The best skill to have when proving that you have the C++ skills is to answer any question about almost anything cold, without notes. A person, whether in person or more likely by phone, will ask a simple question, and then progressively make the questions more difficult. The ONLY way to prepare for such an onslaught of questions is to code, code, and when finished coding, code some more.I have been coding for over 20 years, starting with Fortran and now can easily converse and code in C++, Java, and C#. The language is not the issue. That is a given. The issue is how to quickly recognize how to abstract the question (and eventually a quant model) into a software framework. This abstracting has to be efficient, reliable, scalable, and must solve the problem.The best analogy I can give is that almost anyone with enough time and effort can learn a spoken language, and the syntax to computer languages and the technologies for distributed computing, are in my humble opinion, easier. However, after a very good grasp of the spoken can you write like Shakespeare or Dante or Goethe?That's the issue, not if you can code, but can you develop software that solves the problem a virtuso manner. The ONLY way to acquire this knowledge is to CODE, CODE, and CODE.

QuoteThe ONLY way to acquire this knowledge is to CODE, CODE, and CODE. Nice. That's three times, then the last version will be perfect

Cuch,@ "Depends on the job you applied for. And it depends on othe factors, such as exam questions and the person interviewing you."How about a "desk strategist" job at a leading IB? How useful is STL, etc. in GETTING that job?

QuoteOriginally posted by: AdvaitaCuch,@ "Depends on the job you applied for. And it depends on othe factors, such as exam questions and the person interviewing you."How about a "desk strategist" job at a leading IB? How useful is STL, etc. in GETTING that job?Hi Adva,Knock me over with a feather Difficult question for me to answer, but if you intend developing s/w using C++ then STL is an added advantage, for sure. There is a growing awareness of STL (btw it is C++!) so it could be a precondition. I think DCFC can give a better indication of the market.Is a Desk Strategist like a Quant developer?One thing is sure: STL - just like C++ - is for ever. So learning it will do no harm in the medium and long term.That was my 2 cents on the matter. Edit: No offence but doing something for 'that job' sounds a bit ad hoc. It will take a number of years to become proficient at C++, just like Maths. Some folk think that CS is easy and that you can learn it as go along...

Mark Joshi book is wonderful and very relevant. I could finish half of Joshi's book and found very useful. Very compact (which is its main stength), simple and very dense. I am still reading and trying his code. In my opinion, it very good place to get orientation. If someone can build the technical understanding (subtle points) around MJ book, then it would be good enough to land in good jobs with IBs. I think Duffy and Steve Dalton books in addition will cover all the other aspects in more detail. I like Duffy's course structure but I still found that I can not afford the price and it is quite packed (ideally it would be better if it can be spread over for few weeks to get some practice if it is feasible).Duffy's new book would be a welcome addition. At least, I can afford that..

QuoteDuffy's new book would be a welcome addition. At least, I can afford that.. I think it's good value, 26 chapters and working code and more...No doubt the BookShop will give a good price as well

QuoteOriginally posted by: rsneevasDuffy's new book would be a welcome addition. At least, I can afford that..I'm looking forward to the book too. As a (financial) math undergrad student, we often learn just the numerical algorithms in our math classes so Duffy's first book was a bit over my head with regards to STL, templates, etc.

QuoteOriginally posted by: danfcookQuoteOriginally posted by: rsneevasDuffy's new book would be a welcome addition. At least, I can afford that..I'm looking forward to the book too. As a (financial) math undergrad student, we often learn just the numerical algorithms in our math classes so Duffy's first book was a bit over my head with regards to STL, templates, etc.I must agree with you. At the time 2004 I assumed that people in QF knew the basics of C++. The new book starts from A and goes to Z. And there's quite a bit of numerical code in there as well.It seems that many people in C++ seem to learn the OO stuff (inheritance) first whereas GENERICS is closer to mathematics; I compare a template to a Hillbert space and the instantiated classes correspond to the usual Hilbert spaces a real line, sequences etc.BTW many people have problems with how to set up a project (include files etc.) In new book I have a full chapter on this.