Menu
▾
▴
quantlib-dev
|
From: Luigi B. <lui...@gm...> - 2010-01-12 16:26:02
|
Hi all, we're almost there. A beta for version 1.0 is available at <https://sourceforge.net/projects/quantlib/files/prerelease/>. In everybody's interest, please give this version a try and report any problems you might have. Reports can be either sent here on the list or filed on the Sourceforge bug tracker. Thanks, Luigi -- Everything that can be invented has been invented. -- Charles Duell, Director of U.S. Patent Office, 1899 |
|
From: Piter D. <pit...@pi...> - 2010-01-12 23:18:25
|
> In everybody's interest, please give this version a try and report any > problems you might have. Reports can be either sent here on the list or > filed on the Sourceforge bug tracker. Is there a problem if we report good news too? Test suite run successfully for Windows 7, Boost 1.41 and Visual C+ 2008 Express Edition. 14>Tests completed in 28 m 58 s 14>Test suite "Master Test Suite" passed with: 14> 1689 assertions out of 1689 passed 14> 446 test cases out of 446 passed 14>Build log was saved at "file://c:\Users\piterdias\develop\src\QuantLib-1.0b1\test-suite\build\vc90\Win32\Release (static runtime)\BuildLog.htm" 14>testsuite - 0 error(s), 0 warning(s) ========== Build: 15 succeeded, 0 failed, 0 up-to-date, 0 skipped ========== ------------------------- Piter Dias pit...@pi... |
|
From: N.Cao <ca...@fi...> - 2010-01-13 02:31:37
|
Hi all, I am currently developing the F# version of QuantLib on SourgeForge and it is ranked no.1 among F# free/open-source softwares. I am here looking for more developers to join this project. If you are interested, please send an email to me. For more information, please visit www.quantifa.org. Cheers, Ning Cao |
|
From: Luigi B. <lui...@gm...> - 2010-01-13 08:41:45
|
On Tue, 2010-01-12 at 21:18 -0200, Piter Dias wrote: > > In everybody's interest, please give this version a try and report any > > problems you might have. Reports can be either sent here on the list or > > filed on the Sourceforge bug tracker. > > Is there a problem if we report good news too? No problem. Just don't file them on the tracker :) > Test suite run successfully for Windows 7, Boost 1.41 and Visual C+ 2008 > Express Edition. Ok, thanks. Luigi -- Father's got the sack from the water-works For smoking of his old cherry-briar; Father's got the sack from the water-works 'Cos he might set the water-works on fire. |
|
From: Dirk E. <ed...@de...> - 2010-01-13 12:45:10
|
On 12 January 2010 at 17:25, Luigi Ballabio wrote: | Hi all, | we're almost there. A beta for version 1.0 is available at | <https://sourceforge.net/projects/quantlib/files/prerelease/>. | | In everybody's interest, please give this version a try and report any | problems you might have. Reports can be either sent here on the list or | filed on the Sourceforge bug tracker. Debian builds also passed with flying colours. A few quick remarks though i) the version number propagated from configure.ac is not of the a.b.c form, maybe we should stick with that (eg. 1.0.0 ?) ii) the dynamic library major is still '0' as in /usr/lib/libQuantLib.so.0 -- but the long anticipated API freeze should probably make that 1, no? iii) MarketModels needs a man pages, I will cook something up and commit it (with integration to the examples Makefile etc) unless someone beats me to it. Dirk -- Three out of two people have difficulties with fractions. |
|
From: Ferdinando A. <na...@am...> - 2010-01-13 13:01:15
|
On Wed, Jan 13, 2010 at 1:30 PM, Dirk Eddelbuettel <ed...@de...> wrote: > i) the version number propagated from configure.ac is not of the a.b.c > form, maybe we should stick with that (eg. 1.0.0 ?) I would prefer to stick with 1.0.0 too, but I look forward to Luigi's 1.0 arguments ciao -- Nando |
|
From: Luigi B. <lui...@gm...> - 2010-01-13 13:52:05
|
On Wed, 2010-01-13 at 13:53 +0100, Ferdinando Ametrano wrote:
> On Wed, Jan 13, 2010 at 1:30 PM, Dirk Eddelbuettel <ed...@de...> wrote:
> > i) the version number propagated from configure.ac is not of the a.b.c
> > form, maybe we should stick with that (eg. 1.0.0 ?)
