JANUARY 2016
What We've Done and What We're Doing
Though the defining development of this past month (and year) was the launch of Fadela v0.2.0, there have been a handful of other noteworthy developments.
We have begun to migrate DMM to a proper database management system as the sheer volume of data began to take a noticeable toll on the performance of its flat-file database. Taking advantage of the power and efficiency of SQL, the first major area of its integration has been enhancing the prioritization of playlists' massive related-video data pools. We plan to extend this integration to other areas, as well as enhance and further optimize the prioritization engine once the migration is complete.
Since last month's big launch, the Fadela project has seen extensive brainstorming and experimental proofs-of-concepts as we attempt to further enhance its functionality and user-friendliness. Though years of past iterations suggest that there are no certainties, our main area of focus has been attempting to bridge the existing architecture and the wild frontier of unstructured natural language processing – from a bottom-up approach.
Challenges and Solutions
As SQL begins to play a more prominent role in our projects and the flat-file model's shortcomings grow increasingly apparent, query efficiency may eventually become a concern. Investing in our knowledge of query optimization and other skills are key in anticipating and thus reducing risks of this nature.
Other concerns include the inadvertent introduction of “silly” bugs and underestimation with respect to the complexity of tasks at hand. We believe that time – and the many iterations it will bring – will provide for more effective assessment of the undertakings so that we do not over- (or under-) allocate resources. Meanwhile, comprehensive regression testing will bring trivial bugs to our attention in a speedy fashion, so that they are addressed while the affected areas of projects are still fresh in our minds.