quantlib-dev — Discussion of QuantLib development

[Quantlib-dev] Want to be a developer and just submitted a patch

[Yiping C. <yp...@bu...>]

Hi guys,
    This project is interesting and I want to participate as a developer.
    I have PhD degree in computational physics and several years 
experience in Monte Carlo programming.
    Just submitted a patch adding HyperSphere salvage algorithm to 
PseudoSqrt.
   
    Best,
Yiping

[Quantlib-dev] [ quantlib-Patches-1568796 ] test program of PseudoSqrt

[SourceForge.net <no...@so...>]

Patches item #1568796, was opened at 2006-10-01 11:51
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312740&aid=1568796&group_id=12740

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Yiping Chen (champing)
Assigned to: Nobody/Anonymous (nobody)
Summary: test program of PseudoSqrt

Initial Comment:
Added HyperSphere salvage algorithm. A test program is
attached.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312740&aid=1568796&group_id=12740

[Quantlib-dev] [ quantlib-Patches-1568795 ] Patch for PseudoSqrt 2

[SourceForge.net <no...@so...>]

Patches item #1568795, was opened at 2006-10-01 11:50
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312740&aid=1568795&group_id=12740

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Yiping Chen (champing)
Assigned to: Nobody/Anonymous (nobody)
Summary: Patch for PseudoSqrt 2

Initial Comment:
Added HyperSphere salvage algorithm. A test program is
attached.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312740&aid=1568795&group_id=12740

[Quantlib-dev] [ quantlib-Patches-1568794 ] Patch for PseudoSqrt

[SourceForge.net <no...@so...>]

Patches item #1568794, was opened at 2006-10-01 11:49
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312740&aid=1568794&group_id=12740

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Yiping Chen (champing)
Assigned to: Nobody/Anonymous (nobody)
Summary: Patch for PseudoSqrt

Initial Comment:
Added HyperSphere salvage algorithm. A test program is
attached.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312740&aid=1568794&group_id=12740

[Quantlib-dev] [ quantlib-Patches-1528321 ] Business/252 Day counter

[SourceForge.net <no...@so...>]

Patches item #1528321, was opened at 2006-07-25 13:18
Message generated for change (Comment added) made by lballabio
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312740&aid=1528321&group_id=12740

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
>Status: Closed
>Resolution: Duplicate
Priority: 5
Submitted By: Piter Dias (piterdias)
Assigned to: Luigi Ballabio (lballabio)
Summary: Business/252 Day counter

Initial Comment:
Those are files to implement Business/252 Daycounter,
as used in Brazil.
I included businessDaysBetween into Calendar class in
order to have how to calculate now many business days
there are between two dates (following a Calendar).
Brazil calendar class was rewritten in order to follow
QuantLib standard. I kept backward compatibility.


----------------------------------------------------------------------

Comment By: Luigi Ballabio (lballabio)
Date: 2006-09-29 11:30

Message:
Logged In: YES 
user_id=75450

The patch was applied to the code in the cvs repository.
It will be included in next release.
Thank you.


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312740&aid=1528321&group_id=12740

Re: [Quantlib-dev] [Quantlib-users] QuantlibXL: Defining Futures other thanMAR/JUN/SEP/DEC IMM Month for RateHelper

[Toyin A. <toy...@ho...>]

Hi,

Enclosed is an updated date class to cater for all serial 3M IMM futures.

Toy out...

