Microservices Matters

May 15 2015   12:47AM GMT

Developing apps for space scientists calls for automation

Jan Stafford Jan Stafford Profile: Jan Stafford

Tags:
app technical features
Application development
Cloud-Based Services
Data access
Data analysis
Data Analytics
Data management and storage
Software testing

A photograph from a Mars Rover may be breathtaking, but it will not deliver the complex data space scientists seek. Scientists like Washington University-St. Louis computer systems manager Thomas Stein need broader sets of data in formats that work with modern data analysis software. Stein helped create Analyst’s Notebook, a tool that documents geological findings from space missions and organizes that data in an online offering that is made accessible to scientists and the public.

Look at the data coming from just one instrument, say, the Mars Rover Opportunity. Some scientists are focused on a certain type of data from that one instrument, some on others. Meanwhile, said Stein, the general science community may want broader data from that instrument to do research in other disciplines. In addition, many scientists are doing cross-instrument and cross-mission searches and correlations to study a variety of topics.

“Today’s scientists cannot simply convert an image to a .JPG and use it, because you lose so much of the science quality of the data,” said Stein, who works in the University’s Department of Earth and Planetary Sciences. Analyst’s Notebook helped enable replay and archiving mission images and data, but that information must still be archived in formats accessible to scientists using many different software applications and devices.

Stein’s group works with NASA (National Aeronautics and Space Administration) to archive planetary data for the long term – as in the next 50 to 100 years. “We wanted to develop a value-added tool that helps scientists bind this data in a meaningful way,” said Stein. “By giving them data previews, we’d help them understand what they’re getting before they actually hit the download button.”

Developing software for geological studies of space rocks wasn’t Stein’s intention when he got an after-college job in in the Smithsonian Institution’s Mineral Sciences Department. Yet, it was there that he was asked to develop software for a traveling exhibit on volcanoes. The success of the three applications he delivered led to more projects for the Smithsonian.

After achieving success in these geological software projects, Washington University contacted Stein about programming software for scientists studying “space rock” data from the Giant Magellan Telescope. The immediate problem Stein addressed was a flaw in the way scientists were doing field-testing. “Nobody was taking notes about the decision-making process,” he said. “After a week of field-tests, they realized, ‘Hey, we don’t even remember why did we decided to look at this rock instead of that rock.’”

Of the many challenges for building scientific applications, two in particular really perplexed Stein and the NASA team: unpredictability of data from Rovers and feature glut.

For an orbital mission, an obvious objective is to map the planet systematically, but the Rovers don’t make this process easy, because they, well, rove. “Scientists often don’t know where the rover will drive and what it’s going to find,” said Stein. Another goal is determining the characteristics of natural objects, such as rocks. The scientists need to know where and in what context, which is hard to tell from an image. To deal with this problem, the development team used Microsoft Image Composite Editor, which was built on Microsoft SQL Server. The Editor can be used to create images that aggregate the surroundings of a finding in context mosaic image.

The feature glut issue comes from the length of today’s space missions. “Keeping up with what our users need over 10-15 years is unbelievably hard,” Stein said. “Think of how different the expectations of software users were 15 years ago – nobody asked for one-click ordering online.”

The development team, focused on Opportunity and other NASA Rovers, sought an automated development platform that set up the back end so they could devote more time to building value-added tools specific to planetary data coming from Rovers. “We shouldn’t be building basic code, laboring over documentation and doing cross-platform testing,” he said. Telerik Platform, a cross-platform development suite, was chosen to help the software teams focus on high-level challenges and bypass earlier phases of software development.

A web-based application running on the Microsoft ASP.NET platform, Telerik Platform provides a user interface (UI) that NASA uses for framework controls. In addition, Telerik’s automated test and quality assurance tools reduce the time needed to build a feature. An example is a documentation feature Stein’s team built that enables rapid online searches. “Documentation becomes very difficult when doing rapid application development and dealing with such huge sets of data,” he said. Telerik’s toolset helped him build a feature that enables a user looking for images of a certain target find it quickly online “at the push of a button, instead of the user having to do literature searches.”

Being able to react quickly to user needs is a necessity today, one that automated test and development platforms makes possible. “In reality, I’m still not a computer scientist, I’m a geologist,” he said. “A foundation development tool really helps me not worry so much about the computer science side and focus on the science side.”

 Comment on this Post

 
There was an error processing your information. Please try again later.
Thanks. We'll let you know when a new response is added.
Send me notifications when other members comment.

Forgot Password

No problem! Submit your e-mail address below. We'll send you an e-mail containing your password.

Your password has been sent to:

Share this item with your network: