SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
sergBerger
Topic Author
Posts: 15
Joined: July 22nd, 2008, 10:44 pm

C# C++ Quantlib/xlw

December 21st, 2008, 3:35 pm

Is it possible to use quantlib and xlw with C# (comparison with C++, if possible). I am familiar with C++ and thinking for learning C#, but can't decide which language to focus on. I know C# has less development time and more productivity, but I do want the capability to use quantlib and easy interfacing with Excel. Any comments from experienced people...
 
User avatar
lballabio
Posts: 983
Joined: January 19th, 2004, 12:34 pm

C# C++ Quantlib/xlw

December 23rd, 2008, 8:38 am

There are QuantLib bindings for C#. Look for "QuantLib extensions" on the Sourceforge download page.
 
User avatar
rmeenaks
Posts: 186
Joined: May 1st, 2006, 2:31 pm

C# C++ Quantlib/xlw

December 23rd, 2008, 3:08 pm

See the following site on C# and Quantlib:http://www.qlnet.org/Cheers,Ram
 
User avatar
sergBerger
Topic Author
Posts: 15
Joined: July 22nd, 2008, 10:44 pm

C# C++ Quantlib/xlw

December 24th, 2008, 4:30 pm

thanks guys
 
User avatar
Smiling
Posts: 15
Joined: April 16th, 2007, 1:09 am

C# C++ Quantlib/xlw

January 2nd, 2009, 2:10 am

QuoteOriginally posted by: sergBergerIs it possible to use quantlib and xlw with C# (comparison with C++, if possible). I am familiar with C++ and thinking for learning C#, but can't decide which language to focus on. I know C# has less development time and more productivity, but I do want the capability to use quantlib and easy interfacing with Excel. Any comments from experienced people...C# is definetely the way to go. C++ applications become more and more rare since obvios productivity benefits of .NET languages.Some may argue that C++ usually results in a faster code, but let's just face it - most of the time a quants' code is not the best oneand not the most performant one. If a performance is the issue - the prototype coded by a quant could be easily handed over to prof.developer to rewrite it in C++ or some fancy performance oriented language (e.g. for NVidia's CUDA uses a old C language - but gives you ability to run your calculations on 300 CPUs (which comes at modest price of 500USD - the top end gfx card) at the same time, etc)C# is also much easier to interface with Excel - most of the classes could be made visible as COM Excel add-in by just adding [ComVisible]attribute to the class and regasm-ing the resulting assembly. However once you have more than a dozen functions to maintain - that straightfowardinterfacing approach might become assembly and COM hell (this is a "new technology" spread Microsoft charges us on top of "dll hell" fee we have had to pay for the last 15 years).The good thing - there are several good libraries which could make Excel interfacing much easier - I personally prefer Excel4Net - http://excel4net.com,not very fancy at outside, but very easy to use - no COM knowledge required.
 
User avatar
Cuchulainn
Posts: 62149
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

C# C++ Quantlib/xlw

January 2nd, 2009, 9:26 am

QuoteC# is definetely the way to go. I agree. It is a very productive (n/1 for n >= 3) environment for desktop applications. It is very similar to C++ in syntax plus all the libraries on top.A C++ developer can learn C# in a matter of days/weeks/months. But it lacks many numeric libraries at this moment.
Last edited by Cuchulainn on January 1st, 2009, 11:00 pm, edited 1 time in total.
 
User avatar
Smiling
Posts: 15
Joined: April 16th, 2007, 1:09 am

C# C++ Quantlib/xlw

January 2nd, 2009, 1:28 pm

QuoteOriginally posted by: CuchulainnQuoteC# is definetely the way to go. I agree. It is a very productive (n/1 for n >= 3) environment for desktop applications. It is very similar to C++ in syntax plus all the libraries on top.A C++ developer can learn C# in a matter of days/weeks/months. But it lacks many numeric libraries at this moment.Compared to C/C++ ? Yes, but all those libraries has been built over the last 10? 20? years and 7 years C# was around was not enough time to catch up.Desktop applications? It is pretty much anywhere.I believe that over the following few years there be more and more numerical libraries will be ported to .NET,(or at least .NET -> native code wrappers will be provided) and eventually (in a 5-10 years) most of C++ applications will be either abandoned, rewritten or proxied/emulated (as MVS code nowdays) with a managed code.
 
User avatar
Cuchulainn
Posts: 62149
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

C# C++ Quantlib/xlw

January 2nd, 2009, 1:42 pm

QuoteOriginally posted by: SmilingQuoteOriginally posted by: CuchulainnQuoteC# is definetely the way to go. I agree. It is a very productive (n/1 for n >= 3) environment for desktop applications. It is very similar to C++ in syntax plus all the libraries on top.A C++ developer can learn C# in a matter of days/weeks/months. But it lacks many numeric libraries at this moment.Compared to C/C++ ? Yes, but all those libraries has been built over the last 10? 20? years and 7 years C# was around was not enough time to catch up.Desktop applications? It is pretty much anywhere.I believe that over the following few years there be more and more numerical libraries will be ported to .NET,(or at least .NET -> native code wrappers will be provided) and eventually (in a 5-10 years) most of C++ applications will be either abandoned, rewritten or proxied/emulated (as MVS code nowdays) with a managed code.The libraries are lacking, NOW! In absolute terms. On the other hand, well-designed C++ code (especially if you have the UML blueprints is a sinch to port to C#.You need to qualify the kinds of applications that C# and C++ are best at. C++ has the best libraries (i.e. boost) for computation and networking, not a strong point of MS s/w.My big question would be if C# in its current form will be around in 5 years time? BTW that should not be a reason for not learning it Both management and technical decisions are important here. Tech stuff alone is not the determinat, always. Regarding numerical libraries, I will believe it when I see it (wrappers are awful and too slow for numerics). Who will make them? I only know of a handful of numerical analysts to do this and I get the impression that they all doing high-powered physics and biology PDEs in Fortran and C++.Maybe numerics is too much of a niche market for anyone to 1) be interested in, tech-wise 2) making a business opportunity. As I said, numerics is another ball park.And for parallel, MPI and OpenMP are here to stay.IMO an organisation should use C++ and C# (inclusive OR). I would not relish GUI stuff in C++, for example.
Last edited by Cuchulainn on January 1st, 2009, 11:00 pm, edited 1 time in total.
ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...


GZIP: On