Making it easier to create Atomic Data

[joepio]

Creating Atomic Data is easy to do if you use the tools I've built, but if you want to implement (part of) the spec yourself theres quite a bit of challenges to overcome:

• Hosting the data properly (resolving Subject URLs with JSON-AD) is not always an option. Users should have some altertnative
• Creating Commit resources is difficult, and signing these even more so. Without a library (@tomic/lib for js, or atomic_lib for rust) you're in for a deep dive.

Solutions

• Tables should provide a standard JSON endpoint to which users can POST resources. #37
• Easy Atomic Export Protocol is all about sharing without actually minting subject URLs or signing Commits. Similar to RSS / Atom feeds, but more flexible. The client needs to create the identifiers and host it somewhere. #93
• Importer tools: Atomizer will be a piece of software that helps with converting data to atomic data. It's all about one-time converting without (manually) minting subject URLs or signing Commits - and without the complications of implementing and hosting an Atomic Data Export Protocol endpoint (which is optimized for an ongoing subscription model and a massive waste of time for a one-time conversion)
• Use git to manage data! Many people do this for blogs, and it also makes sense for ontologies. #97
• Find existing Ontologies, Classes and Properties more easily
• ? Share your thoughts below!

[jonassmedegaard]

• Atomizer is all about one-time converting without actually minting subject URLs or signing Commits - and without the complications of implementing and hosting an Atomic Data Export Protocol endpoint (which is optimized for an ongoing subscription model and a massive waste of time for a one-time conversion)

edit by joepio: added to list

[hoijui]

This is a bit over my head still, but if I think of how I'd like to use AD with as little effort as possible, I think: I host my JSON-LD serialized AD data in a github script, and on each commit (that changes the data file) I send it to some API on atomicdata.dev with curl or wget, where it gets hosted. I then optionally direct a PURL for my ontology to the URL where atomicdata.dev serves my data. ahhhh ... This workflow would be for my ontology. I guess you are rather talking about the data, right? :D

[joepio]

@hoijui Interesting, yes, using GitHub as versioning / state management system. JAMSTACK style! I've added it as issue in #97