Enigmatic Comments Re: lab Supplementals..

ZWEI_VIER_ZWEI

Right, a few days back I was in a lecture, and the lecturer was going on about labs, and passing labs, and passing exams, and all that kind of carry-on. And anyway he goes on to mention supplemental exams, and he said something along the lines of "And there are no supplemental lab exams, for reasons which I'm not going to get into, but which reflect quite poorly on the student body"...

Does anyone know what he is talking about? I'm studying Computah Science if that's any help...

cuckoo

hmmm, maybe something to do with plagarism? do you have actual sit down in front of a computer and program it exams in comp sci?

in science we've got to do our labs during the year, or we can be NSed and aren't allowed sit the annual exams.

Crash

i think it might be to do with something along the lines of "yeh cant trust students in a computer room" - if you wouldnt mind telling me which lecturer (either in post or pm) i'd be intrigued.

its a disgrace that in comp sci you have to write programs on paper as opposed to having a practical element to end of year exams - something i'd love to campaign to change but they'd actually have to set up a special lab for it etc. - on a whole the comp sci end of years are a bit ****ed in that respect - they end up easier because they are more forgiving of mistakes, but at the same time, having to write programs is a dirty whore.

Chardee MacDennis

but in fairness its all crap, the exams should be on techniques not the actual writing of a program, do you ever think an emplyer is going to throw you into the room and say "write a FIFO stack from scratch"? i mean come on....

thats all my CS exams have ever been, write this code from scratch, when in the real world we will have resources to use, a good programmer uses all resources available to him/her surely.

Nietzschean

i completely agree, its a bit stupid really, i'd say i know possibly a dozen languages, but i wouldn't be able to write a FIFO stack in all of them without having a glance at some reference materials and having a good debugger....

but then there are guidlines that only something like 20% can be from coursework isn't it? so everything being project based as would be the best imo, or an exam run over like 6 hrs with access to reference materials to produce some desired output...but neither are likely to happen any time soon...

bbrazil

Call_me_al wrote:

do you ever think an emplyer is going to throw you into the room and say "write a FIFO stack from scratch"? i mean come on....

Yes.

It was deletion from a doubly linked list in a job interview, which is a bit more difficult.

Chardee MacDennis

feck, thank god i dont intend going into IT!!

Nietzschean

bbrazil wrote:

It was deletion from a doubly linked list in a job interview, which is a bit more difficult.

Ouch, well that sounds like fun

bbrazil

Nietzschean wrote:

Ouch, well that sounds like fun

Well, I got the job and enjoyed 12 weeks in the states. Lack of a header node and that the list wasn't circular made it a lot more difficult then it should have been (5 lines + search code). Ended up with lots of ifs.

Nietzschean

just curious since i've not done any programming course and can't be arse'd starting a compiler is :

double ** my_stack;
int stack_len =-1;
int start_pos,current_pos;
int stack_count;
int get_item(double & ans){ // Code updated to account for -1 value in stack (bbrazil's noticing skillz ;) )
	if(start_pos == current_pos || stack_len == -1)
	return -1; // Stack empty or not setup yet
	ans = my_stack[start_pos];
	start_pos++;
	if(start_pos == stack_len)
		start_pos = 0;
	--stack_count;
        return 0; // return 0 for success
}
void set_stack(int len) {
	stack_len = len;
	start_pos = 0;
	current_pos = 0;
	stack_count = 0;
	double * tmp = malloc(stack_len * size_of(double));
	my_stack = &tmp;
}
void add_item(double item) {
	if(stack_count == stack_len)
		return;
	my_stack[current_pos] = item;
	current_pos++;
	if(current_pos == stack_len)
		current_pos = 0;
	++stack_count;
}

correct?

bbrazil

Nietzschean wrote:

just curious since i've not done any programming course and can't be arse'd starting a compiler is :

That looks like a ring buffer which is a queue (FIFO) rather than a stack (FILO). Maybe Call_me_al meant a more general deque?

What I'd do for a simple vector based stack is:

int stack[1024];
void push(int i){*(stack++)=i;}
int pop(void){return *(--stack);}

Here's a more general implementation using a linked list:

typedef struct node_{struct node_ *next;int value;} node; node *head=0;
void push(int i){node *t=malloc(sizeof(node));t->value=i;t->next=head;head=t};
int pop(void){node *t=head;int i=t->value;head=t->next;free(t);return i;};

These are untested.

Nietzschean

nice n compact code

rsynnott

crash_000 wrote:

its a disgrace that in comp sci you have to write programs on paper as opposed to having a practical element to end of year exams - something i'd love to campaign to change but they'd actually have to set up a special lab for it etc. - on a whole the comp sci end of years are a bit ****ed in that respect - they end up easier because they are more forgiving of mistakes, but at the same time, having to write programs is a dirty whore.

There's an argument for it. Because you can't test as you go along,or, practically speaking, implement in stages, it requires you to think about the problem slightly differently. And because of this, they are quite forgiving of mistakes.

Call_me_al wrote:

but in fairness its all crap, the exams should be on techniques not the actual writing of a program, do you ever think an emplyer is going to throw you into the room and say "write a FIFO stack from scratch"? i mean come on....

It's not a vocational course; the primary purpose is NOT to teach you exactly what you'd be doing in a job.