Hey Guys!

I am an android developer **a google created OS for mobile environments** and recently some bright sparks ported the boinc client across to Java which is the environment android uses.

Now - the demo project that comes shipped with the source is the seti@home project.

However I want to develop the app so it can support Rosetta@Home because I believe this to be a more suitable cause for the project.

Now - can anybody tell me how, where and who I should speak to - somebody at the project for "boincoid" the boinc client conversion stated to speak to the rosetta admin about getting it ported to java??

The project sourceforge page is here:

https://sourceforge.net/projects/boincoid/

Please help!

Andy

Hopefully one of the Project admins should spot this thread in the next few days and respond, but if you want to hurry things along you could try sending them a private message. I am not sure who the best person to contact is, but a few of the people you can try are:

David Baker (Project Leader)
David EK
Yifan Song

---

How 'strong' are these mobile phones?

---

Andy, I think you would want to license the source code. They will allow you to do that for academic purposes. But I believe you will find it rather difficult to squeeze the application in to a cell phone.

Here is a link to the licensing information
http://depts.washington.edu/ventures/UW_Technology/Express_Licenses/rosetta.php

At one point Dr. Baker mentioned the number of lines of C++ source code. ...was it a half million? I can't recall.

---

Rosetta Moderator: Mod.Sense

I dont need people questioning the concept - it will work and thats the reason guys have spent however long porting the boinc client and the seti project also.

Getting rosetta to work is the natural next step.

So does that mean that each line of c++ needs conversion to java for this to work - its not looking good.

Yes, I wasn't trying to dissuade you from looking further in to it. But, actually, I would think the "natural next step" would be to take a second BOINC project with a very small application, and memory footprint and convert that. And all the simpler if it is a project with open source.

You might find a project on yoyo (or elsewhere) that is already in Java as well.

Have you posted to the BOINC developer's EMail list about your interest in porting applications? ...or perhaps you were intending to more specifically do something for Rosetta. The other idea would be to find someone that doesn't presently have a BOINC project, and help them create the application. The developer's list might be a good place to find such a project. I'll bet there are many that read the posts, intending to do something some day.

---

Rosetta Moderator: Mod.Sense

Hmm - to be honest i would happily port any application which does research into cancer.

I chose rosetta but it could be anything that works alongside boinc - what is yoyo!?

Will you help me find a project to convert or import into this app!?

hmmm, is it possible to compile rosetta as a library - because i have a NDK which allows java to make calls to a .so library?!

Hopefully one of the Project admins should spot this thread in the next few days and respond, but if you want to hurry things along you could try sending them a private message.

So it took me a year to spot this one, must have slipped my searches (which naturally also diminished with time). However I/we do answer to the emails at sourceforge, and have gotten several queries regarding the project in the last year.

For record sake, I'd mention that we did contact Rosetta back then. We participated in the Google Android Contest, and quite frankly thought Cancer would be much more appealing than Aliens.

However, the amount of code, as someone mentioned in this thread, was way larger than we could cope with. We translated ~40K of C to Java, which was hard enough during our studies, but Rosetta was more on the line of ~500K, which was something we couldn't do.

For further inquiry - boincoid@gmail.com

Thanks for the kind words!

Let us know how we can help. It would be great to get R@H running on a couple million snapdragon or A4 processors. Sure these are light weight processors but when you look at the number of them available it is too many to ignore. They could be like ants working on the project.

---

Thx!

Paul

Let us know how we can help. It would be great to get R@H running on a couple million snapdragon or A4 processors...

Thanks!

Most urgent help is bringing Boincoid with the SETI project to completion. We've translated all the code, but haven't gotten around to testing all functionality yet, especially since back then we had no servers to communicate with.

Still waiting innocently to that open-source has-spare-time programmer that will come forward. As an open-source no-spare-time-whats-so-ever programmer, I know that's a long shot... :)

Cheers,
Oded.

Just a really dumb question, if you don't mind. Dumb questions are one of my specialties. Do you envision being able to run this on battery power?

Also, not knowing anything about the Snapdragon or A4 processors (other than the fact that I have an A4 in my MaxiPad) how does it compare to a "theoretical" single core X86 for throughput when doing floating point math?

Just a really dumb question, if you don't mind. Dumb questions are one of my specialties. Do you envision being able to run this on battery power?

Also, not knowing anything about the Snapdragon or A4 processors (other than the fact that I have an A4 in my MaxiPad) how does it compare to a "theoretical" single core X86 for throughput when doing floating point math?

Well, at first stage we envision running only when the phone's connected to the power outlet. And as Prof. Anderson said himself - he was first running BOINC on a 286, so (this is my addition) any complaints about the current hand-held device capabilities is just rant...

In another post I dictated a spec saying the CPU has a floating point unit.


But - CPU's will heat up, batteries will finish their charging cycle quicker, and a lot of other nuisances - but think VISION. Think FUTURE. These devices are going to be a whole lot more powerful, and probably our main computer through out life. BOINC will invade that market one day...

Cheers,
Oded.

Just a really dumb question, if you don't mind. Dumb questions are one of my specialties. Do you envision being able to run this on battery power?

Also, not knowing anything about the Snapdragon or A4 processors (other than the fact that I have an A4 in my MaxiPad) how does it compare to a "theoretical" single core X86 for throughput when doing floating point math?

Well, at first stage we envision running only when the phone's connected to the power outlet. And as Prof. Anderson said himself - he was first running BOINC on a 286, so (this is my addition) any complaints about the current hand-held device capabilities is just rant...

In another post I dictated a spec saying the CPU has a floating point unit.


But - CPU's will heat up, batteries will finish their charging cycle quicker, and a lot of other nuisances - but think VISION. Think FUTURE. These devices are going to be a whole lot more powerful, and probably our main computer through out life. BOINC will invade that market one day...

Cheers, Oded.

I would guess it would be like the gpu processing one day, a special tweaking of the Boinc program to allow it to use the hand help processor better and to its best abilities. Then you and I would select which part of each project we wanted to run, cpu on our desk, gpu on our desk or handheld device. Boinc itself would have to be smaller too though, right now Boinc is designed to be installed on a pc not a smartphone.