subjective_notes.md

General Information

This file serves to publicly note some subjective thoughts which currently do not have a better place to live.

Possible Issues for other Projects:

This section collects observations that I perceive as classical issues for a software project, i.e. things that could be improved.

Owlready

• There should be an official issue tracker
• The code should be automatically checked by some continuous integration service
  • Potential problems waste of energy and ressources through cloud services.
  • Potential solution: configure CI such that only selected branches are tracked, like in .drone.yml in yamlpyowl.
• Change the extension handling in setup.py. Reason: The usage of cython in setup.py is not recommended by the cython documentation.
• Disable the "owlready2_optimized"-warning in driver.py. The user should be warned during installation but not on every use of the library. Also the doc says: "its use is entirely optional", thus a warning every time seems inappropriate.
• Change the representation of owlready2.Thing and its subclasses. Currently this reads like repr(owl2.Thing) → <<property object at 0x7fe8a3417040> 'Thing'>. However "object property" IMHO has a different meaning in the context of OWL, see e.g. owl2-primer.
• Check if newer versions of Reasoners are available. At least Pellet seems to be outdated.
• Call JAVA-Reasoners directly instead of using subprocess. This might be relevant: https://www.py4j.org/index.html
• Update docs https://owlready2.readthedocs.io/ to current version.
  • Ideally, docs should be built automatically when a new version is released.
  • Also, docs should be available for past versions too.
• docs on custom rendering should provide information how to restore the original behavior
  • this can be done with set_render_func(default_render_func), see also https://github.com/Cognonto/CWPK/blob/master/cwpk-29-extracting-object-data-properties.ipynb

404 – A General Impression of the Semantic Web World

Reproducibility and Comprehensibiltiy seems to be a core goal of the whole semantic web idea. However, I discovered a surprisingly high percentage of orphaned projects. An incomplete list (compiled since spring 2021):

• http://www.semanticweb.org/ (time out)

• https://www.semanticdesktop.org/

  • links to: http://www.semanticweb.org/wiki/Semantic_Desktop (time out)

• https://www.w3.org/2001/sw/wiki/Neologism

  • links to: http://neologism.deri.ie/ (time out)

• https://bioportal.bioontology.org/projects/TODE

  • links to: http://ontologicaldiscovery.org/ (time out)

• https://lod-cloud.net/dataset/msc

  • links to: http://msc2010.org/mscwork/
    • links to: http://4store.org/trac/wiki/SparqlServer (404)
    • links to: (404)
      • https://lod-cloud.net/dataset/fu-berlin-eurostat
      • https://lod-cloud.net/dataset/bfs-linked-data
  • links to: mailto:[email protected] (automatic response: unreachable)

• https://github.com/bfo-ontology/BFO/wiki

  • links to: http://ifomis.uni-saarland.de/bfo/discussion (404)

• many of the publicly listed wikibase instances (see also here) are not available (either 404 or 503 problems).

• https://www.isko.org/ko.html (International Society for Knowledge Organization) -> KOS registry -> https://bartoc.org/

  • https://bartoc.org/vocabularies?search=engineering
    • "Sorry, we have some technical problems! Please contact the administrator and include the error message below." (no error message shown)

• https://en.wikipedia.org/wiki/UMBEL → http://umbel.org/ → Online Casinos in Australia

• http://www.visualdataweb.de/webvowl/ Operation stopped in Oct. 2021