Semantic Progress, or lack of !
Have built a white-label location based service on top of a semantic repository (Spring, Jena, Empire, JenaBean, Sesame. MySQL)
- Enterprise level security with OpenId and Oauth (build your own app)
- ACL based authorization
- Umbel structured tagging on user profile and locations
- message and invitation queues
- personal, group, application and cloud access levels
- data cloaking
- semantic query console
Needs more work on
- JPA integration (Empire) as alternative to JenaBean
- gui clean up
- build script, maven
- hosting on GIT ?
Im going to OpenSource it, and write up my development diary – the trials and tribulations.
Iits not been easy, and taken much longer than I wanted –
- I suppose I expected too much from this early “second coming of the lord (the internet!)” and
- learning stuff just plays havoc with schedules .
The 2 biggest issues have been the bleeding edge technologies (and the number of them – there are no standard APIs yet), and performance – if you think you can get near enterprise level performance you need to think hard about what you are trying to do and what you can sacrifice. If you want ACL or logic on semantic or linked data then you have to sacrifice performance. If you want performance, dont marshal it from the text or XML you’ll get back from a sparql endpoint. But there is a third way……
So, once done, or nearly done I’ll reveal all – and the interesting conclusions.