I'm sure there are people on the mailing list that know how to configure
Visual Studio to work with ADMB.
A summary of the discussion will then be uploaded to
http://admb-project.org/documentation/installation and/or
http://admb-project.org/community/editing-tools for future reference. In
the best case, the configuration is similar between VS 2005/2008/2010,
Express or full version.
Thanks for pointing out this missing piece of info,
Arni
On Thu, 8 Jul 2010, Marc Labelle wrote:
> You + colleagues use ADMB lots these days. Anyone developing/running
> pgms from the Visual Studio 2005 IDE ?. Used to run ADMB and call the
> Autodiff library functions directly from VC6, and had it configured
> properly. But that was years ago while at NMFS/Honlab, and the Visual
> Studio interface is now very different. No problems setting it up and
> running it using the command line interface, but would rather do it all
> from the IDE, at a minimum to help me pin down where any run time errors
> comes from. Worked a bit with Dave F lately, but he dislikes IDEs, and
> we couldn't quite figure out how to set up the 'solutions' wit hte
> correct links/paths/etc. Thought you might know of a VS 2005 guru out
> there that could give me a few pointers. Couldn't get useful tips from
> JNoel or admb-project.org. A few other fisheries number crunchers around
> here are also interested in doing it all from the VS 2005-2008 IDEs.
>
> All the best.
>
> ML>
>
_______________________________________________
Users mailing list
Us...@admb-project.org
http://lists.admb-project.org/mailman/listinfo/users
Some advanced VS users can probably tell us about their experiences with
libraries, namespaces, etc., and add that to the recipe.
I agree with your comment on the importance of portability, ease of use,
and support for VS users. My own efforts have indeed focused on
portability (standard compilation scripts for all platforms and compilers)
and ease of use (IDE with a complete and automatic installer). Bill Clark
and Johnoel Ancheta did a great job porting ADMB to Mac OS, and now
OpenSolaris is supported as well. Portability and standards compliance
have been the strong suits of GCC.
Admittedly, the development core team does not include a VS expert, but I
know the ADMB community includes a number of them, who will hopefully
submit useful guidelines.
I'm also aware that you have tons of experience with the internals of ADMB
(the AUTODIF libraries) and how they are used in a large model like
MULTIFAN-CL, so I encourage you to subscribe to the user mailing list
(http://lists.admb-project.org/mailman/listinfo/users) and participate in
the discussions. It's especially valuable to hear from users who know ADMB
from the bottom up and have experience with large models.
All right, VS users, let's hear your stories.
Arni
On Thu, 8 Jul 2010, Marc Labelle wrote:
> Arni,
>
> Thanks for the pointers. Problems remain: It is not running ADMB
> templates using VC6 or later compiler versions, but finding a 'cookbook
> recipe', on setting up the IDE ocrrectly, including statements to use
> the latest Numerical Recipes routines (structures) and user-specified
> numerical libraries with no conficts, converting functions that require
> zero or unit offset matrices or ADMB specific conversion routines,
> problems with using 1+ namespaces, and etc.
>
> In the 'old days' before Visual Studio 2005 got created and the C++
> Standard did not exist, it was a simple matter of specifying #including
> "fvar.hpp", "nrutil.h", and<admodel.h>, and setting links/paths in the
> right boxes (I went through the painfull exercise on two compilers),
> cross your fingers and hope for the best.
>
> Things are a little more complicated now. I understand the admb-org
> wants to promote the use ADMB claiming it can do this and that + more,
> but the 'group' has not done enough (in my view) to make it portable and
> user-friendly for people using different but commonly used computing
> platforms + interfaces (not everyone uses gcc, old Borland compilers,
> and command line interface).
>
> So as I work on the above slowly, thought you might get others involved
> to fix big gaps.