Schemas.pub is a pilot project of the Data Transfer Initiative, built and hosted to promote interoperability and consistency of data exchange, especially for personal data transfer.
There are some data types we exchange over the Internet quite frequently and successfully: calendar items, tasks and contacts for example. The standards for calendar items and contacts were written in 1998, and since then, similar standards are not very common. If you take a look at those standards, they use a distinct data format, which made them a lot more work to define.
Since then, JSON and XML have made basic data formatting much more interoperable and schema languages for those formats have emerged. JSON Schema and other schema languages for JSON can quickly and easily show what some data consumer expects in a particular data file, and XML Schema Definition Language does the same for XML.
Why has the emergence of such widespread data format standards and schema languages not made it easier to define a standard? Why do we have no widely interoperable standard for "an online photo album" or "a restaurant reservation"? Instead of making it easier to define standards, XML/JSON and schema languages have encouraged the proliferation of "unilateral" data definitions -- Web APIs with schemas defined unilaterally by the designer and host of that API. Thus, pretty much every service that hosts common data (like digital photos in albums, or restaurant info and reservation data!) uses different schemas. (Sometimes parts of these APIs use schema.org, which is great and helps a lot with consistency, accessibility, and internationalization of individual fields.)
Is this really a problem? In practice consumers of APIs just translate API data into their own format, sometimes aided by schema definitions and sometimes not. Yet, we think this could be made somewhat better by allowing organizations to publish schemas that have some real thought behind them, alongside tools for using and applying those schemas, and reputational markers to show which data schemas have implementations, interoperability, open licenses or standards governance. We welcome your thoughts.
The pilot version of schemas.pub is a fraction of the functionality and registry size that we envision eventually. We're building the site as we learn about what's most useful to publishers of schemas as well as implementers searching for solid information. We think persistent URLs for schemas are difficult but useful and if your organization could use persistent URLs for stable schemas, contact us.
To get started, search for a schema you're interested in and see how it's defined, some examples, and how it's implemented. If you have a schema you think should be listed, create an account and add it.