<t>VS2005 has issues with Vista, but some of these go away when you turn off Vista's UAC. Or upgrade to VS2008 which is much better with Vista. Vista is soo bad that I don't use it at home, but bootstrap XP from a virtual machine to use VS 2005. But of all these tales of woe its good to hear that wi...

<t>Yes, I've used vsto.1) Ok if you want to use Excel 2007's new GUI.2) Most developers want an add-in in an xll flavour. So they usually go with using an XLL to .net wrapper technology. I've not seen a native C++ XLL for some time with most of my companies clients. The current SDK for producing .ne...

<t>Out of interest are you seeing this problem when the singleton is exported from a DLL at all? I know it should just be a header file, however I had a problem with a templated singleton in vs2005 which lead to multiple singleton instances being created because I forgot to define __declspec on the ...

Yes, but C++/CLI is clunky (i.e) you have to write more code then the c# version. However if you have to mix and match between C++ and C++/CLI then yes a definite bonus.Can C++/CLI run on other platforms then windows, is they a mono version at all?

True there are some excellent threading packages out there. But its not the same as inbuilt C++ runtime support for thread locking, however I accept the software that is produced will have the same end user functionality.

<t>And there are alternatives to writing in assembler, like using GPU based accelarators for performance. Again programed via C++ or extended C language's such as cg.My only gripe with C++ is the lack of inbuilt support for thead constructs, like object level locking that exists in .net and java. </t>

<t>Absolutely, but if you can avoid the performance,code bloat and installation issues of COM then why include it. End.As far as openmp, are you thinking of calling it from within an XLL? MS have other examples using their technology such as Excel Calcuation Server for performing grid based calculat...

<t>The UDF AddIn's in Excel 2007 are COM and that means a lot of bolier plate code moving from the COM parameters to .net types.If you use a thrid-party AddIn it will generate all the bridge code for you, so you write in .net without having to write boiler plate code to perform the XLL type to .net ...

<t>I think it is Excel that drives VB in banks, more then anything else. How well does Excel fit into being used for Algo trading or if it doesn't then what tool do we need for quickly writing an algo trading strategy? C# does seems to be killing of C++ development of XLL's. I don't know anyone who ...

<t>On the subject of memory management, I thought the whole point of the managed approach was that it was faster then C++ at allocation. C++ (unless you have written your own version of new and delete) has to search through a tree to find space for allocation. While the CLR always has a region of un...

Whats memcache, just memory caching or some product?M

<t>I second that, its really like doing a binary read using HDF or netCDF. However the overhead of I/O also makes sense to compress the file (using zip, gz etc) and then uncompress the file in memory as you read it in. I've seen that make the load of files an order of magnitude quicker. This is a te...

<t>MFC is now the new cobol, just like vb6 has become. We have loads of it, but cannot find anyone who wants to code with it at all and MFC is to expensive to code in compared with managed code.VS2008 will allow you to do lots of MFC and C++/CLI, but we'd then need an MFC coder who can program in .n...

If you mean a real quant is a person who developes new models for traders to price new instruments types then yes they do that. But remember they also have to support the models they have already developed and fix any problems.M

The other source of data (not on bloomberg) for vol surface is from super derivatives. This does has the skew (otm/itm calls) you want on swaption, caps and floors. However as usual not cheap, but not anymore then bloomberg I'd expect.