite;
}

_CountCards
// This is the counting method used by the couple of people I've talked
// who have actually counted cards playing Blackjack. It's call "Hi-Lo"
// in "Best Blackjack". I tried several other counting models listed
// in that book, and this performed the best. It gives a simple assessment
// of how far off balance the shoe is with respect to small and large cards.
static const int s_mp_spot_dcount[k10+1] =
//    A  2  3  4  5  6  7  8  9  10
{ 0, -1, 1, 1, 1, 1, 1, 0, 0, 0, -1 };

// Remove the specified set of cards from the shoe
void _CountCards(const Card acard[], int ccard)
{
   if (s_deck.fInfinite)
      return;
      
   while (ccard- > 0)
   {
      Spot spot = acard[ccard].spot;
// Bug Fix: we shouldn't count hidden cards...
#ifdef BUGFIX
      if (spot == kHiddenSpot)
         continue;
#endif
      if (spot > k10)
         spot = k10;
      -s_deck.acspot[spot];
      -s_deck.cspotRemain;
      
      s_deck.dcount += s_mp_spot_dcount[spot];
   }
}

_ScoreHand
// Determine the score for the given cards. When possible, score
// Aces at 11, and return the number of aces thusly scored.
static int _ScoreHand(const Card acard[], int ccard, int *pcAce)
{
   int cAceDummy;
   int score = 0;
   int cAce = 0;

   if (pcAce == NULL)
      pcAce = &cAceDummy;
   *pcAce = 0;
   while (ccard- > 0)
   {
      if (acard[ccard].spot == kAce)
         ++cAce;
      score += s_dnspot[acard[ccard].spot].score;
   }
   
   if (score > 21)
      return scoreBust;
   while (score + 10 <= 21 && cAce > 0)
   {
      score += 10;
      -cAce;
      ++*pcAce;
   }

   return score;
}
 
AAPL
$501.15
Apple Inc.
+2.47
MSFT
$34.67
Microsoft Corpora
+0.18
GOOG
$895.88
Google Inc.
+13.87

MacTech Search:
Community Search:

Software Updates via MacUpdate

Apple HP Printer Drivers 2.16.1 - For OS...
Apple HP Printer Drivers includes the latest HP printing and scanning software for Mac OS X 10.6, 10.7 and 10.8. For information about supported printer models, see this page.Version 2.16.1: This... Read more
Yep 3.5.1 - Organize and manage all your...
Yep is a document organization and management tool. Like iTunes for music or iPhoto for photos, Yep lets you search and view your documents in a comfortable interface, while offering the ability to... Read more
Apple Canon Laser Printer Drivers 2.11 -...
Apple Canon Laser Printer Drivers is the latest Canon Laser printing and scanning software for Mac OS X 10.6, 10.7 and 10.8. For information about supported printer models, see this page.Version 2.11... Read more
Apple Java for Mac OS X 10.6 Update 17 -...
Apple Java for Mac OS X 10.6 delivers improved security, reliability, and compatibility by updating Java SE 6.Version Update 17: Java for Mac OS X 10.6 Update 17 delivers improved security,... Read more
Arq 3.3 - Online backup (requires Amazon...
Arq is online backup for the Mac using Amazon S3 and Amazon Glacier. It backs-up and faithfully restores all the special metadata of Mac files that other products don't, including resource forks,... Read more
Apple Java 2013-005 - For OS X 10.7 and...
Apple Java for OS X 2013-005 delivers improved security, reliability, and compatibility by updating Java SE 6 to 1.6.0_65. On systems that have not already installed Java for OS X 2012-006, this... Read more
DEVONthink Pro 2.7 - Knowledge base, inf...
Save 10% with our exclusive coupon code: MACUPDATE10 DEVONthink Pro is your essential assistant for today's world, where almost everything is digital. From shopping receipts to important research... Read more
VirtualBox 4.3.0 - x86 virtualization so...
VirtualBox is a family of powerful x86 virtualization products for enterprise as well as home use. Not only is VirtualBox an extremely feature rich, high performance product for enterprise customers... Read more
Merlin 2.9.2 - Project management softwa...
Merlin is the only native network-based collaborative Project Management solution for Mac OS X. This version offers many features propelling Merlin to the top of Mac OS X professional project... Read more
Eye Candy 7.1.0.1191 - 30 professional P...
Eye Candy renders realistic effects that are difficult or impossible to achieve in Photoshop alone, such as Fire, Chrome, and the new Lightning. Effects like Animal Fur, Smoke, and Reptile Skin are... Read more

