Re: [Quantlib-dev] Contributions

[Simon I. - S. <Sim...@st...>]

Hi Luigi,

Whilst I'd love to see what many other people are working on, I'm not
sure how this would work... a lot of contributions change the existing
codebase - so you'd either have duplicate code in the ql/contrib folder
or it would be purely for new classes.

If there isn't time for you to test the code - why not have a different
branch of the library which includes the contributions. Effectively two
releases: a full release and a beta version of the next.

Alternatively, you could accept all contributions which compile and
don't break the test-suite...? Admittedly, this might mean that some of
the interfaces change between releases - which can be a pain, unless the
important interfaces are in the test-suite.

Cheers,

Simon




-----Original Message-----
From: Luigi Ballabio [mailto:lui...@gm...] 
Sent: 07 May 2008 16:54
To: QuantLib developers; QuantLib users
Cc: jwm...@ya...; toy...@ho...; ri...@ma...;
all...@ya...; fho...@gm...; mar...@go...;
Simon Ibbotson - Straumur
Subject: Contributions


Hi all,
	as you've noticed by my delay in answering posts on the mailing
lists,
I'm not being able to allocate much time for QuantLib.  Which is more
the pity, since we've received quite a few contributions I'd like to
look at and include in the library.  Moreover, I'd like to put out a
release in the next couple of months.

Putting this all together---what would you say about creating a
ql/contrib folder in the library where I would add contributions just as
they are? The main points would be:

(+) contributions would go in without much effort; we'd just have to
check that they compile against the current code;
(+) they would be out earlier, thus receiving more feedback;
(-) their interfaces wouldn't be considered stable; based on feedback,
they could be reworked so that they fit preferred usage and the design
of the rest of the library.

In time, we'd look at them more carefully and move them to the regular
folder hierarchy.

And now that I think of it, we don't even need a ql/contrib folder---we
can use the existing ql/experimental instead.

What do you think?

Luigi



-- 

There are two ways of constructing a software design. One way is to 
make it so simple that there are obviously no deficiencies. And the 
other way is to make it so complicated that there are no obvious 
deficiencies. 
-- C. A. R. Hoare