Individual contributions, not (yet) integrated with project data.
Please create individual subdirectories here.
- converter platform for open source dictionaries
- partially structured content, only
- unstructured content (no detectable sense information) not parsed, but preserved as rdfs:comment of ontolex:LexicalEntry
- successfully converted to
- 52 language varieties (51 plus spoken English), 108 language pairs, 147 individual dictionaries
- quality of source data is unbalanced, extraction errors in original XDXF data not fixed
- dictionary ecosystem, providing dictionaries and a tool
- retrieved dictionaries restricted to those with explicit open license (GPL, CC-BY, etc.), also including "free-to-use" resources, no legal clearance beyond the original listing on the distributor site
- provides low-level encoding of dictionaries, with index words and definitions. No systematic markup beyond that.
- manual indication of languages.
- shallow, lossy extraction (the format semi-formalized only, no systematic encoding of grammatical features, translations, etc.), providing a selection of open/free-to-use resources that are sound (reliable translation detection, few encoding errors)
- no systematic metadata, parts of it seem to be crowd-sourced and moderate quality
- six dictionaries with online search interface and HTML serialization, mostly crowd-sourced content.
- conversion for Zazaki-German dictionary only
- unter experimental until copyright status is clarified.
For complex converter suites, we recommend the following subdirectories (non-normative).
converter source code
compiled converter source code
if you use shell scripts to preprocess or postprocess data before applying your converter, deposit them here
any libraries your project depends on. You can use maven, etc., to fetch them, instead.
the original data. If copyright-protected or license is unknown, please store it as an encrypted zip archive.
the final data, if any. If copyright-protected or unknown license, please store it as an encrypted zip archive.
Any additional documentation.