## Recommended Posts

Hi Everyone, Im going off on one of my programming tangents and im wondering if anyone can help me figure this out so that i can return to actual work =P. Basically i have a little debug window courtesy of CEGUI and i want to be able to go
gui->Debug << "this is me showing i can type a number: " << 5 << "look i did it!, now for a float" << 0.5f;


so to get started i made myself a debug class and set up operator functions like the following:
void Debug::operator<< (float fl)
{
OutPutToDebug(ToString<float>(fl));
}
void Debug::operator<< (std::string str)
{
OutPutToDebug(str);
}


which did let me do the following: gui->Debug << "this is me showing i can type text: "; OR : gui->Debug << 5; then i went about trying to figure out the return type to allow me to do longer ones, so i tried a few but i figured the best option was either std::stringstream or simply adding up strings. this was my atempt at adding up strings:
std::string Debug::operator<<( float fl )
{
std::string line+=XF::Utilities::ToString<float>(fl);
return line;
}

std::string Debug::operator<<( std::string fl )
{
std::string line+=fl;
return line;
}


and that gave me a whole host of errors like this: Error 2 error C2784: 'std::basic_ostream<_Elem,_Traits> &std::operator <<(std::basic_ostream<_Elem,_Traits> &,const std::_Smanip<_Arg> &)' : could not deduce template argument for 'std::basic_ostream<_Elem,_Traits> &' from 'std::string' c:\documents and settings\chris\desktop\agt\code\project starblazer\gamescreens\menutesting\menutestscreenbuttons.cpp 20 and my atempt at stringstream std::stringstream Debug::operator<< (std::stringstream fl) { return fl; } gave me the errors: Error 3 error C2248: 'std::basic_ios<_Elem,_Traits>::basic_ios' : cannot access private member declared in class 'std::basic_ios<_Elem,_Traits>' c:\program files\microsoft visual studio 9.0\vc\include\sstream 516 im not really sure if im choosing the wrong return type, the wrong structure of the function definition (should i be using const?) or if its simply not possible! Can anyone help me on this one it would be much appreciated! Thanks, Chris

##### Share on other sites
you have to return a referance to the output stream not a std::string
in your case you might want to use std::ostringstream. the general look of the function you are trying to create goes a little something like this.

std::ostream& operator << (std::ostream& os, const double x){	// -- perform output operations here	return os }

the referance to the output stream allows this to happen:
os << someval1 << someval2;

otherwise you would only be able to do this:
os << someval1;
os << someval2;

which is generally not what you would want since the whole point of the '<<' and '>>' operators is so you can 'stream' data

##### Share on other sites
void Debug::operator<< (float fl){    OutPutToDebug(ToString<float>(fl));}void Debug::operator<< (std::string str){    OutPutToDebug(str);}

Should become
Debug& Debug::operator<< (float fl){    OutPutToDebug(ToString<float>(fl));    return *this;}Debug& Debug::operator<< (std::string str){    OutPutToDebug(str);    return *this;}

##### Share on other sites
Wow That was quick thank you Steve132 and CodeCriminal, but im afraid each solution returned an error.

Steve132s solution produced the compile error: Error 15 error C2296: '<<' : illegal, left operand has type 'Debug *' c:\documents and settings\chris\desktop\agt\code\project starblazer\gamescreens\menutesting\menutestscreenbuttons.cpp 20

(although i think i have a solution...ill post if im right EDIT: Nope =[)

on the line:
bool MenuTestScreen::UpgradeTower(const CEGUI::EventArgs &args){		mGUI->mDebug << "Hello " << 5; //here	return true;}

but im assuming thats because Debug is not derrived from any of the stream classes???

