Chronological Thread 

< Chronological >      < Thread >

---

Randall Britten wrote:
> Hi Andrew
>
>
>
> Mike Cooling urged us at the last CellML meeting to support access to
> the CellML api and services from Java, since on his trip he spoke to
> parties who want to use it from their Java software.
>
>
>
> What do you think of http://developer.mozilla.org/en/docs/JavaXPCOM?
>
>
>
> Carey mentioned it to me as an option for accessing Zinc from Java.

Hi Randall,

I am not sure that we would ordinarily want to force people to go
through XPCOM to get to Java; indeed, it has been possible to go through
CORBA (which I expect would be more lightweight than including
everything needed to get the API working through XPCOM) for some time,
but this has not been a popular option either.

I think that a much better approach would be to finish the PCM (Physiome
CellML Mapping) to Java bridge, which would then be usable without
having any dependency on Mozilla (which, I expect, would be much more
practical for most Java based tools). The initial work I did on this
indicated that it shouldn't be very hard to get a basic (non-GC aware
approach, which uses Java's mechanisms for tracking references to Java
objects from native objects without any sort of cycle collection, and
lets the Java GC decide when to release native objects). That said,
getting bridges to work perfectly can be tricky so we should allow a
reasonable amount of time to sort out any issues that may arise (and so
should consider it a reasonably large implementation task).

Best regards,
Andrew

>
>
>
> Regards,
>
> Randall
>

---

• [cellml-dev] Java interface to CellML api, Andrew Miller, 05/05/2008

---

Archive powered by MHonArc 2.6.18.