I want to value portfolios of developed and emerging markets interest rate swaps. I've outgrown Excel, but I'm finding quantlib very dense and too complicated for my needs. Is there anything "in between" that people can suggest? I'd prefer Python. I don't need any optionality. All I need is to create cashflow objects and then discount them on a curve to find current NPV. I'll be doing this on many different portfolios in quasi-real time which is the reason Excel is too clunky. Or really should I invest the time in trying to figure out quantlib.....?

I was/am in nearly the same boat and so far I've written C# wrappers for just about all the simple stuff that you and I need. Quantlib had a solid day counting/interpoation/ stat. distribution support, so you can just leverage that. But yes, from my experience quantlib is overengineered and i'm guessing the people developing it don't have much idea of what is needed in the real world.

Thanks halik I hear you. Am not qualified to judge if its too complex for everybody. I've seen some pretty complex stuff getting done in my bank (which is french and loves maths) and quantlib seems to target that segment. But I am an IRS salesperson dealing with macro hedge fund clients. Directional vanilla and light non-linear is where it's at for me. I need to keep track of valuations of portfolios of fairly simple products. I see a lot of traders around me who need similar - an order of magnitude more capability than Excel, but an order of magnitude less capability than what quantlib seems to offer. And a heavy emphasis on pragmatism to get things done - ie easy docs. Anyway not a criticism - just horses for courses. Perhaps we should start our own vanilla library in something like Python? or a simplified wrapper for Quantlib? Python seems to be growing steadily in the financial community. I am particular impressed with the Pandas timeseries library for example http://code.google.com/p/pandas/. Resolver one also looks interesting, but is locked into Microsoft's world. Anyway any suggestions appreciated.

Try Matlab, now that you can hook in GPU acceleration , makes it compelling in my opinion

And Matlab has a fully functional set of canned financial markets instrument pricers? Or close to?

It seems the Financial Derivatives Toolbox of Matlab should provide you with some of the needed stuff. If you look at potential integration of your tool with powerful visualisation, deployment or inside of a large app, you should definitely keep using an open environment like Python & numpy/scipy. Developing some methods for directional vanilla & lightly non-linear should be pretty quick, robust and scaling with your app.

Why not using the Excel add-in QuantLibXL?QuantLibXLDominik

Have you considered QuantLib-Python? http://sites.google.com/site/tgwena/is a good startup document (still in progress though)