and CodeCriminal soltion produced this compile error: "Error 22 error C2804: binary 'operator <<' has too many parameters c:\documents and settings\chris\desktop\agt\code\project starblazer\menu\debug.h 44"

this solution ran without compile errors:

std::stringstream& operator << (std::string str)		{			streamLine << str;			OutPutToDebug(streamLine.str());			return streamLine;		}		std::stringstream& operator << (float fl)		{			streamLine << fl;			OutPutToDebug(streamLine.str());			return streamLine;		}

but doesnt work logically, which i thought it might not, OutPutToDebug is inserting a new item into a list box, so in the case of debug << "hello" << 5;
we get a list box item saying hello, then the next it will say hello5 hello, and keep adding on.

is there anyway i can know that its the last input, so i would know its just entered debug, and i can clear the string stream and output the whole line?
as it stands here is all my code (each persons version was ran on its own, other versions were commented out for each build):

/*************************************Author: Christopher Pepper 11/12/2009***************************************/#pragma once#include <ostream>#include "CEGUI/CEGUIWindowManager.h"#include <CEGUI/CEGUI.h>#include "OgreVector3.h"#include <iosfwd>#define DEFAULT_CEGUI_SKIN "TaharezLook"#define DEFAULT_CEGUI_FONT "BlueHighway-10.font"class Debug{	private:		bool mInverted;		bool mEnabled;		int mKeepThreshold;		CEGUI::Listbox* mDebugBox;						CEGUI::WindowManager *mWindowManager;		std::stringstream streamLine;public:			//My updated version		std::stringstream& operator << (std::string str)		{			streamLine << str;			OutPutToDebug(streamLine.str());			return streamLine;		}		std::stringstream& operator << (float fl)		{			streamLine << fl;			OutPutToDebug(streamLine.str());			return streamLine;		}	//codecriminals version 		std::ostream& operator << (std::ostream& os, const double x)		{			// -- perform output operations here			std::stringstream streamLine;			streamLine << os;			OutPutToDebug(streamLine.str());			return os; 		}	//Steve132s version		Debug* operator << (std::string str)		{			OutPutToDebug(str);			return this;		}		Debug*operator << (float fl)		{			OutPutToDebug(XF::Utilities::ToString<float>(fl));			return this;		}						void DisableDebug(){mWindowManager->getWindow("Debug")->setEnabled(false);};		void EnableDebug(){mWindowManager->getWindow("Debug")->setEnabled(true);};		void HideDebug(){mWindowManager->getWindow("Debug")->setVisible(false);};		void ShowDebug(){mWindowManager->getWindow("Debug")->setVisible(true);};		void Initialize(std::string skin = DEFAULT_CEGUI_SKIN, std::string WindowTag = "Root");		void OutPutToDebug(std::string output);		/*How many items will the debug store in the debug listbox (please note that the app		will slow down at high numbers)*/		void SetKeepThreshold(float number){mKeepThreshold=number;};#ifdef _DEBUG		Debug() : mInverted(true) , mEnabled(true) , mKeepThreshold(300){};#elif		Debug() : mInverted(false) , mEnabled(false), mKeepThreshold(0){};#endif		~Debug();};

.cpp :