>
> I would prefer to stick with 1.0.0 too, but I look forward to Luigi's
> 1.0 arguments
Looking forward to it? You won't be disappointed. I have a rational,
well-thought and cogent argument for the 1.0 form. It goes like this:
1.0.0 is ugly.
But maybe I'm missing something. What would the explicit second 0 give
us that an implied one doesn't?
Later,
Luigi
--
There is no such thing as public opinion. There is only published
opinion.
-- Winston Churchill
|
|
From: Ferdinando A. <na...@am...> - 2010-01-13 14:15:49
|
On Wed, Jan 13, 2010 at 2:49 PM, Luigi Ballabio <lui...@gm...> wrote: > 1.0.0 is ugly. I agree and I would write about "releasing 1.0" keeping 1.0.0 in the code code is ugly anyway... isn't it? ciao -- Nando |
|
From: Luigi B. <lui...@gm...> - 2010-01-13 13:53:11
|
On Wed, 2010-01-13 at 06:30 -0600, Dirk Eddelbuettel wrote: > ii) the dynamic library major is still '0' as in /usr/lib/libQuantLib.so.0 > -- but the long anticipated API freeze should probably make that 1, no? I thought we could go for a fresh start---previous versions had names such as libQuantLib-0.9.9.so, so they counted as different, right? Or was there a libQuantLib.so.0 symbolic link as well? > iii) MarketModels needs a man pages, I will cook something up and commit it > (with integration to the examples Makefile etc) unless someone beats > me to it. Right. Please go ahead. Luigi -- Newton's Law of Gravitation: What goes up must come down. But don't expect it to come down where you can find it. Murphy's Law applies to Newton's. |
|
From: Dirk E. <ed...@de...> - 2010-01-13 14:55:53
|
My $0.02: 'triplets' give us more rope for intermediary releases like 1.0.1, 1.0.2, etc. If you as release manager don't plan to do those but rather want to jump to 1.1, 1.2, 1.3 -- fine. I'd say that the empirical evidence across open source projects is in favour of 'triplets', but hey, beauty is in the eye of the beholder. Your call. We all will just whine. On 13 January 2010 at 14:52, Luigi Ballabio wrote: | On Wed, 2010-01-13 at 06:30 -0600, Dirk Eddelbuettel wrote: | > ii) the dynamic library major is still '0' as in /usr/lib/libQuantLib.so.0 | > -- but the long anticipated API freeze should probably make that 1, no? | | I thought we could go for a fresh start---previous versions had names | such as libQuantLib-0.9.9.so, so they counted as different, right? Or | was there a libQuantLib.so.0 symbolic link as well? Here is what my Debian box has: edd@ron:~> ls -ltr /usr/lib/libQuantLib* -rw-r--r-- 1 root root 12042164 2009-10-11 07:02 /usr/lib/libQuantLib-0.9.7.so -rw-r--r-- 1 root root 13386964 2009-12-11 20:19 /usr/lib/libQuantLib-0.9.9.so -rw-r--r-- 1 root root 13636428 2010-01-13 00:41 /usr/lib/libQuantLib.so.0.0.0 -rw-r--r-- 1 root root 50923590 2010-01-13 00:41 /usr/lib/libQuantLib.a lrwxrwxrwx 1 root root 20 2010-01-13 06:21 /usr/lib/libQuantLib.so.0 -> libQuantLib.so.0.0.0 lrwxrwxrwx 1 root root 20 2010-01-13 06:21 /usr/lib/libQuantLib.so -> libQuantLib.so.0.0.0 edd@ron:~> We keep just the .so of older release; code built against those stills runs. Libtool (which I still know too little about) in its magic uses three values anyway so we have the 'active' version libQuantlib.so.0.0.0 with two softlinks use while building code. I think that major number ought to be '1' as in /usr/lib/libQuantLib.so.1 -> libQuantLib.so.1.0.0 to signal