Re: [certi-dev] Blinding in C#

[Eric Noulard]

2011/9/11 Andrej Pancik <address@hidden>:

Hi Andrej,

Nice to hear from you again.

> Hi Niwatchai,
>  if you feel like it, you can modify Java libRTI codebase/generator to
> generate C# libRTI. That could get you at least some functionality. If you
> write more about the RTI features your federate might need I may be more
> helpful.
>  Porting Java LibRTI to C# should be pretty straightforward. Java and C# are
> very similar and basically the only parts you would have to rewrite from
> scratch are encoding and socket connection. Other than that it is more about
> search & replace.

Andrej is perfectly right the work he has done for the Java binding:
http://download.savannah.gnu.org/releases/certi/papers/Pancik2010_Thesis.pdf
http://download.savannah.gnu.org/releases/certi/papers/Pancik2010_-_Integration_of_Java_Federations_in_HLA_Compliant_Simulations_Governed_by_CERTI.pdf

could perfectly be done for a potential C# binding.
It would even be a better way to do this because it more "native".

>  Anyway, second approach is to link native code as Eric wrote. I have played
> with it some time ago and it was not as painful as one would expect.

I'll add the main drawback of the native approach is that is it may
not be cross-platform.
I.e. you'll have to have 1 version on windows and may be at least another one
on Unix-like OS.

> As far as I understand, similarly to the first approach you would have to 
> implement
> your own encoding helpers.

More or less true.
User data are not necessarily handled by the RTI so you can send even
non-encoded
data from one federate to another, the RTI does not care because it
conveys user data
as sequence of bytes.

However not encoding user data may makes you loose portability.

> Andrej
> PS: I am not sure if it helps but I would expect both approaches to take 2-3
> weekends of systematical work for person familiar with CERTI.

-- 
Erk
Membre de l'April - « promouvoir et défendre le logiciel libre » -
http://www.april.org