#include "Debug.h"#include "Utilities.h"#include "Logger.h"#include "OgreVector3.h"void Debug::OutPutToDebug( std::string output ){	if(!mDebugBox)	{		XF::Logger::GetInstance()->Log("Error Debug Not Initialized");		throw "Error Debug Not Initialized";	}	if(mEnabled)	{		if(mDebugBox->getItemCount() > mKeepThreshold)		{			if(mInverted)			{				mDebugBox->getListboxItemFromIndex(mKeepThreshold)->setAutoDeleted(true);				mDebugBox->removeItem(mDebugBox->getListboxItemFromIndex(mKeepThreshold));						}						else			{				mDebugBox->removeItem(0);				mDebugBox->getListboxItemFromIndex(0)->setAutoDeleted(true);				mDebugBox->removeItem(mDebugBox->getListboxItemFromIndex(0));			}		}		CEGUI::ListboxTextItem* lItem = new CEGUI::ListboxTextItem(output);		if(mInverted)			mDebugBox->insertItem(lItem,0);		else			mDebugBox->addItem(lItem);	}}void Debug::Initialize(std::string skin , std::string WindowTag ){	mWindowManager = NULL;	std::string buttonSkinType =skin/*+"skin"*/+"/Listbox";	mWindowManager =  CEGUI::WindowManager::getSingletonPtr();	mDebugBox = (CEGUI::Listbox*)CEGUI::WindowManager::getSingletonPtr()->createWindow(buttonSkinType, "Debug");	mDebugBox->setSize(CEGUI::UVector2(CEGUI::UDim(1, 0), CEGUI::UDim(0.15, 0)));	mDebugBox->setPosition(CEGUI::UVector2(CEGUI::UDim(0, 0), CEGUI::UDim(0,0)));	mDebugBox->setAlpha(0.5);	mWindowManager->getWindow(WindowTag)->addChildWindow("Debug");}Debug::~Debug(){	mWindowManager->destroyWindow("Debug");}

##### Share on other sites
You want to return a reference (Debug &), not a pointer (Debug *).

##### Share on other sites
Quote:
 Wow That was quick thank you Steve132 and CodeCriminal, but im afraid each solution returned an error.Steve132s solution produced the compile error: Error 15 error C2296: '<<' : illegal, left operand has type 'Debug *' c:\documents and settings\chris\desktop\agt\code\project

They should be non-member functions.

##### Share on other sites
Thank you jpetrie and SiCrane

Debug& operator << (std::string str){	OutPutToDebug(str);	return *this;}Debug&operator << (float fl){	OutPutToDebug(XF::Utilities::ToString<float>(fl));	return *this;}

the code above compiled and ran! but there is still the problem of the logical error of OutPutToDebug, as it posts to a list box via items (at the moment) so "hello" << 5;
will print

hello
5

when i would prefer it on one line.

is there anyway i can tell when the stream has completed? so that i can then print the full line? or when its first called so that i can tell it to start a new line?

or is this not actually possible? and i should take more of a buffered aproach, print if line is not empty on the draw loop type thing

Thanks for all the support guys!

##### Share on other sites
Yey I got the solution i wanted, i now have a very nice litle debug working with CEGUI list box, probably not the best solution in the world but it'l do i updated my header code to have the following:

private:*--detail omitted--*               std::stringstream streamLine;               bool mStringEmpty;public:*--detail omitted--*                Debug& operator << (std::string str)		{			streamLine << str;			mStringEmpty = false;			return *this;		}		Debug&operator << (float fl)		{						streamLine << fl;			mStringEmpty = false;			return *this;		}

and added the following to be ran every frame:
void Debug::Update(){	if(!mStringEmpty)	{		OutPutToDebug(streamLine.str());		streamLine.str("");		mStringEmpty = true;	}}

im not to keen on the every frame bit or the face im not really using the stream operator itself for what its meant for..but it does produce the desired result! if anyone can offer improvements it would be much appreciated!

is there anyway to know if its the first/laststream? my next challenge is to make it so
mDebug << "hello " << 5;
mDebug << "hello im on a new line " << 4;
prints:
hello 5
hello im on a new line 4
so the first/last thing would still be usefull, but im sure ill thing of something!

*EDIT: For now im happy to use mDebug << "hello " << 5 << D_ENDL;
D_ENDL being an int (98765) that i hopefully wont ever want to print :P
with this change to the float version:
Debug&operator << (float fl)
{
mStringEmpty = false;

if((int)fl == D_ENDL)
{
Update();
}
else
{
streamLine << fl;
}

return *this;
}
*

Thanks to everyone who posted! I would not have got to this solution without u guys but even more importantly you taught me stuff about streams that i can use again in the future to better effect! i haz left u alz good feedback =]

[Edited by - ChrisPepper1989 on February 18, 2010 6:15:52 PM]

##### Share on other sites
You really should try to understand what you're doing. :)

When you write something like std::cout << {value of type T}, it calls an overload for the operator<<. Notice the types of the arguments: operator<< always requires exactly two arguments (because it's a binary operator). We can implement the overload either as a free function with two parameters (left-hand side, then right-hand side), or as a member function with one (the this-object represents the left-hand side, and the single parameter represents the right-hand side).