Halloween – iLovecraft Brings Frightenin...
Halloween – iLovecraft Brings Frightening Stories From Author H.P. | Read more »
The Blockheads Creator David Frampton Gi...
The Blockheads Creator David Frampton Gives a Postmortem on the Creation Process of the Game Posted by Andrew Stevens on October 16th, 2013 [ permalink ] Hey, a | Read more »
Sorcery! Enhances the Gameplay in Latest...
Sorcery! | Read more »
It Came From Australia: Tiny Death Star
NimbleBit and Disney have teamed up to make Star Wars: Tiny Death Star, a Star Wars take on Tiny Tower. Right now, the game is in testing in Australia (you will never find a more wretched hive of scum and villainy) but we were able to sneak past... | Read more »
FIST OF AWESOME Review
FIST OF AWESOME Review By Rob Rich on October 16th, 2013 Our Rating: :: TALK TO THE FISTUniversal App - Designed for iPhone and iPad A totalitarian society of bears is only the tip of the iceberg in this throwback brawler.   | Read more »
PROVERBidioms Paints English Sayings in...
PROVERBidioms Paints English Sayings in a Picture for Users to Find Posted by Andrew Stevens on October 16th, 2013 [ permalink ] | Read more »
OmniFocus 2 for iPhone Review
OmniFocus 2 for iPhone Review By Carter Dotson on October 16th, 2013 Our Rating: :: OMNIPOTENTiPhone App - Designed for the iPhone, compatible with the iPad OmniFocus 2 for iPhone is a task management app for people who absolutely... | Read more »
Ingress – Google’s Augmented-Reality Gam...
Ingress – Google’s Augmented-Reality Game to Make its Way to iOS Next Year Posted by Andrew Stevens on October 16th, 2013 [ permalink ] | Read more »
CSR Classics is Full of Ridiculously Pre...
CSR Classics is Full of Ridiculously Pretty Classic Automobiles Posted by Rob Rich on October 16th, 2013 [ permalink ] | Read more »
Costume Quest Review
Costume Quest Review By Blake Grundman on October 16th, 2013 Our Rating: :: SLIGHTLY SOURUniversal App - Designed for iPhone and iPad This bite sized snack lacks the staying power to appeal beyond the haunting season.   | Read more »

Price Scanner via MacPrices.net

Apple Store Canada offers refurbished 11-inch...
 The Apple Store Canada has Apple Certified Refurbished 2013 11″ MacBook Airs available starting at CDN$ 849. Save up to $180 off the cost of new models. An Apple one-year warranty is included with... Read more
Updated MacBook Price Trackers
We’ve updated our MacBook Price Trackers with the latest information on prices, bundles, and availability on MacBook Airs, MacBook Pros, and the MacBook Pros with Retina Displays from Apple’s... Read more
13-inch Retina MacBook Pros on sale for up to...
B&H Photo has the 13″ 2.5GHz Retina MacBook Pro on sale for $1399 including free shipping. Their price is $100 off MSRP. They have the 13″ 2.6GHz Retina MacBook Pro on sale for $1580 which is $... Read more
AppleCare Protection Plans on sale for up to...
B&H Photo has 3-Year AppleCare Warranties on sale for up to $105 off MSRP including free shipping plus NY sales tax only: - Mac Laptops 15″ and Above: $244 $105 off MSRP - Mac Laptops 13″ and... Read more
Apple’s 64-bit A7 Processor: One Step Closer...
PC Pro’s Darien Graham-Smith reported that Canonical founder and Ubuntu Linux creator Mark Shuttleworth believes Apple intends to follow Ubuntu’s lead and merge its desktop and mobile operating... Read more
MacBook Pro First, Followed By iPad At The En...
French site Info MacG’s Florian Innocente says he has received availability dates and order of arrival for the next MacBook Pro and the iPad from the same contact who had warned hom of the arrival of... Read more
Chart: iPad Value Decline From NextWorth
With every announcement of a new Apple device, serial upgraders begin selling off their previous models – driving down the resale value. So, with the Oct. 22 Apple announcement date approaching,... Read more
SOASTA Survey: What App Do You Check First in...
SOASTA Inc., the leader in cloud and mobile testing announced the results of its recent survey showing which mobile apps are popular with smartphone owners in major American markets. SOASTA’s survey... Read more
Apple, Samsung Reportedly Both Developing 12-...
Digitimes’ Aaron Lee and Joseph Tsai report that Apple and Samsung Electronics are said to both be planning to release 12-inch tablets, and that Apple is currently cooperating with Quanta Computer on... Read more
Apple’s 2011 MacBook Pro Lineup Suffering Fro...
Appleinsider’s Shane Cole says that owners of early-2011 15-inch and 17-inch MacBook Pros are reporting issues with those models’ discrete AMD graphics processors, which in some cases results in the... Read more

Jobs Board

Senior Mac / *Apple* Systems Engineer - 318...
318 Inc, a top provider of Apple solutions is seeking a new Senior Apple Systems Engineer to be based out of our Santa Monica, California location. We are a Read more
*Apple* Retail - Manager - Apple Inc. (Unite...
Job Summary Keeping an Apple Store thriving requires a diverse set of leadership skills, and as a Manager, you’re a master of them all. In the store’s fast-paced, Read more
*Apple* Solutions Consultant - Apple (United...
**Job Summary** Apple Solutions Consultant (ASC) - Retail Representatives Apple Solutions Consultants are trained by Apple on selling Apple -branded products Read more
Associate *Apple* Solutions Consultant - Ap...
**Job Summary** The Associate ASC is an Apple employee who serves as an Apple brand ambassador and influencer in a Reseller's store. The Associate ASC's role is to Read more
*Apple* Solutions Consultant (ASC) - Apple (...
**Job Summary** The ASC is an Apple employee who serves as an Apple brand ambassador and influencer in a Reseller's store. The ASC's role is to grow Apple Read more
All contents are Copyright 1984-2011 by Xplain Corporation. All rights reserved. Theme designed by Icreon.