>From: Wilkie Lai <inf...@ya...>
>To: Ferdinando Ametrano <na...@am...>
>CC: qua...@li...
>Subject: Re: [Quantlib-users] QuantlibXL: Defining Futures other 
>thanMAR/JUN/SEP/DEC IMM Month for RateHelper
>Date: Wed, 20 Sep 2006 10:03:52 -0700 (PDT)
>
>Hi Ferdnando,
>
> > > 1.) In parameters to FuturesRateHelper() one need
> > if you provide the association between
> > non-main-cycle letters/months
> > we will extend QuantLib to support them.
>Here they are:
>Month : 1 2 3 4 5 6 7 8 9 10 11 12
>Letter: F G H J K M N Q U  V  X  Z
>
> > > In the case of IMM futures
> > > it will be the next 4 months, other currencies and
> > > markets can vary.
> > nextIMMDate now default to main cycle, it can be
> > easily extended to
> > non-main-cycle
>Without the non-main-cycle the contract series cannot
>be used as-is, since at any point of time there are
>always the main cycle contacts, combined with the
>recent month contracts.  It's always the recent month
>contacts that provide most liquidity.
>
> > > 3.) Also, the default last date of these contracts
> > as per ISDA definition Section 4.17 IMM Settlement
> > Dates are the 1st
> > delivery dates for contracts written on the
> > International Money Market
> > Section of the Chicago Mercantile Exchange. Are
> > there contract on days
> > different than the 3rd wed of the month?
>These IMM dates are correct and good for CME LIBOR
>contracts, but not for other currencies and markets.
>However when you set up term structure other than USD
>you will have to make use of such future prices as
>well.
>I think the point is there is no standalone future
>object that one can define such that these settings
>can be created at run time, at least not in
>QuantlibXL.
>What will you suggest in this case?
>
>Regards,
>Wilkie
>
>__________________________________________________
>Do You Yahoo!?
>Tired of spam?  Yahoo! Mail has the best spam protection around
>http://mail.yahoo.com
>
>-------------------------------------------------------------------------
>Take Surveys. Earn Cash. Influence the Future of IT
>Join SourceForge.net's Techsay panel and you'll get the chance to share 
>your
>opinions on IT & business topics through brief surveys -- and earn cash
>http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>_______________________________________________
>QuantLib-users mailing list
>Qua...@li...
>https://lists.sourceforge.net/lists/listinfo/quantlib-users

Re: [Quantlib-dev] [Quantlib-users] Stripping a YieldCurve with day tenor formats...

[Toyin A. <toy...@ho...>]

Hi,

Actually that will not work because the BusinessDayConvention parameter is 
not used by the Day Tenor type logic within the advance() function.

The advance() function always checks for holidays regardless of the value of 
the BusinessDayConvention parameter.

Maybe this is where the bug lies, but this is not enough...

You really want to move the date X calendar days forward and then adjust for 
a business day. Thus setting the BusinessDayConvention parameter to 
Unadjusted and actually using this value within the advance() logic still 
leaves you will a day that may land on a holiday (moving X calendar days but 
not adjusted at the end). But then again, that is the definition of calendar 
days eh?

In my case the adjustment at the end does not apply because I am moving to 
an IMM futures date.

But what about the general case?
Are there cases where such movement of calendar days and then an adjustment 
at the end required (for day tenor types)? Maybe the user under this 
scenario applies the adjustment after the calendar day movement him/herself.

Anyway, basically calendar days are not possible at the mo...

Toy out...