Of course, an operator overload is really just a function with a special name that is called in a special way. As such, it has a return value as well. If you write something like foo << bar << baz, the operators are applied left-to-right (operator<< has left-to-right associativity), so foo << bar gets evaluated first. Thus, the whole thing gets translated into operator<<(operator<<(foo, bar), baz), or alternatively foo.operator<<(bar).operator<<(baz), or some combination. :)

For that to work, the first application has to return something that will work with the second application. The way std::ostream (the type of the std::cout object) handles this is to return another std::ostream. That way, you can chain together as much as you'd like: each application returns a std::ostream that's suitable for the next application of the operator. Of course, it can't just return any old ostream; it has to return itself (because you want all the output to go to the same place). It can't even return a copy of itself (ostreams can't even be copied, in the first place), so it returns itself, by reference.

In your case, you want to chain with the Debug object, so you return a Debug&, via 'return *this', as in Steve132's example.

As for the question of how to put multiple things on the same output line, you have some options.

1) You can modify OutPutToDebug() so that it looks for newlines in the strings being logged and breaks up the text accordingly, and otherwise appends the new text to the current line.

2) You can modify OutPutToDebug() so that it always appends the new text to the current line, and then special-case operator<< for std::endl such that it goes to the next line in the GUI display. (Hint: the type of std::endl is std::ostream&(*)(std::ostream&). But it's not the only object of that type, so you'll want to check for equality as well.) This is kind of a hack, though.

3) My preferred solution: you can separate out the concept of a debug message from a debug display. That is, make a separate class called DebugMessage that holds the std::stringstream instance, in which you accumulate the text with operator<< overloads. Tell it about a corresponding Debug instance when you create it, and in the DebugMessage destructor, you have it "flush" to the Debug instance by caling OutPutToDebug().

That looks like:

// Debug has no operator overloads, but still has OutPutToDebug as beforestruct DebugMessage {  std::stringstream buffer;  Debug* display;  DebugMessage(Debug& debug): display(&debug) {}  template <typename T>  DebugMessage& operator<<(const T& t) { buffer << t; return *this; }  ~DebugMessage() { display->OutPutToDebug(buffer.str()); }};// Used like:DebugMessage(gui->debug) << "test: " << 5 << " is bigger than " << 4.2;// See that? We construct an unnamed instance of DebugMessage, use operator// chaining to build up the message, and then at the end of the statement the// instance falls out of scope, so the destructor is called, triggering output.// This only works with the member version of the operator<< overload,// because it relies on modifying a temporary value. Otherwise you'd have to// declare a variable to hold an instance of DebugMessage, and manage its// lifetime, e.g.// {// DebugMessage dm(gui->debug);// dm << "test: " << 5 << " is bigger than " << 4.2;// } <-- note the braces!

This technique exhibits several interesting ideas for good design, and also avoids separately converting each item with ToString(). (By the way, I hope you didn't write that yourself; you can just use boost::lexical_cast if you need it.)

[Edited by - Zahlman on February 18, 2010 7:55:14 PM]

##### Share on other sites
Quote:
 Original post by ZahlmanYou really should try to understand what you're doing. :)

indeed that is always helpful lol!

Quote:
 Original post by ZahlmanWhen you write something like std::cout << {value of type T}, it calls an overload for the operator<< . Notice the types of the arguments: operator<<; always requires exactly two arguments (because it's a binary operator). We can implement the overload either as a free function with two parameters (left-hand side, then right-hand side), or as a member function with one (the this-object represents the left-hand side, and the single parameter represents the right-hand side).

riiight this actually helps alot, ive been messing around with overloading operators before and always when the 2 arguments was appropriate, that is a major penny drop! and a bit of a DOI!

