Questions, ideas, and reaching out

[alexkreidler]

Hey there, this is an awesome project. I've just gotten a chance to check out the demo. I'd like to volunteer some my services: expertise in React, Node, Go, and more.

I have a few questions:

• What are the limitations on input files? Do they need to be generated by LaTex? (e.g. are there some embedded PDF data items that this tool relies on)
• Can you put some of the architecture into docs, so other contributors can more easily get started?

And some ideas (maybe I should add to #32):

• Work on support for other types of papers (e.g. economics, etc)
• Allow users to run their own instance, extracting data, and then contributing back to a main database (if user allows), such that extracted data can be shared. This service should probably use Semantic Search's duplicate feature to detect if users are parsing PDFs at different locations of same item. Hopefully this could be packaged into very easy to use app (e.g. user could download server, run easily, and use browser for interface, or use electron interface)
• Consider adding a "library" UI to manage multiple papers for a given user. This would probably allow better handling of annotations (🖊️ Annotations #36) and more advanced features. However, this could also expand the scope, which one should always be wary of.
• Consider integrating with citation tools like the open source Zotero JS libraries.
• Investigate performance of NodeJS + SQL backend. Definitely fine for an individual instance, but does it scale well as a hosted solution? Would Go with maybe some KV DB work better? What are common data access patterns, and how much can we cache on the client?
• Given that some of the features of this tool deal with understanding terms, Knowledge Base and Semantic Web technologies could be useful in accessing graphs or ontologies that give descriptions of such terms. I've seen a few repos in the allenai organization that deal with this, but I do have some experience here.

Please feel free to let me know if there are any more pressing issues or architectural changes that would take priority. I look forward to collaborating!