From peter at braverock.com  Fri Jun 12 15:36:24 2015
From: peter at braverock.com (Peter Carl)
Date: Fri, 12 Jun 2015 08:36:24 -0500
Subject: [Gsoc-dowd] check
Message-ID: <557AE058.5020506@braverock.com>

Hi Dinesh,

It looks like you're making good progress on the functions list - keep 
going!

Make sure that you're regularly running check against your package. That 
will help you identify issues early, potentially before they get copied 
elsewhere into your code (e.g., errors around standard documentation 
blocks).  Let me/us know if you have any questions about the build and 
check process.

How do you feel your progress is going?  Are you about where you wanted 
to be, ahead or behind?

pcc

From peter at braverock.com  Mon Jun 22 03:39:56 2015
From: peter at braverock.com (Peter Carl)
Date: Sun, 21 Jun 2015 20:39:56 -0500
Subject: [Gsoc-dowd] check
In-Reply-To: <CAAHP=VMeG_oi0rSkyTct0c-y_LEmi+JrGz4Op9EaOQvU-sUjyQ@mail.gmail.com>
References: <557AE058.5020506@braverock.com>
 <CAAHP=VMeG_oi0rSkyTct0c-y_LEmi+JrGz4Op9EaOQvU-sUjyQ@mail.gmail.com>
Message-ID: <5587676C.4030706@braverock.com>

Dinesh,

It looks to me like you're continuing to make steady progress and have 
checked in code for about 1/3rd of the functions on your plan. That's 
great - you've got good momentum at this point.

I have a couple of thoughts for you.  The first I mentioned before - 
make sure that you are incrementally running check against the package 
to make sure that it will build.  When I do that on the code you have so 
far, I'm sometimes seeing typos and minor errors that, if they are 
repeated errors will build up to be a pain to fix at some point.  
Handling that stuff incrementally is generally easier than waiting to do 
it at the end.  Maybe set aside a few hours to get the package to build 
even in it's incomplete state.

You maybe addressed some of that in the last check-in, since currently 
it fails at:

Running examples in ?Dowd-Ex.R? failed
The error most likely occurred in:

 > ### Name: ADTestStat
 > ### Title: Plots cumulative density for AD test and computes confidence
 > ###   interval for AD test stat.
 > ### Aliases: ADTestStat
 >
 > ### ** Examples
 >
 > # Probability that the VaR model is correct for 3 failures, 100 number
 >    # observations and  95% confidence level
 >    ADTestStat(1000, 100, 0.95)
Error in ADTestStat(1000, 100, 0.95) :
   Confidence Interval should be positive.
Execution halted


The second is more of a question - how are you coming with tests? I   
You might use the managers data set from PerformanceAnalytics with 
Dowd's matlab code to generate results to test against where he doesn't 
provide examples.  Like the other issue, this would likely be easier to 
do incrementally rather than having to backfill at the end (which might 
get really tedious).

Let us know if you have any questions or concerns -

pcc

On 06/12/2015 04:26 PM, Dinesh Acharya wrote:
> Dear Peter,
>
> I have completed the functions as planned, to a certain extent. I 
> could have completed more functions from Miscellaneous folder in MMRII 
> Toolbox but I got unexpectedly busy during the first week of June. I 
> hope to do more from them soon.
>
> There are some mistakes and inconsistencies in both source and 
> documentation and in some of those places, I need to do more 
> reading/research. More reading is needed particularly in cases where 
> Dowd's original code appear to have some errors (at least to me). I 
> will let you know if I am unable to sort things out myself.
>
> I will also see if there are mistakes in documentation.
>
> I have also not completed that many tests, which I will start doing, 
> where possible.
>
> In general, there are many places for me to make improvement.
>
> Sincerely,
> Dinesh
>
> On Fri, Jun 12, 2015 at 3:36 PM, Peter Carl <peter at braverock.com 
> <mailto:peter at braverock.com>> wrote:
>
>     Hi Dinesh,
>
>     It looks like you're making good progress on the functions list -
>     keep going!
>
>     Make sure that you're regularly running check against your
>     package. That will help you identify issues early, potentially
>     before they get copied elsewhere into your code (e.g., errors
>     around standard documentation blocks).  Let me/us know if you have
>     any questions about the build and check process.
>
>     How do you feel your progress is going?  Are you about where you
>     wanted to be, ahead or behind?
>
>     pcc
>     _______________________________________________
>     GSoC-Dowd mailing list
>     GSoC-Dowd at lists.r-forge.r-project.org
>     <mailto:GSoC-Dowd at lists.r-forge.r-project.org>
>     http://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/gsoc-dowd
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/gsoc-dowd/attachments/20150621/baae771d/attachment.html>

From peter at braverock.com  Mon Jun 22 15:07:04 2015
From: peter at braverock.com (Peter Carl)
Date: Mon, 22 Jun 2015 08:07:04 -0500
Subject: [Gsoc-dowd] check
In-Reply-To: <5587676C.4030706@braverock.com>
References: <557AE058.5020506@braverock.com>
 <CAAHP=VMeG_oi0rSkyTct0c-y_LEmi+JrGz4Op9EaOQvU-sUjyQ@mail.gmail.com>
 <5587676C.4030706@braverock.com>
Message-ID: <55880878.80801@braverock.com>

It's also probably worth adding functional equivalents into the 
documentation in certain places.  Your example for:

 > Ra <- rnorm(100)
 > NormalQQPlot(Ra)

... is equivalent to:

 > qqnorm(Ra)
 > qqline(Ra)

... in base R.  I prefer that the code in the Dowd package look exactly 
like you've done it - as a replication of his code - but you might also 
add pointers to comparable functions in the documentation.  You could do 
that in "See also" with links and/or in "Examples".

I'm sure there are likely a few other places that might be true. I'll 
try to let you know if I see others...

pcc

On 06/21/2015 08:39 PM, Peter Carl wrote:
> Dinesh,
>
> It looks to me like you're continuing to make steady progress and have 
> checked in code for about 1/3rd of the functions on your plan.  That's 
> great - you've got good momentum at this point.
>
> I have a couple of thoughts for you.  The first I mentioned before - 
> make sure that you are incrementally running check against the package 
> to make sure that it will build.  When I do that on the code you have 
> so far, I'm sometimes seeing typos and minor errors that, if they are 
> repeated errors will build up to be a pain to fix at some point.  
> Handling that stuff incrementally is generally easier than waiting to 
> do it at the end.  Maybe set aside a few hours to get the package to 
> build even in it's incomplete state.
>
> You maybe addressed some of that in the last check-in, since currently 
> it fails at:
>
> Running examples in ?Dowd-Ex.R? failed
> The error most likely occurred in:
>
> > ### Name: ADTestStat
> > ### Title: Plots cumulative density for AD test and computes confidence
> > ###   interval for AD test stat.
> > ### Aliases: ADTestStat
> >
> > ### ** Examples
> >
> > # Probability that the VaR model is correct for 3 failures, 100 number
> >    # observations and  95% confidence level
> >    ADTestStat(1000, 100, 0.95)
> Error in ADTestStat(1000, 100, 0.95) :
>   Confidence Interval should be positive.
> Execution halted
>
>
> The second is more of a question - how are you coming with tests? I   
> You might use the managers data set from PerformanceAnalytics with 
> Dowd's matlab code to generate results to test against where he 
> doesn't provide examples.  Like the other issue, this would likely be 
> easier to do incrementally rather than having to backfill at the end 
> (which might get really tedious).
>
> Let us know if you have any questions or concerns -
>
> pcc
>
> On 06/12/2015 04:26 PM, Dinesh Acharya wrote:
>> Dear Peter,
>>
>> I have completed the functions as planned, to a certain extent. I 
>> could have completed more functions from Miscellaneous folder in 
>> MMRII Toolbox but I got unexpectedly busy during the first week of 
>> June. I hope to do more from them soon.
>>
>> There are some mistakes and inconsistencies in both source and 
>> documentation and in some of those places, I need to do more 
>> reading/research. More reading is needed particularly in cases where 
>> Dowd's original code appear to have some errors (at least to me). I 
>> will let you know if I am unable to sort things out myself.
>>
>> I will also see if there are mistakes in documentation.
>>
>> I have also not completed that many tests, which I will start doing, 
>> where possible.
>>
>> In general, there are many places for me to make improvement.
>>
>> Sincerely,
>> Dinesh
>>
>> On Fri, Jun 12, 2015 at 3:36 PM, Peter Carl <peter at braverock.com 
>> <mailto:peter at braverock.com>> wrote:
>>
>>     Hi Dinesh,
>>
>>     It looks like you're making good progress on the functions list -
>>     keep going!
>>
>>     Make sure that you're regularly running check against your
>>     package. That will help you identify issues early, potentially
>>     before they get copied elsewhere into your code (e.g., errors
>>     around standard documentation blocks). Let me/us know if you have
>>     any questions about the build and check process.
>>
>>     How do you feel your progress is going?  Are you about where you
>>     wanted to be, ahead or behind?
>>
>>     pcc
>>     _______________________________________________
>>     GSoC-Dowd mailing list
>>     GSoC-Dowd at lists.r-forge.r-project.org
>>     <mailto:GSoC-Dowd at lists.r-forge.r-project.org>
>>     http://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/gsoc-dowd
>>
>>
>
>
>
> _______________________________________________
> GSoC-Dowd mailing list
> GSoC-Dowd at lists.r-forge.r-project.org
> http://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/gsoc-dowd

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/gsoc-dowd/attachments/20150622/b189e877/attachment.html>