Quote:
 Original post by ZahlmanIf you write something like foo << bar << baz, the operators are applied left-to-right (operator<< has left-to-right associativity), so <

Ok that clears all that up, i was stabbing in the dark a little, i should have observed the logical! thats really helpful

Quote:
 Original post by Zahlman1) You can modify OutPutToDebug() so that it looks for newlines in the strings being logged and breaks up the text accordingly, and otherwise appends the new text to the current line. 2) You can modify OutPutToDebug() so that it always appends the new text to the current line, and then special-case operator<< for std::endl such that it goes to the next line in the GUI display. (Hint: the type of std::endl is std::ostream&(*)(std::ostream&). But it's not the only object of that type, so you'll want to check for equality as well.) This is kind of a hack, though.

this is what i was sort of going for...especiatly 2 with more so of a hack but...

Quote:
 Original post by Zahlman// Debug has no operator overloads, but still has OutPutToDebug as beforestruct DebugMessage { std::stringstream buffer; Debug* display; DebugMessage(Debug& debug): display(&debug) {} template DebugMessage& operator<<(const T& t) { buffer << t; return *this; } ~DebugMessage() { display->OutPutToDebug(buffer.str()); }};// Used like:DebugMessage(gui->debug) << "test: " << 5 << " is bigger than " << 4.2;// See that? We construct an unnamed instance of DebugMessage, use operator// chaining to build up the message, and then at the end of the statement the// instance falls out of scope, so the destructor is called, triggering output.This technique exhibits several interesting ideas for good design.

I really like this, its neat, interfaced, re-usable and just downright clever, I assume then when things are cleaned up they are cleaned up in order so if i was to go:

void main(){  int a;  int b;  int c;  DebugMessage bob(gui) << "hello";  DebugMessage fred(gui) << "again";//under the hood  clean up a;  clean up b;  clean up c;  delete bob;  delete fred;}

cause if that isnt the case its the only thing that i can see that would stop this from working...

Quote:
 and also avoids separately converting each item with ToString(). (By the way, I hope you didn't write that yourself; you can just use boost::lexical_cast if you need it.)

Well its actually a joint project and my code collegue wrote the ToString function(it uses stringstream)...2bh I generally avoid boost, not out of snubbery but because up untill now my university has frowned on using external stuff, and generally marks more when they are avoided (which is slightly annoying, as gamedev generally teaches me to use the best tools available and dont re-invent the wheel), plus im pretty sure it would be a nightmare as their computers have deep freeze and stuff so re-setting-up and stuff...but otherwise i would probably by now have alot of helper libraries like boost set up as i've had my fare share of writing home rolled: containers, multi-array, conversions...and so on.

Thank you so much for this detailed explanation i feel this is a post i will re-visit many times more to brush up on that particular operator, i may even post up the whole experience on my website in the form of a small tutorial..

Thanks Again, Chris

p.s. I just lost the game thanks to ur signature =[ lol although i do love it =]

[Edited by - Zahlman on February 19, 2010 11:24:39 AM]

##### Share on other sites
Quote:
 Original post by ChrisPepper1989I assume then when things are cleaned up they are cleaned up in order so if i was to go:

Variables and temporary variables within a scope are cleaned up in reverse order: last created is first cleaned up. (There are good reasons for this, and that principle is applied pretty consistently across the language.) However, the scope of a temporary does not extend beyond the statement that it's used in - it can't, because it doesn't have a name and thus can't be used in another statement (how would you call it?).

Notice that in my code sample, I don't write 'DebugMessage bob(gui)'; I write 'DebugMessage(gui)'. This is important. The latter creates a DebugMessage but does not assign it to a variable.