>From: Guowen Han <GH...@dt...>
>To: "Toyin Akin" <toy...@ho...>
>CC: 
>lui...@gm...,qua...@li...,qua...@li...,qua...@li...
>Subject: Re: [Quantlib-users] Stripping a YieldCurve with day tenor 
>formats...
>Date: Wed, 27 Sep 2006 08:41:53 -0400
>
>Hi Toyin,
>
>The default BusinessDayConvention is set to Following for the advance()
>function. If you want calendar days, you should use Unadjusted.
>For example:
>         your_target_date = calendar.advance(your_original_date, period(9,
>Days), Unadjusted))
>
>Hope it will solve your problem.
>
>Guowen
>
>
>
>"Toyin Akin" <toy...@ho...>
>Sent by: qua...@li...
>09/26/2006 11:15 PM
>
>To
>lui...@gm...
>cc
>qua...@li..., qua...@li...
>Subject
>[Quantlib-users] Stripping a YieldCurve with day tenor formats...
>
>
>
>
>
>
>Hi all,
>
>I'm trying to strip a yieldcurve with day tenor formats (ie - 9D for 9
>days).
>
>I expected this to be 9 calendar days forward with an adjustment on the
>9th
>date.
>However quantlib performs 9 market days forward.
>
>The reason I find using the day tenor format useful is in the case where
>we
>have both deposits and futures within the curve and we do not want
>deposits
>to overlap futures (ie - no deposit end date must exceed the first futures
>
>contract as we regard the futures contract more liquid.).
>
>Thus I remove all the deposit instruments that exceed the first futures
>date, however one may find an undefined gap in the curve between the last
>deposit instrument remaining and the first futures contract. I thus add a
>deposit contract starting from spot up to the first futures contract and
>the
>rate is taken as an interpolation of the original deposit rates.
>
>In my case, I currently have a deposit tenor of 153D, but quantlib will
>interpret this as market days and not calendar days. In fact it looks like
>
>the logic within the advance() method of the calendar class always
>interprets the day tenor as market days.
>
>Is it possible to have a calendar days movement option with a final
>adjustment like the Week, Month, Year tenor cases?
>
>Thanks in advance,
>Toyin Akin.
>
>
>
>-------------------------------------------------------------------------
>Take Surveys. Earn Cash. Influence the Future of IT
>Join SourceForge.net's Techsay panel and you'll get the chance to share
>your
>opinions on IT & business topics through brief surveys -- and earn cash
>http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>_______________________________________________
>QuantLib-users mailing list
>Qua...@li...
>https://lists.sourceforge.net/lists/listinfo/quantlib-users
>
>
>
>
>-----------------------------------------
>________________________________________________________
>DTCC DISCLAIMER: This email and any files transmitted with it are
>confidential and intended solely for the use of the individual or
>entity to whom they are addressed. If you have received this email
>in error, please notify us immediately and delete the email and any
>attachments from your system. The recipient should check this email
>and any attachments for the presence of viruses.  The company
>accepts no liability for any damage caused by any virus transmitted
>by this email.

Re: [Quantlib-dev] [Quantlib-users] Stripping a YieldCurve with day tenor formats...

[Guowen H. <GH...@dt...>]

Hi Toyin,

The default BusinessDayConvention is set to Following for the advance() 
function. If you want calendar days, you should use Unadjusted.
For example:
        your_target_date = calendar.advance(your_original_date, period(9, 
Days), Unadjusted))

Hope it will solve your problem.

Guowen



"Toyin Akin" <toy...@ho...> 
Sent by: qua...@li...
09/26/2006 11:15 PM

To
lui...@gm...
cc
qua...@li..., qua...@li...
Subject
[Quantlib-users] Stripping a YieldCurve with day tenor formats...






Hi all,

I'm trying to strip a yieldcurve with day tenor formats (ie - 9D for 9 
days).

I expected this to be 9 calendar days forward with an adjustment on the 
9th 
date.
However quantlib performs 9 market days forward.

The reason I find using the day tenor format useful is in the case where 
we 
have both deposits and futures within the curve and we do not want 
deposits 
to overlap futures (ie - no deposit end date must exceed the first futures 

contract as we regard the futures contract more liquid.).

Thus I remove all the deposit instruments that exceed the first futures 
date, however one may find an undefined gap in the curve between the last 
deposit instrument remaining and the first futures contract. I thus add a 
deposit contract starting from spot up to the first futures contract and 
the 
rate is taken as an interpolation of the original deposit rates.

In my case, I currently have a deposit tenor of 153D, but quantlib will 
interpret this as market days and not calendar days. In fact it looks like 

the logic within the advance() method of the calendar class always 
interprets the day tenor as market days.

Is it possible to have a calendar days movement option with a final 
adjustment like the Week, Month, Year tenor cases?

Thanks in advance,
Toyin Akin.



