Marcelo Forets

The `@system` (and `@ivp`) macros have been super useful and we have used them extensively (see for example the [ReachabilityModels.jl](https://github.com/JuliaReach/ReachabilityModels.jl/tree/master/src/models) library). OTOH I have often found that it is not...

The idea is that if I want that `@system(M*x'' + C*x' + K*x = 0)` define a new `SecondOrderLinearContinuousSystem`, it is sufficient to do something similar to: ```julia register(expr=:(M*x'' +...

> (e.g. allowing for equations with * and without, If this is a problem I'm fine with removing that feature (or keep it but in some other way, like with...

Reopening as there are some ideas left in this thread. The second order system construction is nice, I opened a draft PR https://github.com/JuliaReach/MathematicalSystems.jl/pull/225 . I'll play with it.

Absolutely! We can also add links to the Julia packages that are using MathematicalSystems.

The examples can contain use of the input type (from this library).

@lbenet @jonniedie you work too hard! leave something to discuss in https://juliareach.github.io/juliareach-days-3/ :smile:

I don't have further comments about the implementation, looks good to me :tada: About the evaluation of the new method, it would be interesting to run the [ARCH-NLN RE](https://github.com/JuliaReach/ARCH2020_NLN_RE) on...

I don't know for sure but maybe there is a difference only in models handling "small" numbers; in that sense, the Production-Destruction model is a good candidate.

> I've checked out the suggested for the ProductionDestrucion model and found the following thanks for checking ! > I propose to leave this PR open sounds good to me....