C(omp)++ | A Portal for the Computational Finance Community

Happy Easter...

...once again, a few days late. Also, it is somewhat telling that my latest post was to wish you a happy new year.

Oh well. As usual with these posts, I bring news on my Implementing QuantLib drafts. If you follow the link, you'll find a new version of chapter 3 in which I completed the section on yield term structures. As usual, I'll be grateful for any comments and/or corrections.

Luigi

Happy New Year!

I'm a few days late, I know. I was on vacation after the flurry of activity at the end of 2007; we had a new QuantLib release on December 24th (if you're interested, you can get it while it's hot at quantlib.org). My drafts of Implementing QuantLib got updated, too; there's no new content, but the existing material was modified to reflect changes in the library.

Again, happy 2008.

Luigi

Luigi Ballabio: implementing QuantLib (again)

I really have to come up with a catchier title for these posts.

However: the Implementing QuantLib page now links (thanks to a suggestion from Comp++'s own Daniel Duffy) to a tentative table of contents. Also new is a small appendix describing the code conventions we use in QuantLib---and I use in the book.

As usual, I'll be grateful for any comments and corrections.

Luigi

Luigi Ballabio: Implementing QuantLib

Hi all, Another half chapter of Implementing QuantLib is available. As promised, it describes a few classes I glossed over in the previous chapters. You can find the draft here.

As usual, comments are welcome.

Luigi

Implementing QuantLib Part 2

Hi all,

Just a short post to inform you that I've uploaded another half chapter of my ongoing book---as well as a revision of the previous chapter. You can find the draft here.

As usual, comments are welcome.

Luigi

Luigi Ballabio: A first look into QuantLib

If you happened to have a look at QuantLib, it won't have escaped you that the library sorely lacks documentation. I'm not talking of some kind of class reference; that one is available---mostly because it can be generated automatically. What is missing is a description of the architecture of the library, and of how the various classes interact.

Well, as the song says, rejoice, all ye faithful: such documentation is finally being written. Here is a draft of a chapter. You're welcome to read it; I'll be grateful for any feedback, corrections, and criticisms. In particular, criticisms will help us improve both the docs and the library itself, so fire ahead.

Luigi

Email me!

Download Chapter 2 here

(right-click & save target to download this file)

Luigi Ballabio: Show Me the Code!

As I was checking the new entries in my RSS feeds, I recently stumbled upon this blog entry. Go ahead, read it. I'll wait right here.

Are you back, gentle reader? What do you think? Although the blogger referred in particular to computer science, his argument applies to almost all fields. In fact, Mark Joshi made the very same point in a conversation we had last summer. How much can you trust the results of a quantitative-finance paper if they're not repeatable?

Of course, one could reproduce the results - in a way. Based on the author's description, one could code the same calculations. Like we have that kind of time, right?

In fact, the insight toolkit mentioned at the end of the post is a nice idea. If authors wrote their code on top of such a library and made it available, the advantage would be twofold. On the one hand, they would build upon classes and functions reviewed and tested by scores of other users. On the other hand, readers familiar with the toolkit would find it much easier to read the author's code and could focus on understanding the actual new algorithms rather than the scaffolding.

Need I mention that we have a similar toolkit for quantitative finance?

Thoughts anyone? I'll be glad to read them.

Luigi