the long-awaited API freeze for 1.0. No? With that I'd also make the Debian package name 'libquantlib1-dev' rather the current 'libquantlib0-dev'. But that is all cosmetic at then end of the day. What mattered is that we have distinct packages allowing distinct versions to co-exist: edd@ron:~> dpkg -l | grep libquantlib- ii libquantlib-0.9.7 0.9.7-1+b2 Quantitative Finance Library -- development package ii libquantlib-0.9.9 0.9.9-3 Quantitative Finance Library -- development package ii libquantlib-1.0.0 1.0.0~20100112-1 Quantitative Finance Library -- development package edd@ron:~> And whether that is libquantlib-1.0.0 or libquantlib-1.0 doesn't really matter. Apart from the fact that I created precedent (:-/) but I can always fall back in line for libquantlib-1.1 if you really really howl. | > iii) MarketModels needs a man pages, I will cook something up and commit it | > (with integration to the examples Makefile etc) unless someone beats | > me to it. | | Right. Please go ahead. Will do if I get a few moments. Dirk -- Three out of two people have difficulties with fractions. |
|
From: Luigi B. <lui...@gm...> - 2010-01-13 17:18:12
|
On Wed, 2010-01-13 at 08:55 -0600, Dirk Eddelbuettel wrote: > My $0.02: 'triplets' give us more rope for intermediary releases like > 1.0.1, 1.0.2, etc. If you as release manager don't plan to do those > but rather want to jump to 1.1, 1.2, 1.3 -- fine. Dirk, I'd do both. I'm not discarding triplets; I'm just writing 1.0 instead of 1.0.0 for brevity. From there, new work would lead to 1.1 (short for 1.1.0); but if there were bug fixes, I'd make a 1.0.1 release to address them. Luigi -- Do the right thing. It will gratify some people and astonish the rest. -- Mark Twain |
|
From: Toyin A. <toy...@ho...> - 2010-01-19 04:04:45
|
Hi all,
I have been looking at the class AnalyticHestonEngine::Integration and I find that there are two variables that seem to be declared but never used (AnalyticHestonEngine.hpp).
mutable Real c_inf_;
mutable boost::function1<Real, Real> f_;
Also line 231 within the file AnalyticHestonEngine.cpp returns an expression, but the two (const std::complex<Real>) computed local expressions above it are never used in the return value. Am I missing something?
Finally, can someone verify the following (Line 537 of AnalyticHestonEngine.cpp)
(
boost::function1<Real, Real>(
if_then_else_return ( (_1+1.0)*c_inf> QL_EPSILON,
bind(f, -bind(std::ptr_fun<Real,Real>(std::log),
0.5*_1+0.5 )/c_inf )/((_1+1.0)*c_inf),
bind(constant<Real, Real>(0.0), _1))));
is basically the same as
double dRes = 0.0;
if ((_1 + 1.0) * c_inf> QL_EPSILON)
{
dRes = (-std::log(0.5 * _1 + 0.5) / c_inf);
dRes = f(dRes) / ((_1 + 1.0) * c_inf);
}
else
{
dRes = 0.0;
}
return dRes;
AND (Line 548 of AnalyticHestonEngine.cpp)
(
boost::function1<Real, Real>(
if_then_else_return ( _1*c_inf> QL_EPSILON,
bind(f,-bind(std::ptr_fun<Real,Real>(std::log), _1)
/c_inf) /(_1*c_inf),
bind(constant<Real, Real>(0.0), _1))),
0.0, 1.0);
is bascically the same as
double dRes = 0.0;
if (_1 * c_inf> QL_EPSILON)
{
dRes = (-std::log(_1) / c_inf);
dRes = f(dRes) / (_1 * c_inf);
}
else
{
dRes = 0.0;
}
return dRes;
Best Regards and thanks,
Toyin Akin.
_________________________________________________________________
Do you have a story that started on Hotmail? Tell us now
http://clk.atdmt.com/UKM/go/195013117/direct/01/
|
