EMF-IncQuery project to push forward new EMF API
A new proposal has surfaced at Eclipse – IncQuery is set to revitalise EMF for scalability and big data modelling challenges
With no less than six EMF subprojects set to be part of next month’s Eclipse Juno Release train, you’d think that there wouldn’t be any time to innovate and add extra functionalities to the modelling framework.
But already plans are afoot to push Eclipse’s modelling framework even further, with a new subproject proposed entitled ‘EMF-IncQuery’, which is entrusted with revitalising the framework with a brand new API for querying EMF models.
Numerous challenges to the current EMF set-up have become apparent, namely these three noted in the proposal itself:
- Simple backward getters are missing from the EMF API, which are essential to modelling applications.
- No standard way to define complex queries that concern multi-object configurations.
- Existing EMF query technologies are not optimized for in-memory scalability, and thus have performance issues when used with complex queries and large instance models.
All three are understandably important should the EMF framework want to move forward. IncQuery will be a lightweight framework hopefully eradicating these problems. Efficient support for evaluating backward getters will be provided from the new IncQuery API, plus high-performance incremental query engines to aid complex queries. Finally the project will provide a user-friendly, declarative graph pattern-based query language and tools for easy definition of queries, and a DSL too. This will based on Eclipse’s framework language helper Xtext. The scope is fairly wide at this point, with a lot of emphasis placed upon IncQuery to push the modelling framework moving forward.
Below is an architectural overview for IncQuery:
Technically speaking, the query evaluator based on a graphic pattern matching engine that makes use of the RETE algorithm, whilst the tools will be packaged as an Xtext-based editor with an development environment, making on-the-fly requests possible.
The initial contribution, expected tomorrow, from the Budapest University of Technolgy and Economics will come in the form of plug-ins and core components, before a first build in June. We’ll have to wait a while yet for EMF IncQuery 1.0, as that is expected to land in June 2013. But it’s great to see a project aware of its shortcomings and seeking to erase them early. With IncQuery, speedy querying and a future of scalable robust EMF models will be realised. Check out the full proposal here.