-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share 
your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
QuantLib-users mailing list
Qua...@li...
https://lists.sourceforge.net/lists/listinfo/quantlib-users




-----------------------------------------
________________________________________________________
DTCC DISCLAIMER: This email and any files transmitted with it are
confidential and intended solely for the use of the individual or
entity to whom they are addressed. If you have received this email
in error, please notify us immediately and delete the email and any
attachments from your system. The recipient should check this email
and any attachments for the presence of viruses.  The company
accepts no liability for any damage caused by any virus transmitted
by this email.

Re: [Quantlib-dev] [Quantlib-users] Quantlib Enum: No daily/weekly frequency

[Toyin A. <toy...@ho...>]

Hi,

These frequencies (and more!!) are currently within the dev version of 
QuantLib c++ and thus the functionality is there. However the length of time 
to export this out to Excel is another question.

Also, the new frequencies have not yet been tested enough with the dev 
library. There are portions of code within the library that only deal with 
the original frequency set (for example - code that simply divide the 
(integer) frequency by 12 to translate to a year fraction).

Furthermore, although probably not an Excel issue, there is talk that the 
implementation of these frequencies may change...

Toy out.


>From: Wilkie Lai <inf...@ya...>
>To: qua...@li...,qua...@li...
>Subject: [Quantlib-users] Quantlib Enum: No daily/weekly frequency
>Date: Tue, 26 Sep 2006 20:30:15 -0700 (PDT)
>
>Hi all,
>I have come to realize that enum Quantlib::Frequency
>does not contain frequency like weekly or daily.
>Despite the fact that they are less common there are
>existing products in the market that have to make use
>of such frequencies (eg. Overnight Index Swap).  Is
>there a rationale behind such that they are omitted,
>or would it be possible to put them back in?  I am
>mainly using the library through QuantlibXL so I
>prefer all changes to go through the code base.
>Advice is much appreciated.
>
>Regards,
>Wilkie
>
>__________________________________________________
>Do You Yahoo!?
>Tired of spam?  Yahoo! Mail has the best spam protection around
>http://mail.yahoo.com
>
>-------------------------------------------------------------------------
>Take Surveys. Earn Cash. Influence the Future of IT
>Join SourceForge.net's Techsay panel and you'll get the chance to share 
>your
>opinions on IT & business topics through brief surveys -- and earn cash
>http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>_______________________________________________
>QuantLib-users mailing list
>Qua...@li...
>https://lists.sourceforge.net/lists/listinfo/quantlib-users

[Quantlib-dev] Complex class

[Joseph W. <jo...@gn...>]

The complex and valarray would work, but I'm trying to make something 
analogous to the quantlib Array class which would share as much code as 
possible with it.  Ultimately, I'd also like to create a ComplexSampledCurve 
class, which I'll need to store characteristic functions and exponents.

Also, there aren't any licensing issues with complex arrays, but all of the 
substantial fft classes I've seen are GPL (rather than LGPL), and so I can't 
use those for quantlib.

[Quantlib-dev] Quantlib Enum: No daily/weekly frequency

[Wilkie L. <inf...@ya...>]

Hi all,
I have come to realize that enum Quantlib::Frequency
does not contain frequency like weekly or daily. 
Despite the fact that they are less common there are
existing products in the market that have to make use
of such frequencies (eg. Overnight Index Swap).  Is
there a rationale behind such that they are omitted,
or would it be possible to put them back in?  I am
mainly using the library through QuantlibXL so I
prefer all changes to go through the code base. 
Advice is much appreciated.

Regards,
Wilkie

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

[Quantlib-dev] Stripping a YieldCurve with day tenor formats...

[Toyin A. <toy...@ho...>]

Hi all,

I'm trying to strip a yieldcurve with day tenor formats (ie - 9D for 9 
days).

I expected this to be 9 calendar days forward with an adjustment on the 9th 
date.
However quantlib performs 9 market days forward.

The reason I find using the day tenor format useful is in the case where we 
have both deposits and futures within the curve and we do not want deposits 
to overlap futures (ie - no deposit end date must exceed the first futures 
contract as we regard the futures contract more liquid.).

