Beyond Hierarchical Data: Search and Meta Data as Fundamental Interface Elements

KDE Contributor and Developer Conference

2004 KDE Community World Summit

Speaker: Scott Wheeler

Language: English

Talk is scheduled for: Saturday, 21st of August 2004, 10:00 - 11:00

Check the Schedule for an overview of all talks during this conference.

Data and its abstractions are constantly becoming increasingly complex. Setting, files, networks layouts and online resources have traditionally been organized into similarly increasingly deep and wide hierarchies. At the same time the growing pervasiveness of computing technology is constantly producing a less technical target audience. At a certain point a threshold is reached -- traditional data hierarchies become incomprehensible to users. This talk suggests one potential approach to dealing with this issue.

A number of approaches to mitigate this problem have been attempted with varying degrees of success. Settings databases hidden from "normal" users have been invented to hide less relevant settings, most recently and most often used file lists are commonly employed. However, in the most extreme case of data explosion -- the world wide web -- search (as opposed to the hierarchical data organization of the early web or gopher) has become the de facto method for making sense of this information.

In these days of desktop systems with hundreds of gigabytes of storage, millions of files, thousands of desktop settings and local area networks with dozens to thousands of nodes, we are rapidly approaching this threshold in desktop computing. Search serves as a means of simplifying these hierarchies and presenting them organized be relevance and attempts to make this data manegable again. It has been widely successful specifically on the world wide web.

Meta data is simply what we know about a file -- in my broad definition this includes file names, mime type information, usage patterns, linkage to other data and properties stored within the file itself (i.e. the title of an HTML document). Searching needs information -- as much of it as possible; that's meta data.

KDE provides an excellent framework for putting these together. We already have mime-type recognition systems, a meta-data reading and manipulation layer, component model and interprocess communication tools to build the foundation.

We still require fast meta-data caching and indexing, better data "observation tools" (for collecting file usage patterns), and a number of library elements to make these facilities available to application developers.

In this talk I will discuss how we can use what we have, what we still need to develop and hopefully show a proof of concept implementation of some of the ideas.

About Scott Wheeler

Scott Wheeler

Scott Wheeler is a Linux specialist who has been employed in the SAP LinuxLab in Walldorf, Germany since 2002. He has been active in several areas of KDE development in the last several years, most notably multimedia. He is the author of JuK, FlashKard, KSig, TagLib and assorted hacks throughout the rest of KDE.

Suggestions and inquiries direct please to KDE aKademy Team.