[Soultrader]

Decided to take another stab at learning programming. Its been my number weakness for years and yet Ive been delaying this requirement.

Can anyone recommend any programming language and books on where to start? Primarily interested in being able to code strategies for various platforms. So probably best if I begin with something general and compatible with all the mainstream platforms. Thanks!

[diablo272]

If you want a general idea, I think C++ is a good place to start.

C++ Language Tutorial

As far as a trading-specific language, I think Easy Language is most widely used, though I have no experience with it:

https://www.tradestation.com/support...ylanguage.aspx

I'm a programming newbie, but I've read a lot about getting started...It's just that I never get past the "getting started" phase. I've written a "hello world" program in c++ many, many times.

[programmer]

If you want to be serious :

1) Refresh your math skills(take a course, no excuses...I said serious...).

2) Have a clear goal what you want to attempt.

3) C# is great (Ninja), Easylang is ok(depends on what you want to achieve, thus if you have very very smart ideas (like me ...lol) you will need to exe them on a platform that has no limits ( TS and MC are not the way to go , why ? well these programs cannot calc Multimarket analysis well at all unless you use Global variables, and for that you will need to know C# ...again its all in what you want to attempt).


4) If you are going for the pro stuff (I have been programming for over a decade , so from experience I must stress that the basics is what counts), you will need 1-2 years of experience to even attempt smart coding , if you want easy lang level - 3-6 months - with proper guidance .

[sneo]

as what programmer mentioned, it depends on what you really want to achieve with the programming language you learn.

I'm not too sure if EasyLanguage is really classified as a proper programming language, what I do know from my own experience is:
1. C++:
for speed - writing programs (.exe) that speeds processes that you may want to do routinely. e.g. I use the end program to clean up ;arge xls and csv data files for data analysis.
to be noted - strong in arrays, but dont try if you want to manipulate matrices. Its going to be a real pain.

2. MATLAB:
for manipulation of matrices, mathematical programming
to be noted - not really portable to other computers (the MCR is a joke...the installation is big and you need to have the MCR relating to the same MATLAB version you created your little program). Its really expensive too.

3. R:
for data analysis. its called a statistical analysis environment, but I tend to view it as a general data analysis environment since there are so many packages for different usage asides statistical analysis.
to be noted - its free, but GUI is really limited. console is the main way to do things. If GUI is important, Tinn-R has to be a companion to R. The combination will give a MATLAB-like GUI workspace.

[TomH]

Originally Posted by sneo »

3. R:
for data analysis. its called a statistical analysis environment, but I tend to view it as a general data analysis environment since there are so many packages for different usage asides statistical analysis.
to be noted - its free, but GUI is really limited. console is the main way to do things. If GUI is important, Tinn-R has to be a companion to R. The combination will give a MATLAB-like GUI workspace.

R is more of a scripting language than a serious general programming environment, as mentioned above, but for financial analysis you should be able to implement some serious stuff after a few days with packages like R's quantmod - that would take months from a standing start with C++

If the original poster wants to learn the principles of programming then C or Java are the sorts of things used in undergraduate CompSci classes. However for day to day rapid development I previously use perl and Java, and more recently hack up financial models in R

In the real world, if you learn the principles of loops, conditionals, variable types and functions, then you have the basics of most procedural languages, and converting your knowledge from say VBscript, javascript, Bash, R, perl, and most of the platform specific scripting languages for the trading platform is a matter reading the manual.

[BlowFish]

I would recommend (fairly highly) Easylanguage for numerous reasons.

It is fast to get results which should a) keep you motivated b) allow you to learn quicker.

There are thousands and thousands of studies and code samples. A good way of learning is looking at how other people do it. Again you will get quick practical results modifying someone else's code whilst learning a bit at a time. This leads on to pier support. Just look at TL and ask yourself if I need help which language is best supported?

It is based on Pascal.... some might mock that but personally I think its a strong language. One of its design criteria was that it was easy to learn and has stringent compile time checking (it is usually harder to find bugs at run time, though this is less important with 'scripted' stuff). There will always be a flavour of the month language C++, Java, C Sharp, Ruby etc. Fashion plays a big part in programming languages though many are main stream now. If you can program the languages become less important.

The biggest challenge you are likely to find is not the language itself but the functions (trendy languages will call them methods) you call to do basic stuff, e.g. read data series, plot something, find out the time etc. EL implements this cleanly and more importantly simply.

Nothing wrong with starting a project and just throwing yourself in. (Hello world is a popular one!) despite doing a degree in computer science (in the 70's) I didn't really become accomplished until I actually started writing code professionaly.

[sneo]

Originally Posted by TomH »

R is more of a scripting language than a serious general programming environment, as mentioned above, but for financial analysis you should be able to implement some serious stuff after a few days with packages like R's quantmod - that would take months from a standing start with C++

If the original poster wants to learn the principles of programming then C or Java are the sorts of things used in undergraduate CompSci classes. However for day to day rapid development I previously use perl and Java, and more recently hack up financial models in R

In the real world, if you learn the principles of loops, conditionals, variable types and functions, then you have the basics of most procedural languages, and converting your knowledge from say VBscript, javascript, Bash, R, perl, and most of the platform specific scripting languages for the trading platform is a matter reading the manual.

heh, I dont know whether to call MATLAB and R a programming language or not, but, I prefer to use the word environment XD. as Tom has mentioned, it reduces the time from idea to an actual work.

[BlowFish]

I should add my answer was more from the point of view of learning 'general' programming. If your bias was towards statistical analysis and mathematical modelling I'd probably have to modify my answer