Importance of Algorithmics

This semester I had a course on analysis of algorithms, so i decided to try out a few algorithm based programming contests and sites..specifically www.topcoder.com. The problem is that though i get the logic and generally the method to solve the problems, i usually end up missing some corner case or there is an overflow somewhere, or sometimes i simply misinterpret the problem statement..i havent trained for algorithmics as such, just participated in the topcoder competitions and the local ACM ICPC competition and the like. Could you tell me how important is algorithmics when one goes in for a a job interview or after that..i'd like to know this especially from people here who handle interviews of freshers in their companies and those who have gone through interviews :) Is it worth spending time developing my algorithm based skills or should i go about learning different languages and get involved in projects for designing softwares and all ?

I'd probably say it depends what you want to do.

Quote:

---

I'd probably say it depends what you want to do.

---

I'm asking in general..i guess interviews for freshers are pretty much the same (atleast thats the trend on this side of the globe) Right now i havent really thought what i want to do..have to decide that too :@ damn..

>i havent trained for algorithmics as such, just participated in the topcoder competitions and
>the local ACM ICPC competition and the like.
Those are two different things. Competition coding is really nothing more than knowing the tricks and putting them together quickly. It's a whole different process than professional development.

>Is it worth spending time developing my algorithm based skills or should i go about learning
>different languages and get involved in projects for designing softwares and all ?
Algorithms are fundamental to programming. When I conduct interviews, I spend quite a bit of time going over algorithm theory and problems. If you don't do well there, I'm extremely likely to mark you as a no-hire. Languages aren't a huge deal. The only time I expect proficiency with a language is when the job requires it (compiler implementation, for example). By all means get involved in projects. The only way to get good at developing software is to do it.

Quote:

---

Algorithms are fundamental to programming. When I conduct interviews, I spend quite a bit of time going over algorithm theory and problems.

---

Is intricate stuff related to graphs and never heard before algorithms included in that? Most of the problems that i dont end up getting are in some way or the other related to mathematical relations or to number theory about which i have no idea..are things like number theory and advanced graph theorems that important ?

You should probably know all the existing algorithms, even if you can't implement them. You should be able to understand (very good) the algorithms of linear data structures, trees, graphs and dynamic programming. Examples of graph algorithms would be shortest path, spanning trees, coloring, node search etc.

As for how often you use them out there in the wide world of industrial programming ... I have used them 0 times until now.

>Is intricate stuff related to graphs and never heard before algorithms included in that?
If the job requires it. Generally, I'll concentrate more on the concepts behind the algorithms rather than anything specific. Any code monkey can memorize an algorithm, but a good programmer can tell you how it works, why it works, and use the lessons learned to invent other algorithms. Of course, I would expect you to have heard of the common ones and be able to describe them. Otherwise I'd wonder why you think you're a programmer.

>are things like number theory and advanced graph theorems that important ?
The concepts behind them are important. Let's take sorting as an easy example. If you memorize quicksort, all you have is a sorting algorithm. But if you understand the divide and conquer strategy that quicksort is based off of, you can use it to come up with any number of incredibly useful but unrelated algorithms.

Thanks for the info :)

i think topcoder is a lot of fun :) it really helps to develop valuable skills any programmer can find useful one day. not sure about advanced graph theorems and such, those things you should know when the job requires it. another thing with contest programming is that it's time oriented. i don't think thant in real programming one develops an algorithm in a couple of hours. it all takes lots of testing and time, something you are short of during contests.

@viaxd
Whats your handle on topcoder ? mine is thepenguin. Yes, i usually find myself short of time while participating on topcoder..or i end up making really dumb mistakes. :(

well, with a rating of < 800 im a little embarassed. it just takes a lot of practice i guess.

haha :P well, yea, its all about practise..i ll practise this vacation if i get time from my other commitments, would be blue if i get over these stupid bugs :P

I think algorithm contests are good for interview preparation and for fun.

Quote:

---

Originally Posted by Prelude

>Is intricate stuff related to graphs and never heard before algorithms included in that?
If the job requires it. Generally, I'll concentrate more on the concepts behind the algorithms rather than anything specific. Any code monkey can memorize an algorithm, but a good programmer can tell you how it works, why it works, and use the lessons learned to invent other algorithms. Of course, I would expect you to have heard of the common ones and be able to describe them. Otherwise I'd wonder why you think you're a programmer.

>are things like number theory and advanced graph theorems that important ?
The concepts behind them are important. Let's take sorting as an easy example. If you memorize quicksort, all you have is a sorting algorithm. But if you understand the divide and conquer strategy that quicksort is based off of, you can use it to come up with any number of incredibly useful but unrelated algorithms.

---

I've long since forgotten the specifics of things like quick sort and height-balancing a binary tree, etc, and I do have some bad habits (I basically pick-and-choose my favourite aspects of C and C++) but I've got a good programmer's mindset that lets me solve most problems with logic and reason (if I have a good think and maybe draw some diagrams) and I can get an intuitive grasp of an algorithm.

You could say I've got a good general programming aptitude but lack formal training (I do have a qualification in computing, but I've forgotten a lot of the theory).