@Tom Breloff Quantlib is useful because it has an industry strength calendar/date library containing all the day count conventions and its calendar is maintained (e.g. when the Queen of England turns 100 holiday comes along, you don't need to enter that in your calendar manually). I don't think the underlying methods are particularly complex in Quantlib, though I don't know what would be a good measure of complexity. Quantlib's class structure has a certain amount of complexity, however, as it implements many GOF design patterns, and I don't quite know how they translate to a completely different, functional paradigm. I mainly need the date libraries and certain simple pricing methods, which I can build my more complex calculations on. A good example of how these simpler Quantlib pricings go is described in a simple language here: https://quantcorner.wordpress.com/2011/01/31/us-treasury-notes-and-quantlib/, where the fixedRateBond, bondEngine and the discountingTermStructure are the three main objects working together in some sort of strategy pattern. This particular pattern is given in Julia since it is a functional language, but complexity arises around the class hierarchy of each of these objects.
On Friday, September 11, 2015 at 7:51:29 PM UTC-4, Tom Breloff wrote: > > My guess, based on what little I know of Keno's progress, is that we're > probably within a couple months of a C++ wrapping party... I know I'm > planning on doing some of it myself, just waiting for the release buzz to > calm down before getting my hands dirty. I'm only mildly familiar with > quantlib... How complex are the underlying methods that you care about? It > would be better to implement in pure Julia if reasonable. > > On Friday, September 11, 2015, Ferenc Szalma <[email protected] > <javascript:>> wrote: > >> Yes, watching just now Keno Fischer's talk at JuliaCon 2015 after >> watching yours, I can see the big difference in complexity between the C >> and C++ interfaces. Still it seems like Keno figured C++ out eventually and >> what he shows is pretty impressive. Do you have any info on how stable >> Julia 0.4 is on the necessary newer version of LLVM? Also what do you see >> as other limitations that you mention? >> >> On Friday, September 11, 2015 at 5:48:55 PM UTC-4, Avik Sengupta wrote: >>> >>> While Julia has had a brilliant C interface from the start, calling C++ >>> has been much more difficult. There is work going on under Cxx.jl to make >>> it easy to call C++, but it is still not seamless (it needs a newer version >>> of LLVM compared to base Julia, for one, and there are other limitations). >>> >>> I would hope that Julia has fast native implementation of many of the >>> algorithms in Quantlib. A lot of the code in Quantlib is core mathematical >>> routines, which are already present in base Julia or many of its packages. >>> So given the difficulties in calling C++, I would personally favour a >>> native implementation. But of course, such decisions depends heavily on >>> your situation. >>> >>> Regards >>> - >>> Avik >>> >>> >>> On Friday, 11 September 2015 17:29:32 UTC+1, Ferenc Szalma wrote: >>>> >>>> @Avik Sengupta Thanks for coding up Ito.jl even if it is really just >>>> the start. Comparing Ito's code with the Quantlib C++ code it helps a lot >>>> with understanding how C++ code translates into Julia code, especially for >>>> a newbie in Julia like myself. I just noticed you gave a talk on how to >>>> interface various languages with Julia on YouTube: >>>> https://www.youtube.com/watch?v=AyeArSTzas8 in June and posted only >>>> just a few days ago. Do you think it'd be easier to just call Quantlib C++ >>>> functions from Julia or re-writing that code in Julia would be easier. >>>> Quantlib object are certainly embedded into a complex structure, so >>>> translation may be difficult. I am just trying to find the most effective >>>> way of using Quantlib or Quantlib-like objects and functions with the >>>> least coding time overhead. >>>> >>>> >>>> On Friday, September 11, 2015 at 2:13:30 AM UTC-4, Avik Sengupta wrote: >>>>> >>>>> My hope has been that Ito would become the consolidated financial math >>>>> package, with functionality similar to Quantlib. Unfortunately, while I >>>>> laid down the base, I haven't made as much progress as I would have liked >>>>> (primarily for lack of time), but I do hope to rectify that in the near >>>>> future. Needless to say, contributions welcome. >>>>> >>>>> >>>>> On Friday, 11 September 2015 07:05:39 UTC+1, Ferenc Szalma wrote: >>>>>> >>>>>> Are there any quant finance packages for Julia? I see some >>>>>> rudimentary calendar and day-counting in Ito.js for example but not much >>>>>> for even a simple yield2price or price2yield or any bond objects in >>>>>> Julia >>>>>> packages on GitHub. What is the best approach, using C++ function/object >>>>>> from Quantlib, to finance in Julia? >>>>>> >>>>>
