Logic question...

J

JoeC

I am writing a game and I am having a challenge with my combat
function. All I want to do is find out how to group pieces that are in
the same space. There are two sides and all the units that are in the
same space fight. I want to add up the attack factors and defending
factors in the same space then figure out the odds so I can roll
against an odds table. Basically each piece holds its own x and y loc.


Here is what I have right now:

void fight(vector<unit>& at, vector<unit>& dt,
board * b , terrain * trn ){

vector<fightclass> fighting;

/* adds attacking units to the fightclass vector if units are in the
same space they are to grouped in the same fight class elemet of
the
vector. */

for(int atu = 0; atu != at.size(); atu++){
if(fighting.size() == 0){
fightclass ft;
ft.addAu(at[atu]);
fighting.push_back(ft);
} else {
for(int lp = 0; lp != fighting.size(); lp++){
if(at[atu].getXloc() != fighting[lp].getX() &&
at[atu].getYloc() != fighting[lp].getY()){
fightclass ft;
ft.addAu(at[atu]);
fighting.push_back(ft);
} else {
fighting[lp].addAu(at[atu]);
}
}
}
}
/* Adds defending units to the fightclass array. If x and y locs are
the same as attacking locations (are in the same space) they are
added to array for combat */

for(int dtu = 0; dtu != dt.size(); dtu++){
for(int lp = 0; lp != fighting.size(); lp++){
if(dt[dtu].getXloc() == fighting[lp].getX() &&
dt[dtu].getYloc() == fighting[lp].getY()){
fighting[lp].addDu(dt[dtu]);
}
}
}

// Combat routine

for(int lp = 0; lp != fighting.size(); lp++){ //handles combat
if(fighting[lp].canfight()){
int df = b->GetSpace(fighting[lp].getX(), fighting[lp].getY());
float odds =
fighting[lp].getAtk()/fighting[lp].getDef();
//gets the defense bonus for the terrain in the space where
//combat takes place
int roll = rand() - trn[df].defend();
//get the die roll modified for terrain
odds = fighting[lp].getAtk() / fighting[lp].getDef();
//gets the attack to defence ratio.
if(odds < .5){
MessageBox(NULL, "Fighting! 1:3", "Info!", MB_OK);
return;
}
if(odds < 1){
MessageBox(NULL, "Fighting! 1:2", "Info!", MB_OK);
return;
}
if(odds < 2){
MessageBox(NULL, "Fighting! 1:1", "Info!", MB_OK);
return;
}
if(odds < 3){
MessageBox(NULL, "Fighting! 2:1", "Info!", MB_OK);
return;
}
if(odds < 4){
MessageBox(NULL, "Fighting! 3:1", "Info!", MB_OK);
return;
}
if(odds < 5){
MessageBox(NULL, "Fighting! 4:1", "Info!", MB_OK);
return;
}
if(odds < 6){
MessageBox(NULL, "Fighting! 5:1", "Info!", MB_OK);
return;
}

}
}
for(int lp = 0; lp != fighting.size(); lp++){
fighting[lp].done();
}
fighting.clear();
}

class fightclass{
/* Fightclass holds two arrays of units. The two arrays represent
units
in the same space elgible for combat. Array at represents the
attacking
forces and dt represents the defending units */


int xl;
int yl;

bool sides2; /* indicates if there are units in both attack and
defend
if there are units in both attack and defend then
combat
is to take place between the opposing sides. */

vector<unit>at; //attack units
vector<unit>dt; //defending units

public:
fightclass();
void addAu(unit au); //adds an addacking unit
void addDu(unit du); //adds a defending unit
int getX(){return xl;} //return the x coord
int getY(){return yl;} //returns the y coord
float getAtk(); //returns the total attack factors for one space
float getDef(); //returns the total defending factors for one space
bool canfight(); //if there are both attacking and defending units
void done(); //clears vectores after the turn

};

fightclass::fightclass(){
xl = 0;
yl = 0;
}

void fightclass::addAu(unit u){
at.push_back(u);
xl = u.getXloc();
yl = u.getYloc();

}

void fightclass::addDu(unit u){

dt.push_back(u);

}

bool fightclass::canfight(){
if(at.size() && dt.size()){
return true;
}else {return false;}
}

float fightclass::getAtk(){
float total;
for(int lp = 0; lp != at.size(); lp++){
total +=at[lp].getAttack();
}
return total;
}

I know the logic is not the best but it is the best I can do. This is
tricky and I am pretty confused. I would like to find some way of
making the logic steps easier. There may be some simple error that I
overlooked or it can be totally screwed up.
 
M

Moonlit

Hi,

comp.games.development.programming is probably more appropiate for this
question.

