When the Topaz development team first started working on the framework, they had the daunting task of consolidating an 18 month development cycle into 8 months to meet the December launch date of PLoS ONE. We were able to get PLoS ONE launched on Topaz but in doing so, the Topaz developers took some shortcuts with the framework that had unpleasant implications for future development. So they’ve been busy the last few months with a re-architecture that has finally gone live.
The 0.7 re-architecture includes the following:
- Removal of a layer of web services and increases performance. The web services were moved from Topaz (yes – we had an application layer called "Toapz") and into the PLoS ONE and Mulgara applications. This boosts the speed for storage/retrieval of objects and allows us to remove an application layer. This makes Russ very happy as it simplifies administration of the server stack and application framework (yay!).
- Speeds retrieval of objects through an Object Triple Mapping (OTM). The OTM is analogous to Object-Relational Mapping (ORM) used by Hibernate except the database is made of RDF triples instead of a relational database. This allows for retrieving collections of objects (fast) with one query instead of a single object at a time (slow).
- Adds the Object Query Language (OQL) that is similar to HQL, an object-oriented query language from Hibernate. This query language provides an "object" based query syntax and supports iTQL.
- Adds a REST based interface which will allow developers/users to directly fetch articles and annotations through READ/GET operations. This will allow developers/users to create intuitive URIs to fetch data. In the future, this will be extended to PUT/POST operations.
In addition to all of the behind-the-curtain changes, users can now score articles using three criteria:
- Insight – How thought-provoking a user found the article or how much it advances our scientific understanding.
- Style – How well performed and presented a user considers a study to be.
- Reliability – How secure a user feels the results and conclusion are in a study.
For more information on ratings, take a look at the ratings guidelines.