Chapter 1. Introduction

Table of Contents

1.1. History
1.1.1. About the past
1.1.2. About the present
1.1.3. About the future
1.2. Contact
1.3. About this document

YaTiSeWoBe is a pluggable framework aimed at providing groups of collaborating researchers a mean to exchange time series data, tools to visualize them, and results over a computer network in a user-friendly way.

The Neuro-heuristic Research Group [40] is a team of researchers with different cultures and scientific background, collaborating on different aspects of neuro-science, from electro-physiology to modeling.

In this international environment, the group has started several projects (like ON-ware [42]) aiming at sharing resources and knowledge through Internet.

YaTiSeWoBe is designed to gather a maximum amount of information and knowhow for team members to access and manipulate them.

Some of the requirements are:

And yes, YaTiSeWoBe stands for Yet Another Time Series Workbench!

1.1. History

1.1.1. About the past

The story started back in 1998 when Javier Iglesias was looking for some computer science project to spice his studies in biology. The association with Alessandro Villa (NHRG) and Marco Tomassini (INFORGE) ended up in a software dubbed RasterViewer. This program was available as a standalone Java application, and a simplified Java applet version used in the context of DAN [42].

RasterViewer was deployed in several groups associated to the Neuro-heuristic Research Group in the field of neuro-sciences. As the name suggests, the main purpose of RasterViewer was to display raster plots for time series data acquired during experiments in electro-physiology. Apart from being a first experience in software development for the author, RasterViewer was the opportunity to test and validate some interesting features and solutions to dynamically display and manipulate time series data on screen, among which we find the importance of good graphics and printing capabilities.

Despite being integrated in XNBC [55] as an external module for time series visualization in 2000, an important neuro-mime simulator package, RasterViewer slowly died, and is not maintained anymore.

1.1.2. About the present

The experiences with RasterViewer lead to define a list of requirements for a new application that could be able to help a team of researcher colleagues to share with the same platform data, methods and results in the general field of time series analysis, with a bias toward neuro-sciences. Given the multi cultural and diverse components of the NHRG, we had to develop portable and internationalized solutions that can be of great interest. The increasing interest for the Java language [30] from Sun Microsystems [47], with its enormous advantages for cross-platform and internationalized developments, the numerous libraries one can find in that language, the web and XML are many factors that converge and make possible the writing of YaTiSeWoBe.

With years running, the problematic of data visualization in neuro-sciences is still an issue. The more data is acquired (by recording or simulation), the more troubles we find in managing all that information with sometimes contradictory needs like having accessible and protected repositories. The lack of standardized data format is an important obstacle for data sharing, not to mention data analysis methods implementations.

By 2003, Javier Iglesias undertook a second implementation of the RasterViewer concept, based on the new developments in computer science and data management, and mainly around the web and the eXtensible Markup Language (XML) [56]. YaTiSeWoBe is a tentative to integrate the problems of a team of coworkers that share time series data. Above the graphical display of this information, we designed a modular and flexible way to handle data [Section 3.2, “Handling data”], and we are working on implementing the different propositions, even if users are not forced to share our vision. We take a particular care at leaving room for contributions from other groups, by designing modular software, and open data formats based on accepted standards like XML.

1.1.3. About the future

There is no single line of YaTiSeWoBe or this documentation that is not subject to change in the next days. The clean implementation of YaTiSeWoBe concepts alone will require man-months of work, but we also pretend reaching different goals: the acceptation of YaTiSeWoBe in our field of neuro-sciences, the cross-pollination with other fields where time series are used (like finance or meteorological forecast), …

Since almost a few years, INFORGE and NHRG have been collaborating in the seek for an open platform on-top of the web, that could help users from different cultural, scientific, … backgrounds to exchange resources: data, knowledge, computing power, … In this landscape, YaTiSeWoBe itself is only a "poor Richard's" piece of a project called ON-ware for which we developed a prototype/demonstrator [42]. If that project was able to take up, we could go further than YaTiSeWoBe by transcending fields of competence and have larger, less linked groups of users collaborating.

YaTiSeWoBe is an Open Source software released under the GPL license Section E.1, “GNU General Public License”. It is expected to take the project to SourceForge.net [46] as soon as it is in a stable state and start to attract attention from contributing users.

You can consult the Section 6.8.2, “Forseable releases” to get more details on the short term issues that the development will tackle.