Re: [Quantlib-dev] Intra-day date representation ?

[Dirk E. <ed...@de...>]

On 7 August 2014 at 12:34, Luigi Ballabio wrote:
| Unfortunately C++11 is out of the reach of a lot of people still
| suffering under the heel of their IT groups :)
| (Anecdotal evidence: at StatPro Italia, out boxes are still on the
| long-term support Ubuntu release from April 2012---we'll be switching
| to the April 2014 one, eventually, but that's lower priority that a
| lot of other things on our plate---so we're using gcc 4.6.2 from March
| 2012. And that's Ubuntu: if we were using RedHat, for instance, that
| would be more even conservative.)

(Seconded from over here in R-land. While we can now use C++11 on packages, I
also know eg via the friends running RStudio that there are still RHEL 4
installation out there.  So to be _used_ we may need to be conservative in
what we depend upon.)

(That said there is crazy new shit around. Seen the C++ Streams library using
C++14?  http://jscheiny.github.io/Streams/  Just wow.)
 
| Anyway: I'd love all of the things on your list (C++, multithreading,
| boost) and I'd be glad to start a 2.0 branch, but I have no idea how
| to go about a lot of problems yet: see my talk at Dusseldorf
| <http://vimeo.com/79724861> for the major ones.
| 
| So: let's start the discussion. I'd love to hear from people with some
| ideas about the above...

Why don't start by generalizing Date :) (Only half-kidding, maybe incremental
approaches are fine rather than grand rewrite that then may stall ...)

Dirk

-- 
http://dirk.eddelbuettel.com | @eddelbuettel | ed...@de...