Codename One for Big Data

[Chuck Mosher]

I am new to Codename One, looks fantastic.

I've signed up for a Basic account, and enrolled in a number of the courses, thanks for the great set of lectures !

I am interested in building a data explorer app for academic and industrial research in the seismic exploration industry, which has been slow on take-up of mobile technology. The good news is that I maintain a public domain Eclipse/Java code base for managing seismic data, JavaSeis.org, with a rich set of tools for managing Big Data on Big Computers ;-)  JavaSeis powers many of the largest computing systems on the planet, one of the many little known secrets of the oil industry. Since so much capital is expended on computing, networking, and storage in our industry, JavaSeis provides a unique point of access with big iron computing vendors. Several of these leading vendors have expressed interest in providing the back end compute and storage that would be required to support an experimental marketplace for seismic data.

So, in a nutshell, I have the backend logic, vendors willing to support the backend, and a number of relatively unsuccessful web sites (e.g. http://JavaSeis.org   http://JavaSeisData.com). I also have a bunch of ugly attempts at Xcode for both Mac and iPad. As a Java developer, I was delighted to finally stumble across your product.

With that background I am looking for advice on proceeding with yet another JavaSeis experiment. Initially the main target will be academic and industrial researchers who need access to vary rare public domain datasets. I envision the JavaSeis app as a means for exploring available datasets, with the ability to download data subsets for more elaborate analysis on traditional compute servers.

So, with rose-colored glasses on, I would start by working through tutorials and examples, and then work my way up to a prototype data explorer app. My initial questions are:

1. Am I right in thinking that the Codename One platform is a good (and apparently unique) choice for an app of this type ?

2. Will Codename One be around long enough to complete what will realistically be a multi-year effort ?

3. Java has not created any performance issues for scientific computing on Intel/AMD/Linux systems - will the same be true for mobile systems ?

4. Will it be possible to access back-end services from external vendors (e.g. Amazon Cloud, Heroku, NetApp, Panasas, ... ) using the Codename One platform ?

Thanks in advance for any advice, I hope to frequent this forum - Initially as a consumer, but hopefully also as a contributor.

Best regards,

Chuck Mosher

JavaSeis.org

[Shai Almog]

Hi Chuck,
this looks very interesting.

1. I hope it is and if not we should probably improve it in those areas.
In terms of data visualization (graphs and charts) we need to improve and these are things we are targeting for a 2.1 time frame.

2. We do hope so and we have bet heavily on this. It is open source though.

3. From our benchmarks its pretty fast on iOS and beats Objective-C for some use cases. It won't be as fast as C but you can always use native code for the heavy number crunching.

4. Sure. You can use webservices freely and integrate with 3rd party libraries.

[Chuck Mosher]

Shai,

Your prompt and informative replies on this forum and the quality of your tutorials clearly support your answers.

As a side note, JavaSeis was originally developed as a joint project between ARCO (Atlantic Richfield) and Sun Microsystems at the dawn of Java back in the 90s. JavaSeis is now the base technology that powers SeisSpace from Landmark / Halliburton, the leading seismic processing system in the oil industry. Just goes to show that these type of collaborations can bring a lot of value to industry. It would be fitting to take JavaSeis to the next level in collaboration with a Sun alumnus ;-)

Chuck

[Shai Almog]

Thanks.