--


Regards, Ron AF Greve

http://moonlit.xs4all.nl

JoeC said:
I am writing a game and I am having a challenge with my combat
function. All I want to do is find out how to group pieces that are in
the same space. There are two sides and all the units that are in the
same space fight. I want to add up the attack factors and defending
factors in the same space then figure out the odds so I can roll
against an odds table. Basically each piece holds its own x and y loc.


Here is what I have right now:

void fight(vector<unit>& at, vector<unit>& dt,
board * b , terrain * trn ){

vector<fightclass> fighting;

/* adds attacking units to the fightclass vector if units are in the
same space they are to grouped in the same fight class elemet of
the
vector. */

for(int atu = 0; atu != at.size(); atu++){
if(fighting.size() == 0){
fightclass ft;
ft.addAu(at[atu]);
fighting.push_back(ft);
} else {
for(int lp = 0; lp != fighting.size(); lp++){
if(at[atu].getXloc() != fighting[lp].getX() &&
at[atu].getYloc() != fighting[lp].getY()){
fightclass ft;
ft.addAu(at[atu]);
fighting.push_back(ft);
} else {
fighting[lp].addAu(at[atu]);
}
}
}
}
/* Adds defending units to the fightclass array. If x and y locs are
the same as attacking locations (are in the same space) they are
added to array for combat */

for(int dtu = 0; dtu != dt.size(); dtu++){
for(int lp = 0; lp != fighting.size(); lp++){
if(dt[dtu].getXloc() == fighting[lp].getX() &&
dt[dtu].getYloc() == fighting[lp].getY()){
fighting[lp].addDu(dt[dtu]);
}
}
}

// Combat routine

for(int lp = 0; lp != fighting.size(); lp++){ //handles combat
if(fighting[lp].canfight()){
int df = b->GetSpace(fighting[lp].getX(), fighting[lp].getY());
float odds =
fighting[lp].getAtk()/fighting[lp].getDef();
//gets the defense bonus for the terrain in the space where
//combat takes place
int roll = rand() - trn[df].defend();
//get the die roll modified for terrain
odds = fighting[lp].getAtk() / fighting[lp].getDef();
//gets the attack to defence ratio.
if(odds < .5){
MessageBox(NULL, "Fighting! 1:3", "Info!", MB_OK);
return;
}
if(odds < 1){
MessageBox(NULL, "Fighting! 1:2", "Info!", MB_OK);
return;
}
if(odds < 2){
MessageBox(NULL, "Fighting! 1:1", "Info!", MB_OK);
return;
}
if(odds < 3){
MessageBox(NULL, "Fighting! 2:1", "Info!", MB_OK);
return;
}
if(odds < 4){
MessageBox(NULL, "Fighting! 3:1", "Info!", MB_OK);
return;
}
if(odds < 5){
MessageBox(NULL, "Fighting! 4:1", "Info!", MB_OK);
return;
}
if(odds < 6){
MessageBox(NULL, "Fighting! 5:1", "Info!", MB_OK);
return;
}

}
}
for(int lp = 0; lp != fighting.size(); lp++){
fighting[lp].done();
}
fighting.clear();
}

class fightclass{
/* Fightclass holds two arrays of units. The two arrays represent
units
in the same space elgible for combat. Array at represents the
attacking
forces and dt represents the defending units */


int xl;
int yl;

bool sides2; /* indicates if there are units in both attack and
defend
if there are units in both attack and defend then
combat
is to take place between the opposing sides. */

vector<unit>at; //attack units
vector<unit>dt; //defending units

public:
fightclass();
void addAu(unit au); //adds an addacking unit
void addDu(unit du); //adds a defending unit
int getX(){return xl;} //return the x coord
int getY(){return yl;} //returns the y coord
float getAtk(); //returns the total attack factors for one space
float getDef(); //returns the total defending factors for one space
bool canfight(); //if there are both attacking and defending units
void done(); //clears vectores after the turn

};

fightclass::fightclass(){
xl = 0;
yl = 0;
}

void fightclass::addAu(unit u){
at.push_back(u);
xl = u.getXloc();
yl = u.getYloc();

}

void fightclass::addDu(unit u){

dt.push_back(u);

}

bool fightclass::canfight(){
if(at.size() && dt.size()){
return true;
}else {return false;}
}

float fightclass::getAtk(){
float total;
for(int lp = 0; lp != at.size(); lp++){
total +=at[lp].getAttack();
}
return total;
}

I know the logic is not the best but it is the best I can do. This is
tricky and I am pretty confused. I would like to find some way of
making the logic steps easier. There may be some simple error that I
overlooked or it can be totally screwed up.
 
J

JoeC

Thanks, I didn't know about that group. I will ask the question there.
Still it is programming and I am trying to learn programming by
programming games.
 
J

Jerry Coffin

[ ... ]
void fight(vector<unit>& at, vector<unit>& dt,
board * b , terrain * trn ){

vector<fightclass> fighting;

/* adds attacking units to the fightclass vector if units are in the
same space they are to grouped in the same fight class elemet of
the
vector. */

for(int atu = 0; atu != at.size(); atu++){
if(fighting.size() == 0){

A minor detail, but this is better written as:
if (fighting.empty()) {
Given that this is a vector, it probably doesn't make any difference,
but IMO, it's still worth doing.
fightclass ft;
ft.addAu(at[atu]);
fighting.push_back(ft);
} else {
for(int lp = 0; lp != fighting.size(); lp++){

Looking just at the tiny bit starting here:
if(at[atu].getXloc() != fighting[lp].getX() &&
at[atu].getYloc() != fighting[lp].getY()){

I see what looks like an opportunity for a bit of improvement: create a
class (or struct) to represent a position, and then use it wherever you
have positions. In that case, this would become:
if (at[atu].pos == fighting[lp].pos)
// do whatever

I'd also attempt to use consistent naming -- whether you decide on
'pos' or 'position' or 'location' or 'coordinate' or whatever, attempt
to use the same name everywhere you have to deal with that kind of
thing. In the code above, you've got getXloc and getX, both apparently
doing the same kind of thing. While it's obvious after the fact that
both do the same kinds of things, when you're writing code, it's nice if
you can guess ahead of time what the name of a member is going to be
instead of having to check every time whether it's location, position,
etc. in this particular case.
fightclass ft;
ft.addAu(at[atu]);
fighting.push_back(ft);

In this case, the 'whatever' part can be simplified a bit as well. I'd
create a fightclass ctor that takes a unit as its argument, so it
creates a fightclass object with a single attacking unit, and then use
it here:
fighting.push_back(fightclass(at[atu]));
} else {
fighting[lp].addAu(at[atu]);
}
}
}
}

At least if I'm reading the intent correctly, I'd consider this much
enough for one function, though the function might include the following
chunk as well:
/* Adds defending units to the fightclass array. If x and y locs are
the same as attacking locations (are in the same space) they are
added to array for combat */

for(int dtu = 0; dtu != dt.size(); dtu++){
for(int lp = 0; lp != fighting.size(); lp++){
if(dt[dtu].getXloc() == fighting[lp].getX() &&
dt[dtu].getYloc() == fighting[lp].getY()){
fighting[lp].addDu(dt[dtu]);
}
}
}

Down to here. To implement the logic as I understand it, I'd make it two
functions, but to implement the logic I think you really want, I'd make
it one. See comments below for more on that.
// Combat routine

for(int lp = 0; lp != fighting.size(); lp++){ //handles combat

Starting from here:
if(fighting[lp].canfight()){
int df = b->GetSpace(fighting[lp].getX(), fighting[lp].getY());
float odds =
fighting[lp].getAtk()/fighting[lp].getDef();
//gets the defense bonus for the terrain in the space where
//combat takes place
int roll = rand() - trn[df].defend();
//get the die roll modified for terrain
odds = fighting[lp].getAtk() / fighting[lp].getDef();
//gets the attack to defence ratio.

.... and going down to here probably deserves to be yet another function.
if(odds < .5){
MessageBox(NULL, "Fighting! 1:3", "Info!", MB_OK);
return;
}
if(odds < 1){
MessageBox(NULL, "Fighting! 1:2", "Info!", MB_OK);
return;
}
if(odds < 2){
MessageBox(NULL, "Fighting! 1:1", "Info!", MB_OK);
return;
}
if(odds < 3){
MessageBox(NULL, "Fighting! 2:1", "Info!", MB_OK);
return;
}
if(odds < 4){
MessageBox(NULL, "Fighting! 3:1", "Info!", MB_OK);
return;
}
if(odds < 5){
MessageBox(NULL, "Fighting! 4:1", "Info!", MB_OK);
return;
}
if(odds < 6){
MessageBox(NULL, "Fighting! 5:1", "Info!", MB_OK);
return;
}

This chunk probably also deserves to be a function -- but I'd write it a
bit differently, something like:

void show_odds(double odds) {
//
// For the moment I'm assuming that odds is always in the range 0..6.
// For the moment, I've cheated, and conflated 1:3 and 1:2 together into
// "outnumbered".
//
static char *strings[] = {
"Outnumbered", "1:1", "2:1", "3:1", "4:1", "5:1"};

string output("Fighting! ");
output += strings[(int)odds];
MessageBox(NULL, output, "Info!", MB_OK);
}

