Switch to Manchester Syntax

[mrnolte]

For reviews, it would be nice to have the ontologies and the changes in Manchester Syntax, which is designed to be human-readable.
It is also the Ontology Language used by Protégé, e.g., to define complex concepts. It might be neccessary to convert the Manchester Syntax to XML for online releases (via the CI).

[sasjonge]

In issue https://github.com/ease-crc/soma/issues/251we have problems with some GCI's. It looks to me, as if GCI's are deleted in the manchester syntax, so this could be a problem in this issue, too.

[mrnolte]

Uli Sattler writes that GCI's are not proper manchester syntax - I will look into that into more detail next week

[mrnolte]

Protégé tells me the same - either we resolve the GCIs, which seem to cause problems elswehere, too (#251 ) or maybe use functional syntax / turtle syntax instead? Still way easier to read than XML

[mpomarlan]

+1 for functional syntax :)

[mrnolte]

This is not closed yet! The .owl files are not in another format yet, as we need to wait for all .owl-file regarding pull requests to be completed. Otherwise, it will be a huge pain to merge these PRs.

[mrnolte]

We could still use Manchester Syntax instead of Functional, if we rewrite the GCIs - and this seems to be possible without leaving EL: for C_1 subclass C_2, where C_1 and C_2 are complex concepts, one would introduce new Concepts A_1 and A_2 and assert C_1 = A_1, A_1 subclass A_2, A_2 = C_2, which gets rid of the CI.
I would prefer this.

[daniel86]

someone also needs to make sure KnowRob can load the ontologies in the other format! might not be a lot of work, but code must be modified I am pretty sure

[mrnolte]

someone also needs to make sure KnowRob can load the ontologies in the other format! might not be a lot of work, but code must be modified I am pretty sure

This has been thought through already: We only want to switch the syntax in the git files (for readability in the git changelog / PRs), the CI will then translate them back to rdf xml. This has already been implemented here: #264

Therefore, the published files (via the URIs) will still be in rdf xml.