Thus I remove all the deposit instruments that exceed the first futures 
date, however one may find an undefined gap in the curve between the last 
deposit instrument remaining and the first futures contract. I thus add a 
deposit contract starting from spot up to the first futures contract and the 
rate is taken as an interpolation of the original deposit rates.

In my case, I currently have a deposit tenor of 153D, but quantlib will 
interpret this as market days and not calendar days. In fact it looks like 
the logic within the advance() method of the calendar class always 
interprets the day tenor as market days.

Is it possible to have a calendar days movement option with a final 
adjustment like the Week, Month, Year tenor cases?

Thanks in advance,
Toyin Akin.

[Quantlib-dev] [offtopic] request for assistance developing fin. instrument with information as underlying asset

[J. P. B. <jpb...@gm...>]

Hello,
   Please forgive me for the offtopic message, but I'm hoping that
this will be of interest to people developing QuantLib.
   I'm developing software to create digital financial instruments
with information as the underlying asset (as opposed to physical
underlying assets, or promises to deliver physical assets, or other
underliers).  The software is at http://infoeng.sourceforge.net, and
the 0.2.2 version of the server (ICWS) has been released.  The first
information currency client is icsvn (currently at version 0.0.5),
with more to come.  icsvn can be used to create information currency
based on a  commit to a subversion repository, and eventually this
will be a way that developers of open-source software reap additional
benefits from their work, and better  coordinate their efforts in
collaborative development.  I've also released an Internet-draft
documenting information currency documents and operations at
http://infoeng.sourceforge.net/information-currency-rfc.txt .
   However, there is a lot of development that needs to be done before
there is a market where QuantLib developers can sell information
currency representing their work.  There's lots of software to be
developed, and then there are the commercial entities  (underwriters,
brokers, market-makers, etc.) that need to be created so that
information currency markets can be used to manage information.
   I'm writing this email to ask for help - if you are interested,
send me an email so that we can talk further, or download the software
and dive in!  People who can provide insight into the startup process
for an underwriter / broker firm would be particularly valuable.
   Thanks in advance for your help, and thanks for tolerating this
offtopic message!

   J. Patrick Bedell
   jpb...@gm...

Re: [Quantlib-dev] QuantLib-dev Digest, Vol 4, Issue 8

[Mark j. <ma...@ma...>]

surely the standard library complex class and the stl valarray class
should be sufficient and will have no licensing issues

mark


On 26/09/06, qua...@li...
<qua...@li...> wrote:
> Send QuantLib-dev mailing list submissions to
>         qua...@li...
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://lists.sourceforge.net/lists/listinfo/quantlib-dev
> or, via email, send a message with subject or body 'help' to
>         qua...@li...
>
> You can reach the person managing the list at
>         qua...@li...
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of QuantLib-dev digest..."
>
>
> Today's Topics:
>
>    1. Implementing complex arrays (Joseph Wang)
>    2. One more thing about fft and complex classes.... (Joseph Wang)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Mon, 25 Sep 2006 00:20:21 -0500
> From: Joseph Wang <jo...@gn...>
> Subject: [Quantlib-dev] Implementing complex arrays
> To: qua...@li...
> Message-ID: <200...@gn...>
> Content-Type: text/plain;  charset="us-ascii"
>
> I'm in the process of adding some functionality to deal with Levy processes
> which involves implementing fast fourier transforms and the ability to deal
> with characteristic functions and PIDE's.
>
> The question I have is that these functions require complex arrays and I was
> wondering the best way of implementing those.  The thought occurs to me that
> we can create an ArrayTemplate class and then
>
> typedef ArrayTemplate<Real> Array
> typedef ArrayTemplate<complex<Real> > ComplexArray
>
> Alternatively one can to
>
> class ComplexArray {
>    Array real;
>    Array imaginary;
> }
>
> Or perhaps
>
> template <class T>
> class ComplexTemplate {
>     T real;
>     T imaginary;
> };
>
> typedef ComplexTemplate<Array> ComplexArray;
>
> Thoughts?
>
>
>
>
>
>
> ------------------------------
>
> Message: 2
> Date: Mon, 25 Sep 2006 00:27:24 -0500
> From: Joseph Wang <jo...@gn...>
> Subject: [Quantlib-dev] One more thing about fft and complex
>         classes....
> To: qua...@li...
> Message-ID: <200...@gn...>
> Content-Type: text/plain;  charset="us-ascii"
>
> If anyone knows of any fft classes or complex classes that are BSD compatible
> let me know.  I'm having to do some wheel reinvention because all of the
> fft's I've found (namely fftw) are GPL'ed and hence incompatible with BSD.
>
> The stub for the fft was code that I'm using was explicitly marked public
> domain.
>
> I've also found that there is surprisingly little fft code that actually tries
> to do something that is object oriented.  Most of the code I've seen is in C
> or Fortran and is very procedural.
>
> Once I get a basic FFT working, the next step is to start implementing PIDE
> code, and to do some experiments with non-uniform FFT's.
>
>
>
>
>
> ------------------------------
>
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys -- and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>
> ------------------------------
>
> _______________________________________________
> QuantLib-dev mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-dev
>
>
> End of QuantLib-dev Digest, Vol 4, Issue 8
> ******************************************
>


-- 
Assoc Prof Mark Joshi
Centre for Actuarial Studies
University of Melbourne
My website is www.markjoshi.com

[Quantlib-dev] One more thing about fft and complex classes....

[Joseph W. <jo...@gn...>]

If anyone knows of any fft classes or complex classes that are BSD compatible 
let me know.  I'm having to do some wheel reinvention because all of the 
fft's I've found (namely fftw) are GPL'ed and hence incompatible with BSD.

The stub for the fft was code that I'm using was explicitly marked public 
domain.

I've also found that there is surprisingly little fft code that actually tries 
to do something that is object oriented.  Most of the code I've seen is in C 
or Fortran and is very procedural.

Once I get a basic FFT working, the next step is to start implementing PIDE 
code, and to do some experiments with non-uniform FFT's.

 

[Quantlib-dev] Implementing complex arrays

[Joseph W. <jo...@gn...>]

I'm in the process of adding some functionality to deal with Levy processes 
which involves implementing fast fourier transforms and the ability to deal 
with characteristic functions and PIDE's. 

The question I have is that these functions require complex arrays and I was 
wondering the best way of implementing those.  The thought occurs to me that 
we can create an ArrayTemplate class and then

typedef ArrayTemplate<Real> Array
typedef ArrayTemplate<complex<Real> > ComplexArray

Alternatively one can to 

class ComplexArray {
   Array real;
   Array imaginary;
}

Or perhaps

template <class T>
class ComplexTemplate {
    T real;
    T imaginary;
};

typedef ComplexTemplate<Array> ComplexArray;

Thoughts?

[Quantlib-dev] I find bug ql/PricingEngines/Lookback/analyticcontinuousfixedlookback.cpp

[killereco <kil...@gm...>]

I find bug "ql/PricingEngines/Lookback/analyticcontinuousfixedlookback.cpp."

solaris 10
g++ 3.4.3

I see compfile error When comaile quantlib..

analyticcontinuousfixedlookback.cpp:130: error: expected 
primary-expression before numeric constant

That source is this

      Real SS = underlying()/minmax();

I guess that "SS" is aleady define in gcc 3.4.3 at solaris.
So I Change 'SS' -> 'lSS'

[Quantlib-dev] Fast Fourier Transform conventions

[Joseph W. <jo...@gn...>]

I've started coding up an FFT algorithm in QuantLib, and I was wondering what 
are the standard conventions in finance for the constants (i.e. factors of 
pi) in front of the transforms.

[Quantlib-dev] actact (Bond) yearFraction() function...

[Toyin A. <toy...@ho...>]

Hi All,

I've been playing with some of the Daycounter classes, viewing their results 
and I believe I'm getting some strange numbers with the actact (Bond) 
(ActualActual(ActualActual::Bond)) class.

