Database Reference
In-Depth Information
Karpištšenko:
Right now we are a small tightly-knit team of twelve people
and are looking to grow quite intensely from here. The main interest for
people who have joined the team so far is the possibility to do something
meaningful instead of working on advertising or yet another web or mobile
app. They can truly impact how humanity interacts with the environment and
the oceans. They can start solving the problems everyone thinks about in their
spare time. So that has been a great motivation.
For some, of course, there is excitement in the tremendous technical chal-
lenges necessary to make such a big data platform happen. We have to figure
out how to integrate different sensors and machine-generated data streams in
a trusted manner. We also have to make sure those data streams are acces-
sible interactively. And, of course, if you work at Planet OS for a while, you
start to learn about the planet more than you ever knew. It's an opportunity to
broaden your horizon by not only focusing on human activities and the impact
we have, but also helping to reduce the risk we create in disregarding the ocean
and the environment. Essentially, we want to build a sustainable future.
Gutierrez:
When did you realize you wanted to work with data?
Karpištšenko:
As usual, it was a progression. I've been in the software busi-
ness for 14 years professionally. In the early Internet days, I did a lot of ser-
vice development work. I used to work for a European environmental agency
that worked with various governmental organizations to develop systems to
handle information. Another example is a public transport ticketing system.
As I built these systems, I started to see some repetition in how services and
products were created.
This awareness helped to shift my interests very much toward the border
where the real world meets the software world—software embedded in
devices. As I worked on this border, I started to understand that there's so
much complexity and dynamics in the environment that it's very hard to reflect
reality well in simple software models. From this, I got very excited about
complex systems and sensor networks. I even did a bit of academic work in
that direction. Eventually, that progressed into some interesting projects with
FuturICT—a program that brings together complex-system analysts, scien-
tists, and professionals to build economic and social models.
As you work with complex systems, be it in the embedded world or with
systems like Skype, you start noticing emergent behaviors. You start to notice
that the software design, architecture, and initial assumptions you put into a
system can live a very long life and have unexpected consequences and behav-
iors years later as the system scales. To understand what was happening and
why it was happening, I was motivated to start using model-driven develop-
ment. Model-driven development allows you to you describe a system at a
higher abstraction level and then later have the compiler generate code into
byte code or assembler code. To do that, however, you need to have some
type of feedback loop. This feedback loop has to rely on data.
