Hi Sebastien,

thanks for your summary for 4.2 release. Looking at the two features which involve me, parallel and identification, I can surely commit for the mid october deadline for parallel. For identification, I am a bit worried first for a number of details that may require some further check and discussion with Nikolay, second to my other schedules.

I am not sure, also in terms of time available to development and testing in the next 4 weeks, whether it is worth rushing or it may be wiser thinking of post-pone its official release to 4.3.

Since it is a rather 'isolated' featurs, we may resume deicision on it looking at the stage of development we will have reached in mid october. What do you think?

best wishes Marco

On Thu, 16 Sep 2010 16:25:40 +0200, Sébastien Villemot sebastien.villemot@ens.fr wrote:

Hi,

I summarize in this e-mail the main release goals for Dynare 4.2, with the name of the responsible person(s) and the associated ticket numbers. Note that there are many other tickets – in particular bugs which need to be fixed – but I only list here the outstanding intended new features :

• Symbolic computation of the first order conditions of the Ramsey problem (Sébastien) [#5]

• Use of a seed to make estimation deterministic (Michel, Stéphane)

[#85]

• Estimation with unexpected structural changes (Stéphane) [#13]

• Partial information (George) [#21]

• Markov-Switching BVAR (Houtan, Michel) [#24]

• Automatic generation of the stationary model, given the trends (Houtan) [#19]

• Local identification (Marco) [#31]

• Parallelization system (Marco, Sébastien) [#55]

• Constrained parameters for OSR (Michel) [#58]

• Possibility of using the bytecode representations for most tasks (Ferhat, Sébastien) [#11]

• User guide: refresh outdated parts, add new chapter on optimal policy (Tommaso, Michel) [#17]

The goal is to release the 4.2 version by the beginning of november, which means that the various features should be ready by mid-october.

Note that by “ready” I mean that the features must be working reasonably well, have an interface in the MOD-file, be documented (at least on the wiki, preferably in the reference manual), and come with test files in the appropriate directory. Of course I can help you concerning these various aspects, in particular for the interface (which require preprocessor changes) and for the build system.

Let me know if you think that you cannot achieve completion of the features that you are responsible for by the mid-october deadline. We will then decide whether to postpone the feature for 4.3 or to delay the release of 4.2.

Best,

Hi Sebastien

I added a wiki with a manual page for PI and there are test files in the tests subdirectory.

see: http://www.dynare.org/DynareWiki/PartialInformation

I am not sure whether if and if, how to amend the html manual pages.

Could you please check if we have all we need, thogh, I may add and attach another, test results document.

Best regards George

On 16/09/2010 15:25, Sébastien Villemot wrote:

Hi,

I summarize in this e-mail the main release goals for Dynare 4.2, with the name of the responsible person(s) and the associated ticket numbers. Note that there are many other tickets – in particular bugs which need to be fixed – but I only list here the outstanding intended new features :

• Symbolic computation of the first order conditions of the Ramsey problem (Sébastien) [#5]

• Use of a seed to make estimation deterministic (Michel, Stéphane) [#85]

• Estimation with unexpected structural changes (Stéphane) [#13]

• Partial information (George) [#21]

• Markov-Switching BVAR (Houtan, Michel) [#24]

• Automatic generation of the stationary model, given the trends (Houtan) [#19]

• Local identification (Marco) [#31]

• Parallelization system (Marco, Sébastien) [#55]

• Constrained parameters for OSR (Michel) [#58]

• Possibility of using the bytecode representations for most tasks (Ferhat, Sébastien) [#11]

• User guide: refresh outdated parts, add new chapter on optimal policy (Tommaso, Michel) [#17]

The goal is to release the 4.2 version by the beginning of november, which means that the various features should be ready by mid-october.

Note that by “ready” I mean that the features must be working reasonably well, have an interface in the MOD-file, be documented (at least on the wiki, preferably in the reference manual), and come with test files in the appropriate directory. Of course I can help you concerning these various aspects, in particular for the interface (which require preprocessor changes) and for the build system.

Let me know if you think that you cannot achieve completion of the features that you are responsible for by the mid-october deadline. We will then decide whether to postpone the feature for 4.3 or to delay the release of 4.2.

Best,

---

Dev mailing list Dev@dynare.org https://www.dynare.org/cgi-bin/mailman/listinfo/dev