Dear Everyone,

as you may or may not know, the Blosc compressor has become the basis for some novel, innovative technological experiments in the PyData space. Especially the Bcolz and Bloscpack projects which provide a way to perform out-of-core computations on column based datasets have become particularly interesting for the analysis of medium-sized time-series datasets.

In this post, we would like to convince you to give us some money to foster the project, development and accelerate growth of our community. Historically, it has always been a difficult endeavour to monetize open-source development and so, below is a non-exhaustive list of potential models that we are considering:

Direct sponsoring / Donations

This involves paying either a single lump-sum or monthly installments to foster continued development and innovation. This type of sponsoring isn't bound to any specific goal or feature and would allow us for example maintain and release the projects regularly.
Feature-driven sponsoring

Paying for specific features to be implemented, bugs to be fixed or paying to have a voice when it comes to prioritizing items in the issue-tracker(s).
Hiring us as freelancers for Blosc/Bcolz projects

This means that you hire one or both of us to implement a project that uses bcolz inside your company. Any bugs we find or improvements that need to be made would flow back into the open source code-base.
Hiring us as part-time freelancers for general projects

This means you hire one or both of us as part-time freelancers for two to three days a week to work on general projects. These can be related to Python and data or open-source work on other projects. This would allow us to spend the remaining days on Blosc/Bcolz.
PhD positions

There are still a few interesting theoretical aspects to be unlocked, for example certain mathematical properties of the shuffle filter and a compressed extension of the external-memory-model (EMM) to analyse the runtime of Blosc style out-of-core algorithms and Bcolz operations in general.

We welcome any feedback regarding the above options and please do tell us about any additional models that may be interesting to us or for you.

With best wishes and looking forward to your input,

Francesc Alted and Valentin Haenel

Comments