A few random thoughts on my January 2014…
So it’s been a crazy busy month!
Managed to get a lot of work done in MarkLogic land. As you may have seen I released MLJS 1.2 and spoke at the MarkLogic User Group London (MUGL).
Noticed a few vaguely interesting things…
Merging of dataset types in single solution
For starters, there’s a lot of people needing to merge structured (CSV, RDBMS), unstructured (twitter, facebook, etc.) and semantic (relationships, graphs, facts) data.
I’m working on something at the moment pulling together many sources of information to provide a single browseable hub of an organisations data.
Sounds a bit of a scary prospect – but actually it’s only taking 5 days total, including ingestion, data quality, denormalisation and user interface build!
Naturally it’s so quick because the UI is built on MLJS. Managed to find a handful of bugs in MLJS during the PoC, but not really had a need to add any new functionality.
Only change I’ve made is for performance reasons. Turns out the way I’m storing tweets is fairly inefficient. Just my fault, not MarkLogic Server’s.
On the bright side, this must mean I’ve got most of the commonly needed functions of MarkLogic exposed via MLJS and it’s widgets for application builds – I’ve not had to add anything substantial.
A few goodies will wind themselves in to the Feb release now though:-
- Ability to extract metadata for search result rendering and geospatial markers drawn on a map (longitude, latitude from geo constraints)
- Suggestions / auto complete in the search bar widget (I couldn’t sleep tonight…) and search context
- MLJS Core to include the suggest() function for the above
- Vastly improved heat map display (now I’ve actually read the heatmap.js API …)
- Auto recalculation of heatmaps as you zoom/pan
Adoption of MLJS
During the MUGL meeting there were a fair few people who had used MLJS.
I also went back through my email to see how many people had asked me questions on the MLJS API. Managed to find 8 internally, 4 other internal likely users, and a handful externally.
There are also at least 10 PoCs/demos built by ourselves (MarkLogic pre-sales), partners and customers that use MLJS as the UI layer.
Even more impressively – these have all happened since the 1.0 release in October 2013!
I’m hoping the Feb 2014 release will cement this popularity thanks to it’s focus on visualisations.
I hope to provide 90% of the common widgets needed in MarkLogic systems, including the MLJS Workplace that will bring all this functionality to mere mortals via drag and drop page laying out.
Semantic technology is mainstream
For the first time we’re actively being engaged worldwide around semantic applications.
I know of customers in Media, Financial Services and Public Sector that are all wanting to use MarkLogic’s semantic capabilities to improve and simplify their application development.
A lot of this centres around modelling organisations, people, and other concepts like roles and knowledge they have. General idea being information discovery and collaboration.
Some very interesting concepts coming out, some of which I hope to share soon.
Anyways, best get off to bed… or watch a West Wing… let’s face it, it’s gonna be West Wing!