Linked list implementation question

I am preparing for job interviews and was going over the implementation of a linked list. specifically appending to the end of the list.

when I initially wrote it out I had a public variables that held a pointer to the first node, and the last. so adding was simple, take the last node->next = newNode and the last node = newNode.

badabing bada boom.

i was looking up other peoples implementations, and every single one's(including wikipedia, which as we all know is always right) did not have a last node variable, but instead looped through the list until the tempNode->next ==NULL, and then that is your last one.

Logically the latter implementation makes sense but I cant see a reason for doing it. the small space the last pointer takes up, and the extra code to read and write to it, is still far less than having to loop through a list of 1000+ elements.

So, I figure I'm misunderstanding something. any help is appreciated.

Like you, I subscribe to the "also keep the last pointer" camp, if entries are to indeed be entered at the end. Otherwise, I just keep the anchor and add to the front.

badabing, bada boom.

It seems like the most logical approach.. but I have seen to many loop through the whole list approaches to disregard it. any idea why they do it?

They don't care about performance?

What kind of job interview are you going to that you worry about linked lists?!?
Anyways, in C++, the standard says that for any container you have access to the end() in amortized constant time. So if you implement it in C++, do everyone a favor and use the readily-available stl version. And if you implement your own container, adhere to the parts of the standard which are applicable. This way, the pain you do to your fellow programmers when using your code will be minimized (argh. libxml++ implemented size() in O(n). Bastards. I spent two days looking for that "bug".)