The process BitJazz Inc. has developed for inventing new technologies,
whether for our own products or for corporate clients,
consists of three major phases: specification, research, and development.
Much of this process is standard research and development protocol.
However, our method is surprisingly unique in that,
rather than approaching a problem with a box of familiar but inappropriate tools
and sort-of solving a somewhat different problem by imitating known solutions to similar problems,
we solve the real problem,
even if this means we have to develop new data-exploration tools and mathematical techniques to do so.
This unique approach typically results in the creation
of valuable innovative intellectual property for us and our clients
raising significant barriers of entry to potential competitors.
Our invention process begins with a thorough assessment of needs for the project at hand,
in which we ascertain the detailed specifications of the project.
Following the definition of the specific project,
we determine how the project fits into the whole picture,
which often permits us to find a more-efficient solution to a larger problem,
or, conversely, reduce a seemingly general problem
to a specific case that has a more-efficient solution.
We then take the unusual step of outlining the ideal solutions to the problem,
irrespective of whether those solutions are considered feasible
given current tools and technology,
since we often discover that, contrary to expectations,
the real problem is no harder to solve
than the compromise problems for which ready solutions are known to exist.
Only after describing ideal solutions do we undertake a survey of existing tools and solutions
to determine whether we can cut development time and costs
by borrowing from prior techniques.
Before we can begin empirical research on the problem,
we need to collect data
which is truly representative of the data which will be handled by the finished product.
Armed with representative data and an understanding of the ideal solutions to the problem,
we then take the unusual step of developing our own software tools for data exploration
and our own mathematical tools for data modelling.
Through our past work we have developed suites of highly sophisticated tools
which can be easily adapted or extended to suit the unique characteristics of most new projects.
In the data exploration phase, we use our proprietary software tools
to look for unexpected patterns and assess how well the data fits the expected patterns.
If necessary, we modify our software data exploration tools to accommodate new findings.
Having attained an understanding of the statistical nature of the data,
we then use our proprietary mathematical tools to develop custom models for the data
in accordance with the ideal solutions to the problem at hand.
If necessary, we extend our mathematical tools to deal with new complications.
The technology development phase begins with the creation of a fully functional prototype
that meets or exceeds the quality specifications.
The prototype is typically much slower and bulkier than the finished product,
and is often designed to run on our development platform
rather than on the target product platform.
The prototype is then applied to the research data
and the test results scrupulously compared to the research predictions.
Any discrepancies, no matter how small, are investigated
until we understand exactly why and how the prototype differs from the mathematical model.
In the refinement phase of development we optimize the technology for the target platforms,
yielding an embodiment of the technology that meets all the specifications.
Once the new technology is perfected,
we begin the productization phase.
When developing new technologies for corporate clients,
we do everything necessary to ensure that our technology works correctly and efficiently
in the product environment.