The following is meant primarily as pseudo-code, not really C++. Chances
are that I've left out a few things that really need to be there (like
the calls to fightclass::done). Still, adding things like that shouldn't
be too horrible.

void fight(vector<unit> &at, vector<unit> &dt, board *b, terrain *trn)
{
vector<fightclass> fighting;

add_attacking_units(at.begin(), at.end(),
std::back_inserter(fighting));

add_defending_units(dt.begin(), dt.end(),
std::back_inserter(fighting));

std::vector<double> odds_values(fighting.size());

std::transform(fighting.begin(), fighting.end(),
odds_values.begin(), bind1st(figure_odds, trn));

std::for_each(odd_values.begin(), odds_values.end(), show_odds);
}

For the moment, I've used your function signature, but chances are
pretty good that (for example) all the parameters should refer to const
values. Perhaps the attacking/defending forces can be killed so those
vectors can be modified, but it's almost certain that the board and
probably terrain as well shouldn't be changed -- unless you're getting a
lot more sophisticated than most, and taking into account things like
artillery changing the terrain slightly.

At least as I undertand the idea here, I'd probably also change the
basic algorithm a bit. Unless I'm misreadin it, right now you're adding
a fightclass object to the fighting vector for every attacking unit,
whether there's a defending unit at the same position or not. Then
you're adding defending units everywhere there are attacking units. I
believe your real intent is that units only fight each other where
they're at the same location, so you really want an intersection of the
two -- i.e. a set of fight objects representing those places there's
both an attacker and a defender in the same place.

If that's the case, I'd consider doing things a bit differently. One
possibility would be to keep each force in a multiset, with the units
sorted by position. To move a unit, you remove it from the multiset, and
then insert it back in at the new position. In this case, finding the
intersection between two forces (i.e. the units they have in the same
positions) might be as simple as a single call to std:set_intersection.

I'm not at all sure why board and/or terrain is a pointer either -- it
appears these could be references to vectors, or something on that
order. For the moment, I'm leaving the alone, but I suspect some
reexamination of this point might be worthwhile.

Applying those comments, we could end up with a function something like:

void fight(multiset<unit> const &at, // attacking units
multiset<unit> const &dt, // defending units
board const *b,
terrain const *trn)
{
vector<fightclass> fighting;

find_fighting_units(at.begin(), at.end(),
dt.begin(), dt.end(),
std::back_inserter(fighting));

vector<double> odds_values(fighting.size());

transform(fighting.begin(), fighting.end(),
odds_values.begin(), bind1st(figure_odds, trn));

for_each(odd_values.begin(), odds_values.end(), show_odds);
}

Though it's more or less off-topic, it also seems like using MessageBox
for showing the odds is a poor idea -- if there are many intersections
between the forces, the user is going to have to click through a lot of
boxes, and as soon as he clicks OK on one, he loses that information, so
it's hard for him to any kind of overall picture.
 
J

JoeC

Thank...


Jerry said:
[ ... ]
void fight(vector<unit>& at, vector<unit>& dt,
board * b , terrain * trn ){

vector<fightclass> fighting;

/* adds attacking units to the fightclass vector if units are in the
same space they are to grouped in the same fight class elemet of
the
vector. */

for(int atu = 0; atu != at.size(); atu++){
if(fighting.size() == 0){

A minor detail, but this is better written as:
if (fighting.empty()) {
Given that this is a vector, it probably doesn't make any difference,
but IMO, it's still worth doing.
Thanks fixed that one.
fightclass ft;
ft.addAu(at[atu]);
fighting.push_back(ft);
} else {
for(int lp = 0; lp != fighting.size(); lp++){

Looking just at the tiny bit starting here:
if(at[atu].getXloc() != fighting[lp].getX() &&
at[atu].getYloc() != fighting[lp].getY()){

I see what looks like an opportunity for a bit of improvement: create a
class (or struct) to represent a position, and then use it wherever you
have positions. In that case, this would become:
if (at[atu].pos == fighting[lp].pos)
// do whatever
I have been considering that.
I'd also attempt to use consistent naming -- whether you decide on
'pos' or 'position' or 'location' or 'coordinate' or whatever, attempt
to use the same name everywhere you have to deal with that kind of
thing. In the code above, you've got getXloc and getX, both apparently
doing the same kind of thing. While it's obvious after the fact that
both do the same kinds of things, when you're writing code, it's nice if
you can guess ahead of time what the name of a member is going to be
instead of having to check every time whether it's location, position,
etc. in this particular case.

That is true. Naming conventions are important. I just wrote this
program at different times where some objects left over from older
projects.
fightclass ft;
ft.addAu(at[atu]);
fighting.push_back(ft);

In this case, the 'whatever' part can be simplified a bit as well. I'd
create a fightclass ctor that takes a unit as its argument, so it
creates a fightclass object with a single attacking unit, and then use
it here:
fighting.push_back(fightclass(at[atu]));
} else {
fighting[lp].addAu(at[atu]);
}
}
}
}

Intersting, I will have to look into that one. There are times I don't
do that because it is easier for me to read. But yes, it is a good
idea.
At least if I'm reading the intent correctly, I'd consider this much
enough for one function, though the function might include the following
chunk as well:
/* Adds defending units to the fightclass array. If x and y locs are
the same as attacking locations (are in the same space) they are
added to array for combat */

for(int dtu = 0; dtu != dt.size(); dtu++){
for(int lp = 0; lp != fighting.size(); lp++){
if(dt[dtu].getXloc() == fighting[lp].getX() &&
dt[dtu].getYloc() == fighting[lp].getY()){
fighting[lp].addDu(dt[dtu]);
}
}
}

Down to here. To implement the logic as I understand it, I'd make it two
functions, but to implement the logic I think you really want, I'd make
it one. See comments below for more on that.

OK that was the intent of writing this. I couldn't figure out how to
condense this to functions.
// Combat routine

for(int lp = 0; lp != fighting.size(); lp++){ //handles combat

Starting from here:
if(fighting[lp].canfight()){
int df = b->GetSpace(fighting[lp].getX(), fighting[lp].getY());
float odds =
fighting[lp].getAtk()/fighting[lp].getDef();
//gets the defense bonus for the terrain in the space where
//combat takes place
int roll = rand() - trn[df].defend();
//get the die roll modified for terrain
odds = fighting[lp].getAtk() / fighting[lp].getDef();
//gets the attack to defence ratio.

... and going down to here probably deserves to be yet another function.
if(odds < .5){
MessageBox(NULL, "Fighting! 1:3", "Info!", MB_OK);
return;
}
if(odds < 1){
MessageBox(NULL, "Fighting! 1:2", "Info!", MB_OK);
return;
}
if(odds < 2){
MessageBox(NULL, "Fighting! 1:1", "Info!", MB_OK);
return;
}
if(odds < 3){
MessageBox(NULL, "Fighting! 2:1", "Info!", MB_OK);
return;
}
if(odds < 4){
MessageBox(NULL, "Fighting! 3:1", "Info!", MB_OK);
return;
}
if(odds < 5){
MessageBox(NULL, "Fighting! 4:1", "Info!", MB_OK);
return;
}
if(odds < 6){
MessageBox(NULL, "Fighting! 5:1", "Info!", MB_OK);
return;
}

This chunk probably also deserves to be a function -- but I'd write it a
bit differently, something like:

This is all experiemtal. I am not sure how to implenmt an odds table.
All this is just test to see if the prvious code worked. What I want
to do is implemnt an odds table

1:3 1:2 1:1 2:1 3:1 4:1 5:1
0 ae ae ae ae ae ad ad
1 ae ae ae ae ad ad dd
2 ae ae ae ad ad dd dd
3 ae ae ad ad dd dd de
4 ae ae ad dd dd de de
5 ae dd dd dd ad de de

ae =attacker eliminated
ad = attacker dispersed
dd = defebder dispersed
de = defender eliminated
void show_odds(double odds) {
//
// For the moment I'm assuming that odds is always in the range 0..6.
// For the moment, I've cheated, and conflated 1:3 and 1:2 together into
// "outnumbered".
//
static char *strings[] = {
"Outnumbered", "1:1", "2:1", "3:1", "4:1", "5:1"};

string output("Fighting! ");
output += strings[(int)odds];
MessageBox(NULL, output, "Info!", MB_OK);
}
I have to lower odds for when I expand the game.
The following is meant primarily as pseudo-code, not really C++. Chances
are that I've left out a few things that really need to be there (like
the calls to fightclass::done). Still, adding things like that shouldn't
be too horrible.

void fight(vector<unit> &at, vector<unit> &dt, board *b, terrain *trn)
{
vector<fightclass> fighting;

add_attacking_units(at.begin(), at.end(),
std::back_inserter(fighting));

add_defending_units(dt.begin(), dt.end(),
std::back_inserter(fighting));

std::vector<double> odds_values(fighting.size());

std::transform(fighting.begin(), fighting.end(),
odds_values.begin(), bind1st(figure_odds, trn));

std::for_each(odd_values.begin(), odds_values.end(), show_odds);
}

For the moment, I've used your function signature, but chances are
pretty good that (for example) all the parameters should refer to const
values. Perhaps the attacking/defending forces can be killed so those
vectors can be modified, but it's almost certain that the board and
probably terrain as well shouldn't be changed -- unless you're getting a
lot more sophisticated than most, and taking into account things like
artillery changing the terrain slightly.

