Clone

HTTPS: git clone https://vervis.peers.community/repos/yEzqv

SSH: git clone USERNAME@vervis.peers.community:yEzqv

Branches

master

data-publishing.mdwn

How should semantic documents, such as Idan files, be organized in this wiki and in general, and how should they be published/released?

The purpose of this project is to answer this question and create the tools needed for a solution.

What is Published

I’d like to begin by examining the kinds of object published by Kadma and Kiwi, which currently are - if I didn’t miss anything - the only projects publishing Smaoin information.

Ontologies: Definitions for describing things in specific knowledge domains
Vocabularies: Definitions of dedicated tools (e.g. Idan templates)
Domain data: Objects which are part of Smaoin and Idan, such as Idan context variables for use in templates, Unicode character names, global Boolean aliases T and F, etc.
Utility data: Objects defined using the domain tools, e.g. sets of useful basic Idan templates, English localization for Smaoin and for Idan keywords, etc.
User data: Things made by the community published in an organized manner, e.g. localizations, additional Idan templates, collection of ontologies, etc.

How to Publish

Direct download
Web rendering
Web API
Torrents
Distributed file system

Should Rel4tion have a single central distribution point, or each project can have its own files? Idea: Allow distribution packages and files to be marked as such, or somehow identified, and create a release page which collects the various releases.

But what about overall, should people upload files to a central system like Hackage for Haskell, or should there be some distributed package / search system and an installer based on collecting content from trusted friends?

Current State

For now I can just use this wiki for distribution, and make the bigger plans later when the big picture is clearer. The page [[/languages/idan] under [[/languages]] has a list of Idan files in this wiki, and the older “ttl” prefixed files. That, for now, is the way to find these files.

What I do want to take care of at this early stage is the kinds of files published. Ontologies, domain data, user data… how should all of that be organized? Should Idan provide the related files instead of Kadma, after all?

Let’s think bigger. In the future there may be many ontologies and many applications using them. But not every single application provides a whole new ontology for itself. The idea to share them and develop them together. An ontology project can either use hosting services here, or have its own website and wiki. Therefore it makes sense to have the files in separate sources!

Maybe each project can have a Data section, categorized by the different types of data. For example, there must be a clear difference between data objects which are part of Idan’s domain (e.g. the context variables), and data objects defined on top of flexible vocabularies and which are completely optional (e.g. localizations of Idan keywords).

Idea: A project will provide the domain content under one section, and any extras will either be a separate distribution and collaboration project, or under some kind of “contrib” section.