Basically, when building a FixedCoupon schedule (for example), apart from 
the first and last coupon periods (which may be stubs), the cashflows 
in-between will use the same start/end dates for the refStart/refEnd dates.

Now given a period such as 7/6/2002 -> 21/12/2002, the actact (Bond) will 
compute a yearFraction value of 0.5

However, it will also compute a value of 0.5 for any start date in the range

7/6/2002 - > 6/7/2002

with the same EndDate : 21/12/2002.

Within my (hacked) environment, when I pass in the unAdjusted schedule dates 
to the refPeriod dates. I obtain a more likely (to me) set of results.

Basically, is the correct operation of the actact (Bond) dayCounter, given 
the above set of dates, correct? Does it depend on the market (ie Bond 
cashflows vs swap fixed leg rate cashflows).

I'm wondering whether the other Daycounter's also have differences in output 
if it is correct to utilise unAdjusted coupon dates for the refDates.

Thoughts...
Toy out.

Re: [Quantlib-dev] memory access violation in ql/MonteCarlo/lsmbasissystem.cpp

[Klaus S. <kla...@fr...>]

Hi Nando,

thanks for the hint. I guess that the template metaprogramming I used here to 
speed up the code is causing the "compiler-headache" even though the code 
works with various g++ versions (3.3, 4.0 and 4.1) and VS 8 Express. Anyway, 
I've replaced the code by a more conventional implementation and that should 
work for VS 7 as well. (Unfortunately I don't have access to VS 7 to check 
it). 

cheers
 Klaus

On Monday 11 September 2006 11:37 am, Ferdinando Ametrano wrote:
> Hi Klaus,
>
> the test-suite compiled with Visual Studio 7 crashes in the basket
> option test (memory access violation).
>
> Compiling in debug mode results in a run-time Debug Error reported at
> line 62 of ql/MonteCarlo/lsmbasissystem.cpp: "Run-Time Check Failure
> #0 - The value of ESP was not properly saved across a function call.
> This is usually a result of calling a function declared with one
> calling convention with a function pointer declared with a different
> calling convention."
>
> Why this error doesn't show up in VC8 escapes me...
>
> Could you please check what is going wrong?
>
> thank you
>
> ciao -- Nando
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job
> easier Download IBM WebSphere Application Server v.1.0.1 based on Apache
> Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> QuantLib-dev mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-dev

Re: [Quantlib-dev] Period class... some minor updates...

[Ferdinando A. <na...@am...>]

Hi Toyin

> Currently I am in need of the new frequency types and thus have integrated
> this into my framework, however if, as you say, there is a bigger picture of
> using the Period class everywhere, then I'm happy just to keep this local
> until such time that the Period class is replicated everywhere.
I did the same with some Frequency I needed, but I hope to remove them
from next release, or at least they will born deprecated.
if this won't happen I'll include your new Frequencies too.

> By the way, the new (work in progress) schedule class doesn't work with the
> daily frequency type (the constructor that takes the boolean endOfMonth
> parameter). It generates repeated daily dates...
thank you for pointing that out

ciao -- Nando

Re: [Quantlib-dev] Period class... some minor updates...

[Toyin A. <toy...@ho...>]

Hi Nando,

I agree with you that the Period class is much more flexible and powerful.
I can see this within the new Schedule constructor(s).

Currently I am in need of the new frequency types and thus have integrated 
this into my framework, however if, as you say, there is a bigger picture of 
using the Period class everywhere, then I'm happy just to keep this local 
until such time that the Period class is replicated everywhere.

In actual fact, I use the Frequency enum type as a placeholder for the user 
to easily identify a period of time, but internally I convert this to the 
Period type and then call the new Schedule constructor.

By the way, the new (work in progress) schedule class doesn't work with the 
daily frequency type (the constructor that takes the boolean endOfMonth 
parameter). It generates repeated daily dates...

Toy out.