I can see the differences from my code and can see the weakness of my
programming. I am not fimilliar with this kind of programming. Are
these algorithms? I am a bit confused and would have not come up with
this. Are there smaller experiment programs I can write to better
understand how to do things like that?
At least as I undertand the idea here, I'd probably also change the
basic algorithm a bit. Unless I'm misreadin it, right now you're adding
a fightclass object to the fighting vector for every attacking unit,
whether there's a defending unit at the same position or not. Then
you're adding defending units everywhere there are attacking units. I
believe your real intent is that units only fight each other where
they're at the same location, so you really want an intersection of the
two -- i.e. a set of fight objects representing those places there's
both an attacker and a defender in the same place.

Yes, I just couldn't figure out how to do that. I simply want units in
the same space to add up attack factors then add up defedning factors
then have them fight. I put each attaking unit in some vector. Then
see it they were in the same locations as other units. Then I went to
the defending units and compared them with the other array and if they
are in the same location, put them together int he fightclass.
If that's the case, I'd consider doing things a bit differently. One
possibility would be to keep each force in a multiset, with the units
sorted by position. To move a unit, you remove it from the multiset, and
then insert it back in at the new position. In this case, finding the
intersection between two forces (i.e. the units they have in the same
positions) might be as simple as a single call to std:set_intersection.

Again, I am not a very good programmer and not awaire of many of these
techniques. I wrote a dungeon program where my map was an array of
spaces that could hole the graphic for the space, the character then
potentially other things. That became overly complex and difficult to
remove the player from one space then putting it in aother. Just
having the units hold its own location then have them display themselvs
is much simpler but grouping them togert for fighting is a real
challenge.
I'm not at all sure why board and/or terrain is a pointer either -- it
appears these could be references to vectors, or something on that
order. For the moment, I'm leaving the alone, but I suspect some
reexamination of this point might be worthwhile.
I don't need vectors because they don't change in size. I made them
pointers because I want to create them in one file then use them in
another. It is ieasier to pass them to functions than regular arrays.


Still I am going to save this post in hopes that I can improve how I
add
 
B

Bart

Moonlit said:
comp.games.development.programming is probably more appropiate for this
question.

Why? His question was more or less about program logic, using mostly
standard C++ (save for one function). That's more on-topic than many
other posts here.

Regards,
Bart.
 
P

Phlip

Bart said:
Moonlit wrote:

Why? His question was more or less about program logic, using mostly
standard C++ (save for one function). That's more on-topic than many
other posts here.

Only the OP should guess which group will provide the best answer.
 
N

Nathan Mates

Only the OP should guess which group will provide the best answer.

Posters are entitled to guess, true. But, correcting guesses is the
job of newsgroup regulars, and prefectly reasonable.

Nathan Mates
 
P

Phlip

Nathan said:
Posters are entitled to guess, true. But, correcting guesses is the
job of newsgroup regulars, and prefectly reasonable.

The only topicality criteria relevant to a questioner is what newsgroup will
provide the best question. Once we suggestion one, we can no longer declare
whether an equally topical question belongs here or there. That was the
context.

When you offer another newsgroup, always reinforce that topicality is in the
poster's best interests. The answer "your question is off topic here go
away" doesn't increase the odds of useful participation.
 
J

JoeC

Phlip said:
Only the OP should guess which group will provide the best answer.

I think MY code works. It is a bit more complex and cumbersom than I
would like but I am pushing ahead on next part of the game.
 
P

Phlip

Moonlit said:
Thanks Nathan :)

You correctly indicated a good newsgroup. Nobody doubted that.

Except Bart, who claimed you shouldn't have bounced.

I defended you, by saying you didn't bounce, you suggested.

Then Nathan attempted to refute my defense, which he didn't understand.

So I think your thanks are due me, not Nathan. ;-)
 
M

Moonlit

Hi,



Regards, Ron AF Greve

http://moonlit.xs4all.nl

Phlip said:
The only topicality criteria relevant to a questioner is what newsgroup
will provide the best question. Once we suggestion one, we can no longer
declare whether an equally topical question belongs here or there. That
was the context.

