search

mozillascience / code-research-object

Project between GitHub, figshare and Mozilla Science Lab.
https://mozillascience.github.io/code-research-object/
68 stars 8 forks source link

What is needed to reuse data? #14

Open codersquid opened 10 years ago

codersquid commented 10 years ago

This topic is closely related to #2.

In the context of figit and other tools, what would we ask for authors to put in a repo to get packaged up with the data? The dataprotocols project is working towards a standard data package description in data-package.json, There is some interesting discussion in https://github.com/dataprotocols/dataprotocols/issues/110 about use of data-package.json and package.jsonld.

Along with a README or other helpful files, an author could include a data package description file.

Since I think the concept will be confusing to many people, we could provide some minimum recommendations for what to include in a README as a first step, and then perhaps have pointers to more advanced methods (such as how create a package definition file).

jbenet commented 10 years ago

Hey @codersquid,

Since I think the concept will be confusing to many people,

Have you tried npm init ? It's super simple, even for people who program very little. I will be building the equivalent data package registry tool once we converge on a standard

(for a preview, see data + datadex, which are now part of the Dat Project, and will change a lot soon).

joyrexus commented 10 years ago

@jbenet wrote:

I will be building the equivalent data package registry tool once we converge on a standard

Note also existing tooling such as dpm and datapackage-json.

Sidenote: curious about the relation between datadex and the dat-registry project?