Hacker News Reader

Comment by debanjan16 | Hacker News Reader

Students learning Algorithms for the first time should never touch CLRS. It is the worst thing that a beginner can pick up. It is good as a reference text.

Beginners should start with this. This book "actually" teaches you how algorithms are designed and how to break down problems into chunks and solve them using induction/recursion.

Your mind will just be blown.

View on HN

What infuriated me about CLRS was that so much of the content was in the exercises - that had no answers you could check your results against. There are solutions guides you can look up now, but that they would publish such a hostile math book in the first place rubbed me the wrong way.

At some point Cormen had a page of shame where he would name people who asked him for solutions to some of the problems. I remember it was removed with a note that it was a bit much. Though, I suspect it was pointed out to him that this made him look bad.

I think this says a lot about the logic of the authors, though.

i frequently see him post on Quora--he appears as a pompous, abrasive and highly anti-social individual who acts like a gatekeeper to knowledge. He grudges the readers their knowledge by making it extremely boring and difficult to break into. Many 'authors' are like that. Which is feels like a conspiracy.

> so much of the content was in the exercises

I have this reaction to some Coursera courses (Odersky's Scala course did this to me). I mean it's wonderful if you get the "ah hah" to figure out the exercise, and it's probably the most meaningful learning, but it shouldn't be the ONLY way.

It's like working on some super expensive niche car; if you have the magic tool that fixes your issue it's trivial, if you don't it's almost impossible.

I wouldn't say that, it's just the audience is different. It's basically a math theory text for getting into academic CS.

I think the main key is the teacher. I followed the teacher in the class and the course was using CLRS, and I did not feel anything hard in homework and exams as long as I fully understood what was taught during the class time.

CLRS barely touches the intuition part and that makes it worst from a beginner's perspective.

> Students learning Algorithms for the first time should never touch CLRS. It is the worst thing that a beginner can pick up.

I can second that. CLRS is a great reference work for people who already know algorithms, and I can absolutely understand why CS Professors use it to prepare courses and exercises.

But as an introductory text for newcomers? No. Absolutely not.

And if not a beginner I would go straight to chapter 5 "Design of Algorithms by Induction". One of my favorite algorithms books.

Notably, this material is based on an earlier paper on the design of algorithms by induction, also by the author: http://akira.ruc.dk/~keld/teaching/CSS_e10/Manber88.pdf

I also recommend Jeff Erickson's book.

I didn't know what CLRS is. Seems to be "Introduction to Algorithms" by Cormen, Leiserson, Rivest and Stein (CLRS).

Maybe I'm different, but I used Kleinberg/Tardos and CLRS in my undergraduate algorithms class, and I preferred CLRS to KT and the other alternatives (Algorithm Design Manual, etc.), though KT was great too. I've heard from others that KT was better for them for learning how to actually design algorithms as well.

I had a similar experience. It may be that I have a math background and CLRS feels more like a math book than do other algo texts, but it was a breath of fresh air for me after looking at other popular texts that felt so much more handwavy. It felt so much more concise and the math arguments felt more compelling to me anyway. I mean, I am glad though that different texts exist because different people seem to benefit from different texts.

Agreed! I also recommend Grokking Algorithms for a friendly overview. There are also some great topic-specific books, such as Who's #1? and In Pursuit of the Traveling Salesman.

I loved CLRS! It was my first algorithms book in undergrad.

A beginner can and should pick it up if they love math and programming.

What about people who took a basic alg/ds class in the past but forgot most of it many years later?

I’m in this boat too, and would love suggestions to freshen up my algorithm skills.

Reply on news.ycombinator.com