Juergen Neumann believes that we need to organise a conference on this topic 
to bring the various stakeholders together and agree on the key foundations of our goals.
A conference similar to Volker Grassmuck's 'Wizards of OS' event. We suggest that 'OS' could stand for 'Open Semantics'. However, we are not really concerned about the title at the moment.
See Wizards of OS
Before the conference, I (Juergen) think that we should all write a joint position paper. This paper should explain the basic rules and ideas that define the characteristics of the tool, no matter how it is used. The paper could then exist separately from the implementation. For me, it would just show what we have learned so far about what a tool like this must be able to do to work as well as possible.
To determine its effectiveness, we must first define our goals and establish how we will measure them. While different stakeholders will have some common ideas and goals, they will also have some that differ. Therefore, it will be crucial to identify these similarities and differences at an early stage.
The description and development of the tool(s) should be largely independent of the planned business scenarios and usage models. For example, I (Juergen) disagree with Jack's suggestion that the 'knowledge graph' developed jointly should become the property of the platform operator. To me, that would be predatory, akin to the practices of META and ALPHABET. I believe that it (the knowledge of the world) should be in the public domain/much more part of the commons, to use Volker Grassmuck's terminology.
At this point, we don't care what kind of fantasy business anyone creates using the tool. We don't want to hear any stories about them, either, because we don't think they're important. If you create a useful, well-functioning tool, others will find many great uses for it. Our only requirement is that we all work together under a suitable free/libre open-source licence.
As I (Juergen) understand it, you (Ralf) already have a clear idea of which tools and features you want to link together and how you want to do it to achieve your goal. See Federated Lepiter
For me (Juergen), a good place to start would be to put this 'on paper'. Initially, a few bullet points would suffice, which we could then prioritise and develop together. For example:
* Typen- und Instanzen * Assocs sind Topics (First-Class-Object, Assoziatives Datenmodell) * Kognitiv adäquat * Values vs. Entities * Versionierung * Local First * Kollaborativ (Rollen- und Rechte, ACLs) * Das Selbe ist das Selbe (ist eindeutig referenzierbar, "gibt es nur einmal") * Es sollte "create", "move", "copy", "edit", "hide" und "delete" Operationen geben. * "same as", "foo" ist gleichbedeutend mit "bar" * "merge" gleicher Topics (Heuristic) * Proxy Objects als Verweis auf "externe" Objekte (Linked Data) * Import/Export vorhandener Ontologien und Instanzen * Welche Standards existieren, die wir bedienen|von denen wir profitieren können? * "UND" und|oder "ODER" Operationen müssen möglich sein (Properties!) * Facetts = Unterschiedliche Properties eines Topics, je nach Betrachter|Zweck * Parsing von Texten und automatische "Typisierung" * Inferenz / AI Assistenten * föderierbar * LATCH = Sortierung nach Location, Alphabet, Time, Category, Hierarchy * ...
We are sure you can think of other characteristics, too. In the next step, we could further define, prioritise and formulate the common points in the glossary. To ensure compatibility, it would probably be best to work in English from the outset.
~
See Modes in Monochrome Interface Elements by Tuomo. dmx