Hi Fabian, all, I see different arguments here, I would like to comment on them separately.
1. it is hard to maintain a large codebase. This is in general true. It is not clear to me which have been the features that have not been implemented because of the kres components. The ontonet and reasoners components are stable, there have been not new features implemented, true, because nobody needed new features (me included), as far as I know. Rules have several open issues that have not been solved, I don't have nothing against moving it away from the supported codebase (that would require to remove some features from reasoners, but it can be done easily). Is there any modification to commons that have not been implemented because of the fallback on the kres components? Personally, this problem should be discussed when we have a clear issue and see if there is anybody in the community that wish to cover the changes needed on this part of the codebase (I am one of them) instead of assuming that nobody would do that. In other words, if there is no development (evolution) on some part of the code base this doesn't mean that it cannot be maintained (there is nobody that is maintaining it). 2. Set a clearer focus for Stanbol Here the problem looks different. I argue that there is the feeling that those modules are not really needed by the community and that the community would like to not have them at all in order to have a clearer view on what Stanbol is. Let me try to explain why those components are there, in my opinion. Stanbol is a framework for providing semantic capabilities to CMSs. It uses RDF as the reference model for data representation. The kres part have been designed to allow the usage of ontologies, rules and reasoning for the interpretation of RDF data according to some application's task. The big picture would be: enhance content (get some RDF from a content), bind that content to a conceptual representation (ontology), infer some new data (rules, reasoning). Possibly most of the users don't use ontologies or rules for achieving their goal. I do. Is this big picture that is in question? If yes, I would like to discuss it (and to accept fairly to be in the minority ;) ) Finally, I want to clarify that I do not have any personal motivation for obstacle your proposal, I just want the community to make a clear statement about that, and help as I can to have a better understanding of this interesting issue. Bests, Enrico On 27 May 2013 08:43, Fabian Christ <[email protected]> wrote: > Hi, > > the discussion came up that we have to maintain a large code base in > Stanbol. > > This has historical reasons as a lot of stuff was added to Stanbol > during the funding period of the IKS project. Now that IKS has ended > we see that development is still in progress for the enhancer but less > is happening in other parts. > > When trying to explain what Stanbol is, I try to separate between > Enhancement services and Knowledge Representation and Reasoning (KReS) > services. > > Enhancement: > - Enhancer + Engines > - Entityhub > - Contenthub > > KReS: > - Ontologymanager > - Reasoner > - Rules > > A large code base with Enhancer and KReS has the downside that > changing something in commons could affect many components that need > to be changed and tested. Keeping all those components in compatible > sync is a time consuming maintenance act. With less active > developments in some parts this work can hardly be done. > > Perhaps it is time to set a clearer focus for Stanbol. For most people > this would be the Enhancer and I would agree with this. As a > consequence we would need to reduce the number of maintained > components in Stanbol. This number can of course be increased at any > time again when there are active maintainers. Technically this would > mean to move the stuff that is not maintained at the moment to > '/contrib'. > > What do you think? > > -- > Fabian > http://twitter.com/fctwitt > -- Enrico Daga -- http://www.enridaga.net skype: enri-pan