When you offer another newsgroup, always reinforce that topicality is in
the poster's best interests. The answer "your question is off topic here
go away" doesn't increase the odds of useful participation.
I didn't see anyone say in this thread "your question is off topic here go
away"
I merely suggested a better newsgroup.

And yes if I quote from the FAQ "Ultimately this means your question must be
answerable by looking into the C++ language definition as determined by the
ISO/ANSI C++ Standard document, and by planned extensions and adjustments".
Well it isn't answerable by looking in the C++ language definition. So
basically the question is off topic. Yet you didn't see an OT in my header,
simply because I have better things to do than to police newsgroups. Besides
that if you look at my responses in the past you will notice that I try to
help whether OT or not OT.

I suggested a better newsgroup and thats all there is, the rest is a result
of your imagination, really.

Regards, Ron AF Greve.
 
P

Phlip

Moonlit said:
I merely suggested a better newsgroup.

I never said you didn't. However...
And yes if I quote from the FAQ "Ultimately this means your question must
be answerable by looking into the C++ language definition as determined by
the ISO/ANSI C++ Standard document, and by planned extensions and
adjustments". Well it isn't answerable by looking in the C++ language
definition. So basically the question is off topic. Yet you didn't see an
OT in my header, simply because I have better things to do than to police
newsgroups. Besides that if you look at my responses in the past you will
notice that I try to help whether OT or not OT.

Ultimately, the FAQ was written by a human, and it is not authoritative.

Ultimately, the term "ultimately" means "at the end of a chain of events".
So if I ask about the Class Factory Pattern, or MVC, or huffman compression,
and I ask in platform-neutral terms, then ultimately the exact answer will
depend on the C++ ISO Standard. Hence, by the criteria of the
non-authoritative FAQ, such questions are on-topic.

This newsgroup is not , and posters should always pick the
narrowest newsgroup for their questions. So questions that only relate to
interpretation of The Standard belong there, not here.
 
J

Jerry Coffin

[ ... ]
I didn't see anyone say in this thread "your question is off topic here go
away"
I merely suggested a better newsgroup.

I'm not even sure that's true -- at least to me, it seems that most of
the discussion has centered primarily around how to express his ideas in
C++, not so much about the fundamental design of the game itself. That
being the case, I think the questions are really more topical here --
though it may be that enough game programming is done in C++ that it
would also fit well there.
 
J

Jerry Coffin

[ ... ]
This is all experiemtal. I am not sure how to implenmt an odds table.
All this is just test to see if the prvious code worked. What I want
to do is implemnt an odds table

1:3 1:2 1:1 2:1 3:1 4:1 5:1
0 ae ae ae ae ae ad ad
1 ae ae ae ae ad ad dd
2 ae ae ae ad ad dd dd
3 ae ae ad ad dd dd de
4 ae ae ad dd dd de de
5 ae dd dd dd ad de de

ae =attacker eliminated
ad = attacker dispersed
dd = defebder dispersed
de = defender eliminated

enum results { AttElim, AttDisp, DefDisp, DefElim };

results table[5][7] = {
{ AttElim, AttElim, AttElim, AttElim, AttElim, AttDisp, AttDisp},
{ AttElim, AttElim, AttElim, AttElim, AttDisp, AttDisp, AttDisp},
// ...

At least as I understand it, you'd then have to convert your odds to an
integer (something along the lines I used in show_odds might work).
You'd apparently index into the array above using the value on the die
in one dimension and the odds in the other dimension. Alternatively, you
might be able to simply multiply the odds by the die throw, and figure
out the result mathematically (your table looks like the relationship is
basically linear, so this probably wouldn't be terribly difficult.
I can see the differences from my code and can see the weakness of my
programming. I am not fimilliar with this kind of programming. Are
these algorithms? I am a bit confused and would have not come up with
this. Are there smaller experiment programs I can write to better
understand how to do things like that?

As it worked out, did this code more or less "bottom-up" -- i.e. looked
at your one long function, broke it down into some smaller functions,
and then composed a function that would call those in turn.

Generally, I'd advise doing more or less the opposite of that: start by
writing a description of what needs to happen in relatively general
terms. Especially at first, I'd advise breaking out pencil and paper for
this part of the task. Doing worry a lot about it being particular neat
or tidy, just get the general ideas of what needs to happen. In a lot of
cases, something on the order of an outline or bulleted list works
reasoanbly -- i.e. try to come up with a _few_ works that generally
describe a step.

Then take each of those steps, and do roughly the same thing again --
break it down into a number of simpler steps, each of which can still be
described in a few words.

