About this Research Topic
As science is ever-increasingly dependent on computing, compute platforms themselves are emerging as the new substrate of innovation. Beyond the conventional "desktop" paradigm lies new domains that might dwarf anything researchers typically have available on premise. Nonetheless, programming beyond the desktop is complex and less understood by scientific programmers. Awareness of these platforms is vital to unlocking new discoveries, in particular awareness beyond commercial and private interests. Amazon Web Services, for example, is not the only solution for science. Against this backdrop, this Research Topic focuses on the state of the art of open-source platforms and compute in the early 2020s in several key fields: neuroscience, bioinformatics, genomics, and analytical science. Our goal is to provide an overview specifically of open source solutions and approaches to this new frontier and provide readers with information to choose solutions that can better enable their research.
Most technical/scientific researchers are more interested in exploring new algorithms than engaging in software engineering. Using a Jupyter notebook to hack an algorithm is one thing - - deploying a fully fledged program that is cloud-ready is something else entirely. Paradoxically, these software engineering tools and concepts are key to the sharing of numerical experiments and replication of results that underpin science. While many (and counting) commercial and private companies provide clouds and services, the use of this in science is limited. Partially this is cost-related, partially this is complexity-related. Mostly this is because researchers might simply not know what is available. In fact, the scientific community has already come up with several solutions for compute platforms. While many of these are general purpose, the practical reality is that most are deployed in a very domain-specific way. A platform used in genomics might work just as well in image processing for example. The core purpose of this Research Topic is to provide a view- from- above of the state of scientific compute platforms in the early 2020s. The hope is to cross-fertilize the usage of existing platforms more widely.
Ideally, a researcher who has some data and some algorithm that consumes this data should be able to consult this Research Topic and learn/discover what tools and platforms are available for their use. Contributors that are part of a team with an existing scientific compute platform (and here platform is taken to mean any abstraction that allows other researchers to manage data and execute some mapping on this data) are welcomed to submit an overview of their work. The goal is both to provide a catalog of existing platforms and also to educate researchers about the use of these platforms and tools. Particular attention should be applied to the _ generalization_ of a tool or platform beyond the domain in which it might have been developed or deployed. Submissions should not be user manuals or guides, but rather a presentation of platform scope, design, and applicability to a scientific problem.
Most technical/scientific researchers are more interested in exploring new algorithms than engaging in software engineering. Using a Jupyter notebook to hack an algorithm is one thing - - deploying a fully fledged program that is cloud-ready is something else entirely. Paradoxically, these software engineering tools and concepts are key to the sharing of numerical experiments and replication of results that underpin science. While many (and counting) commercial and private companies provide clouds and services, the use of this in science is limited. Partially this is cost-related, partially this is complexity-related. Mostly this is because researchers might simply not know what is available. In fact, the scientific community has already come up with several solutions for compute platforms. While many of these are general purpose, the practical reality is that most are deployed in a very domain-specific way. A platform used in genomics might work just as well in image processing for example. The core purpose of this Research Topic is to provide a view- from- above of the state of scientific compute platforms in the early 2020s. The hope is to cross-fertilize the usage of existing platforms more widely.
Ideally, a researcher who has some data and some algorithm that consumes this data should be able to consult this Research Topic and learn/discover what tools and platforms are available for their use. Contributors that are part of a team with an existing scientific compute platform (and here platform is taken to mean any abstraction that allows other researchers to manage data and execute some mapping on this data) are welcomed to submit an overview of their work. The goal is both to provide a catalog of existing platforms and also to educate researchers about the use of these platforms and tools. Particular attention should be applied to the _ generalization_ of a tool or platform beyond the domain in which it might have been developed or deployed. Submissions should not be user manuals or guides, but rather a presentation of platform scope, design, and applicability to a scientific problem.
Keywords: Compute platforms, containers, open source, data sharing
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