>From: "Ferdinando Ametrano" <na...@am...>
>To: "Toyin Akin" <toy...@ho...>
>CC: lui...@gm..., qua...@li...
>Subject: Re: [Quantlib-dev] Period class... some minor updates...
>Date: Tue, 12 Sep 2006 09:47:28 +0200
>
>Hi Toyin
>
>it hasn't been discussed yet, but I'm actually in favor of deprecating
>(and then removing) Frequency adopting instead a more pervasive usage
>of Period. Frequency has been in QuantLib before Period, but nowadays
>I cannot see any good reason for retaing it, as it is less flexible
>than Period. The only plus might be the "Once" Frequency which doesn't
>have a clear equivalent Period.
>
>There is already ongoing work in QuantLib to allow Period as
>alternative to Frequency.
>
>As for your specific proposal: once we introduce EveryFifthWeek and
>EverySixthWeek, why not introduce EverySeventhWeek, etc?
>This is one of the reason I prefer Period which doesn't have this 
>problem...
>
>look forward to feedback
>
>ciao -- Nando
>
>On 9/12/06, Toyin Akin <toy...@ho...> wrote:
>>
>>Hi Luigi,
>>
>>Is it possible to add the following additional Frequency types to 
>>quantLib?
>>(Assuming that the implementation is not too shabby)
>>
>>I've updated the files based on the most recent version within CVS.
>>
>>Thanking you in advance...
>>
>>Toy out.
>>
>>
>>
>>-------------------------------------------------------------------------
>>Using Tomcat but need to do more? Need to support web services, security?
>>Get stuff done quickly with pre-integrated technology to make your job 
>>easier
>>Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
>>http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
>>
>>_______________________________________________
>>QuantLib-dev mailing list
>>Qua...@li...
>>https://lists.sourceforge.net/lists/listinfo/quantlib-dev
>>
>>
>>
>>

Re: [Quantlib-dev] Period class... some minor updates...

[Ferdinando A. <na...@am...>]

Hi Toyin

it hasn't been discussed yet, but I'm actually in favor of deprecating
(and then removing) Frequency adopting instead a more pervasive usage
of Period. Frequency has been in QuantLib before Period, but nowadays
I cannot see any good reason for retaing it, as it is less flexible
than Period. The only plus might be the "Once" Frequency which doesn't
have a clear equivalent Period.

There is already ongoing work in QuantLib to allow Period as
alternative to Frequency.

As for your specific proposal: once we introduce EveryFifthWeek and
EverySixthWeek, why not introduce EverySeventhWeek, etc?
This is one of the reason I prefer Period which doesn't have this problem...

look forward to feedback

ciao -- Nando

On 9/12/06, Toyin Akin <toy...@ho...> wrote:
>
> Hi Luigi,
>
> Is it possible to add the following additional Frequency types to quantLib?
> (Assuming that the implementation is not too shabby)
>
> I've updated the files based on the most recent version within CVS.
>
> Thanking you in advance...
>
> Toy out.
>
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
>
> _______________________________________________
> QuantLib-dev mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-dev
>
>
>
>

[Quantlib-dev] Period class... some minor updates...

[Toyin A. <toy...@ho...>]

Hi Luigi,

Is it possible to add the following additional Frequency types to quantLib?
(Assuming that the implementation is not too shabby)

I've updated the files based on the most recent version within CVS.

Thanking you in advance...

Toy out.

[Quantlib-dev] memory access violation in ql/MonteCarlo/lsmbasissystem.cpp

[Ferdinando A. <na...@am...>]

Hi Klaus,

the test-suite compiled with Visual Studio 7 crashes in the basket
option test (memory access violation).

Compiling in debug mode results in a run-time Debug Error reported at
line 62 of ql/MonteCarlo/lsmbasissystem.cpp: "Run-Time Check Failure
#0 - The value of ESP was not properly saved across a function call.
This is usually a result of calling a function declared with one
calling convention with a function pointer declared with a different
calling convention."

Why this error doesn't show up in VC8 escapes me...

Could you please check what is going wrong?

thank you

ciao -- Nando