When you've gotten to the point that everything is broken down into
relatively simple steps, the coding is (hopefully) pretty easy. The top-
level outline/list transforms more or less directly into the top-level
function. Probably nearly every line there will end up as a function
call. Each of the subordinate descriptions becomes another function as
well -- until each function has been broken down to the point that it's
a trivial little bit of code you can read/write/understand quite easily.

A few points: I'm really serious about the pencil and paper part -- if
you're at the computer, it's tempting to write code instead. At some
point, you might consider using one of the cool Computer Aided Software
Engineering (CASE) tools to draw UML or something like that -- but for
the moment, I wouldn't advise it. Despite decades of work, CASE tools
still really aren't the best for quickly roughing out a design; once you
have a relatively complete design, they can work nicely for making
incremental improvements, playing around with the effect of various
changes, and so on -- but that comes a bit later as a rule.
Yes, I just couldn't figure out how to do that. I simply want units in
the same space to add up attack factors then add up defedning factors
then have them fight. I put each attaking unit in some vector. Then
see it they were in the same locations as other units. Then I went to
the defending units and compared them with the other array and if they
are in the same location, put them together int he fightclass.

Okay -- for the moment, that's probably perfectly fine. One minor
refinement that might be helpful would be after you've done the above,
scan back through all the fightclass objects and simply remove any that
have an empty defender vector. That may not be the most elegant way to
produce the result, but at least that should be the result you're
looking for -- sometime later you can come up with a cleaner way of
getting the result.
 
J

JoeC

Jerry said:
[ ... ]
I didn't see anyone say in this thread "your question is off topic here go
away"
I merely suggested a better newsgroup.

I'm not even sure that's true -- at least to me, it seems that most of
the discussion has centered primarily around how to express his ideas in
C++, not so much about the fundamental design of the game itself. That
being the case, I think the questions are really more topical here --
though it may be that enough game programming is done in C++ that it
would also fit well there.

Not the whole game but this function. I am not a profetional
programmer but I do the best I can. I am wrighting this conbat routine
and it seems very complex and I have trouble keeping track of what I am
doing. That is not good. I am wondering if there is an easier way to
do what I am doing. I have been adding to my game, what I am doing
appears to work.
 
J

JoeC

Phlip said:
I never said you didn't. However...


Ultimately, the FAQ was written by a human, and it is not authoritative.

Ultimately, the term "ultimately" means "at the end of a chain of events".
So if I ask about the Class Factory Pattern, or MVC, or huffman compression,
and I ask in platform-neutral terms, then ultimately the exact answer will
depend on the C++ ISO Standard. Hence, by the criteria of the
non-authoritative FAQ, such questions are on-topic.

This newsgroup is not , and posters should always pick the
narrowest newsgroup for their questions. So questions that only relate to
interpretation of The Standard belong there, not here.

I am just trying to write a game and I feel I am getting a bit over my
head but still being a hard head, I make it work and push on. I am
sure what most programmers do. The only way to get experience is to do
it. It is one thing to learn syntax but it is another to apply good
programming practice in large projects. What my problem is is that I
program in a vaccuum, I don't have people with more experience letting
me know that there are better ways to do what I am trying to do. So
what I do is put this stuff out there and hope some one who has seen
simmiliar problems can give me some advice. I have already gotten some
good advice for other parts of the program. I would like to get this
part done then the game will be complete and I can expand it or go back
and improve the code so that it is easier to expand. If somone would
like to see the whole program (e-mail address removed) and I will send the
latests version program out.
 
N

Nathan Mates

Not the whole game but this function. I am not a profetional
programmer but I do the best I can. I am wrighting this conbat
routine and it seems very complex and I have trouble keeping track
of what I am doing. That is not good. I am wondering if there is
an easier way to do what I am doing. I have been adding to my game,
what I am doing appears to work.

Managing "who's where" is best separated from the "combat routine."
Most games have some code (and data structures) dedicated to keeping
track of units on the map. This means that units know where they are,
and the map manager also knows which units are where. If your map is
just a 2D array of a fixed, "reasonable" size (under a million
entries), you can just make a 2D array of entries, and each entry is
just a linked list of all units in that cell. If you're working on
PCs, just burn some memory upfront, and work on optimizing it later--
plenty of data structures exist to manage sparse datasets.

You should also separate out code to to map manager so that you can
say "find all units within radius R of point X,Y" and it'll go do that
for you. That's code for the map manager, not a "combat routine," as
you'll probably want your UI code to be able to do the same thing--
clicking on a square selects all units in that square, or shift-drag
selects all units in a user-specified rectangle. If you're going to
use the same kinds of functionality in multiple places, then break it
out to a common place.

Nathan Mates
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,744
Messages
2,569,482
Members
44,901
Latest member
Noble71S45

Latest Threads

Top