package explorer - Page 333

Apache Commons Math optimization

- by alex

Does anyone have any experience with the Apache Commmons Math optimization package? More specifically, the Nelder-Mead method implementation? Is it pretty high quality?

Does oneway declaration in Android .aidl guarantee that method will be called in a separate thread?

I am designing a framework for a client/server application for Android phones. I am fairly new to both Java and Android (but not new to programming in general, or threaded programming in particular). Sometimes my server and client will be in the same process, and sometimes they will be in different processes, depending on the exact use case. The client and server interfaces look something like the following: IServer.aidl: package com.my.application; interface IServer { /** * Register client callback object */ void registerCallback( in IClient callbackObject ); /** * Do something and report back */ void doSomething( in String what ); . . . } IClient.aidl: package com.my.application; oneway interface IClient { /** * Receive an answer */ void reportBack( in String answer ); . . . } Now here is where it gets interesting. I can foresee use cases where the client calls IServer.doSomething(), which in turn calls IClient.reportBack(), and on the basis of what is reported back, IClient.reportBack() needs to issue another call to IClient.doSomething(). The issue here is that IServer.doSomething() will not, in general, be reentrant. That's OK, as long as IClient.reportBack() is always invoked in a new thread. In that case, I can make sure that the implementation of IServer.doSomething() is always synchronized appropriately so that the call from the new thread blocks until the first call returns. If everything works the way I think it does, then by declaring the IClient interface as oneway, I guarantee this to be the case. At least, I can't think of any way that the call from IServer.doSomething() to IClient.reportBack() can return immediately (what oneway is supposed to ensure), yet IClient.reportBack still be able to reinvoke IServer.doSomething recursively in the same thread. Either a new thread in IServer must be started, or else the old IServer thread can be re-used for the inner call to IServer.doSomething(), but only after the outer call to IServer.doSomething() has returned. So my question is, does everything work the way I think it does? The Android documentation hardly mentions oneway interfaces.

Search Results

Search found 14633 results on 586 pages for 'package explorer'.

Page 333/586 | < Previous Page | 329 330 331 332 333 334 335 336 337 338 339 340 | Next Page >

- by alex

- by Dan Menes

- by Kazoom

- by Bilal

- by Carlos

- by crashpoint_zero

- by C. Ross

- by Mihalis Bagos

- by Melody Friedenthal

- by alureonissue

- by Vitaly

- by Jim Linny

- by Karsten W.

- by malcolms

- by Nev_Rahd

- by HH

- by Tener

- by bigstylee

- by Mick

- by Zoltán

- by St.Shadow

- by ensnare

- by Keikoku

- by nmeth

- by dell

< Previous Page | 329 330 331 332 333 334 335 336 337 338 339 340 | Next Page >