The bit at the end was about an oddity in C++; apparently, this technique works with member functions (the surprise, even though I've done this before and not really thought about it), but not with free functions (which I could have already figured out). If you write the operator overload as a free function, then you won't be able to call it with a temporary. (Actually, I suppose it could be worked around, but now I'm really getting into the murky depths of C++...)

That's why I showed the other examples with braces. The point of the braces is to forcibly create another scope for two statements: the declaration of a DebugMessage variable, and then its use. At the closing brace, only the DebugMessage instance is out of scope, so only it gets cleaned up - so the reverse-order thing doesn't matter. :)

Quote:
 I generally avoid boost, not out of snubbery but because up untill now my university has frowned on using external stuff, and generally marks more when they are avoided (which is slightly annoying, as gamedev generally teaches me to use the best tools available and dont re-invent the wheel)

But they're letting you use CEGUI? ;)

(P.S. You were missing a quote end-tag, so I fixed that for you.)

##### Share on other sites
Thanks for clearing that up, that will really be helpfull knowledge now and in the future! Sincerly thank you!

and thanks for fixing the quote tag =]

As for the boost/cegui thing its more of an old habbit plus unsurity, for every other assignment we wern't allowed to use external libraries and only for this assigment are they letting us do pretty much whatever we like.

The only requirement of the project is to use Ogre and so as a result Ogre and therefore CEGUI is already set up on the university computers which makes it easy enough to integrate and not have any problems.

Any other library we try and integrate almost certainly ends up being alot of hassle to work both at home and on the university computers! im sure we could do it (i.e. just taking the boost library with us on a usb key) but at this point we already have accumulated enough of our own helper code (like the toString thing) that it just seems easier to use our own code then try and set up Boost on the university computers (plus have it compile and run on the day we present it, in another room on another computer =]).

It doesnt make a whole lot of sense i know but i do honestly believe that for this project we are best off leaving it out for now.

Again i want to say a big thank you i will be switching the code from my slighly hacky set up to your nice clean version soon enough!

##### Share on other sites
Just out of curiosity, which university are you at? I was considering going to lancashire to do a course that focuses on computer game programming next year. but its always good to see what else is out there.

Thanks

##### Share on other sites
It might be more than you were looking for, but scrolling down in this old thread might be beneficial.

It sounds like what you need is to replace the streambuf (the facility used by the language runtime to move formatted stream data to its sink) rather than replace the iostream (the part used to format stream data). A little bit of boilerplate code and using std::clog to log to a debug window will instantly win the respect and admiration of your peers and copious quantities of free liquor in bars. Except for the latter. Usually.

##### Share on other sites
@Bregma thank you for the link, it looks very interesting, i am very happy with Zahlmans solution and at the moment im going with that.

@CodeCriminal, i looked at lancashire when i was looking at universities and it seemed pretty good, i dont remember it having a game programming course at the time, it had a good game design, im guesing the programming one is new? I go to Bolton university and i sort have mixed feelings, i have enjoyed most of my time here, it has been alot of fun, the univeristy is relatively small so u get to know everyone and you will have a very friendly relationship with the lecturers who are very down to earth. However if im honest i wouldnt necessarily reccomend it, especially to anyone with prior coding knowledge, its great if you dont know anything and want to start from scratch, but if not its very possible you wont actually learn anything. I came from a college course with an extremly good lecturer so when i came to bolton i was bored for 2 years and only now in my third year do i feel like im pushing myself, and if im honest those 2 years made me lazy. The fact i have to ask how to use stream operators properly after 2 years at uni worrys me but if im honest, the gerneral reputation of computer games programming courses (in the UK anyway) makes me feel its the same across the board.

The shining star of the university really is the lecturers, if you do want to learn more and if your having trouble they have the time to spare, they really are fantastic and Phil Carlisle is one of those, hes ex-industry (team-17) a speaker at GDC and he does really know his stuff, and i think its hard to find anyone of that calibre and i would be tempted to recomend the uni on him alone!

