0055 - 085 8112 2187

Altavila Condominium

Simple concept, strong message and a fast turnaround.: this hotsite is created for MDMY Construtores, for a private development project in Caucaia, Ceará. It is the first in a line of sites for MDMY. The site will be expanded in the future; it will grow to maturity together with the building project itself.
See Live

Sunset Beach Hotel

A new website concept for the Sunset Beach Hotel in Cumbuco, Ceará. A big-screen design; sporting a separate kite-surfers 'landing area', with an interactive scrolling panorama to bring you as close to this kite-surfing playground as possible. On-site Facebook integration, and (in development) an online reservation system. We are showing the original design; on request we also produced a second design to cater for a warmer and more conventional feelling. We like both - you be the judge.
See Live 1st Version
See Live 2nd Version
See Live Kitesurf Landing Page

Bread & Beans

A website for a terrific restaurant-and-coffee-place in the center of cumbuco.
See Live

0031 Pousada

As the original design for this website was a good one and fit well with the house style of 0031, this website is not a new design but an upgrade of an existing one - why not. Modernized using much larger photo's and bringing the form technology up-to-date. The site is also included into our Silcre Web Authoring System now; a situation that fits better for a site that has grown to be over 100 live pages large.
See Live

Ephorus Fetch Federation

The Ephorus Fetch Federation project created a federation of web spidering services, coordinated by a centralized controller. The project can process long lists of web sites and builds an index using a logically endless number of machines to retrieve, analyse and index text data. The controller distributes work over many machines, choosing the ones that are not actively working to start up the next site to be fetched. The fetching process is self-sustaining, yet the process allows operator intervention; for instance to dynamically add filters. Because of the sheer size of the indexed data, persistence is completely distributed and 'noSQL' . This project was created using core java for the actual worker services and Servlet API 2.5 / Jersey REST for the front-end. The controller uses a multicast-based service Lookup borrowed from the Jini infrastructure; for the rest all intra-service communication, including starting and stopping, is done using REST API's. All data and indexes are stored in the Hadoop HBase noSQL database. The Fetch Federation went live in the last week of June 2011.

Ephorus Teacher-UI

The Ephorus Teacher-UI is the inbox for a teacher to monitor the plagiarism detection process for documents submitted by students. It connects to a distributed document processing pipeline. Because of the massive amounts of documents uploaded and scanned for plagiarism, the system uses a technique called 'sharding'': it handles millions of uploads and downloads and distributes document metadata over several database servers in a deterministic and intuitive manner. It uses cloud storage for actual document content. Besides being a web-ui it generates reports in PDF and statistics in Excel format. This project was created based on Servlet API 2.5; MySQL, a modified version of EclipseLink JPA to support sharding, plain SQL for performance-sensitive operations, WS-REST using Jersey, and plain Javascript and JQuery. The project is has been handed over to the Ephorus product development group for further integration with the Ephorus pipeline and to start migration and the ramp-up process.

Ephorus Search Component

This component is a simple and small ATOM feed / OpenSearch compliant search aggregator. The project supports Ephorus' plagiarism detection process: it searches content using many different search engines; internal ones and search engines external to the organisation. The component is a simple Servlet API 2.5 wep application using no database at all. In order to have maximum control, the component controls all threading concerns itself and does not rely on any type of container-supplied resource apart from the URL of the configuration. Its start configuration isolates threading issues per search engine and each search engine is configured for 1000 threads each. Configuration can be managed centrally and changes are picked up with a 10 second delay without restarting the service. The search component went live ithe 21st of september 2011