- From: Sebastian Samaruga <ssamarug@gmail.com>
- Date: Tue, 31 Oct 2017 20:57:18 -0300
- To: cory-c@modeldriven.com, W3C Semantic Web IG <semantic-web@w3.org>, public-rww <public-rww@w3.org>, DBpedia <Dbpedia-discussion@lists.sourceforge.net>, pragmaticweb@lists.spline.inf.fu-berlin.de
- Message-ID: <CAOLUXBsE01rK6v1FqkgONKdv0EX3wfb++ptHttk0f-jvwENFVA@mail.gmail.com>
Hi. It's really encouraging to find people interested into achieving similar goals. Mostly when the whole community seems to be looking at other side and the support comes from an organization like the OMG. I'll try to respond inline each part of your message: On Oct 28, 2017 1:10 PM, "Cory Casanave" <cory-c@modeldriven.com> wrote: Sebastián, We may have some goals in common. We are currently working on a reference implementation for “Semantic Modeling for Information Federation” (SMIF), an OMG standard in progress. Here is a link to the current draft: https://github.com/ModelDriven/SIMF/blob/master/NextSubmission/ SMIFSubmissionMasterDocument.pdf I've tried to read the specification. It's really 'formal' in shape but a little bit narrow less in scope. It defines lots of artifacts and their roles in the way of 'usual' models of models approach of trying to solve anything. My approach would be to have a metamodel that could render any model (including the model itself) The foundation meta-model is designed to support a variety of modeling, data languages and technologies as well as mapping semantics. There is already one commercial implementation of the UML/OWL binding (which makes it much easier to create and reuse SW models): https://www.nomagic.com/ product-addons/magicdraw-addons/cameo-concept-modeler-plugin If a foundation meta-model is meant to be built for business integration, both in the producers side of the coin as in the consumers side everything should be homogeneus as for being able to regard resources comming or going into an operation (or exchange as I'm building an ESB) as the same kind of thing, regardless they are database records, XML files, CSV or whatever. That's why I'll be using Functional Programming for this along with Java 8. Operations include the representation side of the coin as verbs as in CRUD resorts on the formats being exchanged. We are building out the more runtime aspects of federation ( also in Java 😊). The mapping semantics that define the “pivot” through concepts to various data representations is a key feature, as are higher representation of concepts like roles and first-class relationships. We are not quite ready to post the initial implementation, the target is end-of-year. One area we would like to collaborate on is the analytics side, currently the analytics front-ends are all very SQL (or at least table or name/value) oriented. The 'pivot' representation for various integrated sources should be the ability to treat a DB, REST or whatever representation as the same without resorting in layers of layers of mappings. Federation could be thought of as if I update a (database) resource a (RESTful) resource is updated (invoked) accordingly. Analytics should come its way if 'federated' queries may be rant over 'homologated' meta-models of sources. Here I agree that there should be no mappings to the underlying sources but leverage the homogeneity given from FP. Given a functional ESB approach I'm pointing to, analytics would be pipes of transforms in exchanges. Like you, we are looking for collaborators on both the business and technical sides. It may be that the projects could be mutually supportive. Of course I'll be pleased in collaborating in this endeavor. What I'll warn you is that my 'formalism' and, even, my skills levels may be very bellow the ones beared by those actually participating. And what I could offer for now are *very fuzzy* early drafts (even ones without any layout at all) where I'm dumping thoughts regarding a future implementation: https://github.com/CognescentBI/BISemantics Best, Sebastián Samaruga. Cory Casanave *From:* Sebastian Samaruga [mailto:ssamarug@gmail.com] *Sent:* Friday, October 27, 2017 11:26 AM *To:* W3C Semantic Web IG <semantic-web@w3.org>; public-rww < public-rww@w3.org>; DBpedia <Dbpedia-discussion@lists.sourceforge.net>; pragmaticweb@lists.spline.inf.fu-berlin.de *Subject:* "Team" members wanted Team members wanted (for a project that doesn't yet exist): Hi, I'm just creating 'CognescentBI' as an organization into GitHub for initially hosting the proyect 'BISemantics' where I'll be pushing all the development efforts regarding a Business Integration platform for the Web of Data. https://www.github.com/CognescentBI/BISemantics I'm looking for collaborators, members, contributors and comments of all kind and sources. Please watch for updates or follow: http://exampledotorg.blogspot.com >From Blogspot there will be accessible a forum group link and Twitter updates from which to keep updated. For being honest, I already have no code at all. The whole project is planned to be implemented in Java and with a strong 'semantics' orientation. For now all I have is a bunch of Word documents specifying what's my 'vision' of a Semantic Web of Data. So please be patient. I'm trying to take this very slowly doing as much analysis and design as posible before actually implementing anything. So, if you dare, try to skim through the document files in the repo and, reading between lines, try to figure out what are the ingredients I'll try to mix up. I now it's cumbersome for someone to be asked to figure out what others (example: me) are trying to mean when stating something. That's my main problem when I think something I wrote is worthwhile to share with the world and I think it will be evident for others to realize what's my point. That's why I'm going 'social' into development. It's evident for me I'm needing to be clear in the most possible ways and that comments from peers are the only way to find out where I must turn out to achieve this. Best Regards, Sebastián.
Received on Wednesday, 1 November 2017 09:22:25 UTC