Its also kind of helpful that they cover maths and physics but if im honest they could do without the 3 modules on them, They are sort of taught very academically and truth be told you sit there thinking ok its lovely that you just spent 3 lessons explaining how mr so and so derrived this formula but in reality we could just plug it in and it work, or alternatively use a physics engine and get used to using third party stuff. An overview is very helpfull and you need to know the basics but some of the stuff is really unessesary, or at least it shouldnt be marked...plus the guy teaching us Astar had no idea what it really was and got us to do it by hand (well he thinks i did it by hand :P, good thing he wanted it in Excel hehe)

Lancashire sounds pretty good and knowing what i know now id either go with them or a computer science degree/Software enginering degree (theres a pretty good CS one at salford and they have some pretty awesome game stuff there like the CAVE room) however all in all it doesnt really matter, go somewere that feels right and u think ull have fun at cause u only get uni once and you may aswell try to enjoy it and then work your arse off in the third year for that first =P. In reality proving you can do the three years well and providing a very good portfolio is what you need, what you learn on the actuall course is nothing compared to what you can learn browsing game dev, playing with your own code, self teaching, talking to peers and your first year in industry (from what i hear and imagine) and i get the impression your already doing all of that (apart from the industry bit)!

Go to an open day and see if you can see yourself being there, meet a few lecturers, make sure they are approachable, have a look at the halls of residence and good luck, the right university or not im sure youl do fine! p.s. first year halls here was fantastic, then some other guy took over and now they are all "strict" and it was very stressfull dealing with them in the second year!

Let me know how it goes!

Rough overview of Bolton course:

1st year:
the whole first year covers the building blocks, if,switch, for,while, classes, OO (each with its own 1 hour lecture....really could have bashed my head against a wall as im sure u no, it doesnt need that)

theres also a module for maths and a module on physics

the year finishes with the requirement to make a small 2d shoot up kind of game (like space invaders but not as...fun...complex...interesting) but u are allowed to go off on a tangent and make it a bit more interesting

2nd year:
GHAP was cool we had to write a game for XNA but it was a little bit spoon fed, this is actually because the lecturer was smart and realised most of the class would struggle to pass without it, but that in itself is slighly annoying

theres a module on mathamical physics, which covers rotational physics mainly, torque and angular velocity and you just further learn to hate the fact that everything is in symbols and its not always easy to find out what the symbols mean, but the lecturer is pretty cool and you write a little thing in flash...and learn to both hate (for its code & debugging) and love that (for the fact it is actually pretty easy to just throw something together).

Theres also an incredibly pointless module on Career development but i think thats getting reworked to be a bit more usefull, its mostly the old "this is how you write a C.V for the games industry" but there was a load on setting up ur own business which was pointless to anyone who didnt want to set up there own business...or even to those who did if im honest

in the second semester you right a 3D game that is your first real challenge, but its made annoying by the fact you have to write the physics for it and the game never look how you wanted it to cause physics will only cause you pain and wasted time...then again im not a big physics programmer (u may have guessed im not a huge fan of physics)

The final year is huge, a double award project module that runs the entire year, a module that requires u to write a pretty big game and another module on writing C.Vs, developing your portfolio exept this one is usefull but finishes with an exam that you really could do without with the major work load :p

that pretty much sums it up, like i say though it does have lecturers that know what there talking about, my project mentor is a cool guy who wants to do a pHD with me on input stuff and so his office is always full of interesting toys on input and phill and him are great guys to talk to about the games industry as they are both so pasionate about it and you learn so much simply from talking to them. If i do decide to stay and you come to it, ill say hello, i may also rope you in for some android development im planning on doing next year lol

## Create an account

Register a new account

• ### Forum Statistics

• Total Topics
628381
• Total Posts
2982360

• 10
• 9
• 15
• 24
• 11