First things first…
Zeitgeist never intended to replace the hierarchical file-system. It is not doable. And it cannot replace the classic file browser either. Zeitgeist is more or less a layer over the file hierarchy to reduce browsing into folders etc. It is just an alternative experience of easily finding your files whatever type they may be.
This being said…
Zeitgeist is based on and works with chronologically logged data of the users working with their personal computer.
Our focus currently is on supporting activity (files,websites,notes and mails) management by tagging (manually and automatically), chronological organizing and bookmarking. This is all provided by the current UI plus searching. An RDF model will also be developed to provide for interoperability. This way, Zeitgeist will reduce and ease the experience for going back to information already used thus orienting oneself in one’s information, resuming work and task switching.
This is already tackled by common desktop search engines, but they rather fail to find and present the user’s information in the context of their, possibly repeated, usages and exploit the thereby implicitly or explicitly established relationships.
The journaling data can be understood as providing extensions to or a generalization of the “recently used” information access method common to most operating systems, where extension is meant time‐wise and spans potentially all data types and information which one touches on one’s desktop, including, for example, text documents, pictures, web resources, instant and email messages, but also contact information, calendar items and other information items related to planning.
As opposed to the intimacy expressed by many people, and in particular knowledge workers, interacting with their personal computing devices throughout the everyday and across private, professional and educational domains, today, it is striking to observe that this very personal computing environment is not really prepared and in fact offers only limited support in answering the central questions of orienting oneself: “What did I do?” (retrospective perspective), “What am I currently doing?” (current perspective, where ‘current’ turns out to be a very vague term), and “What have I planned for the future?” (prospective perspective).
Now the fun part…
The engine (in python 
) is stable, we have a D-Bus interface working (thanks to Siegfried), running constantly for one day I have around 18 MB memory usage (logging and sending to UI via D-Bus). There is the classic Journal UI and a new “Project Viewer” based on project Gothenburg concept.
We fixed a lot of bugs and memory leaks and still on it but not many new features will be coming soon. Jason Smith is working on a new project in Mono that uses the Zeitgeist engine.
A first release could be soon. Just some fixes and tweaks…
Is there a possibility of seeing the zeitgeist engine integrated as a view mode on nautilus? While it’s not a replacement for hierarchical file-systems it can be a useful complement para nautilus…
Agreed. This is especially the case if you want to limit your history/search to a particular directory or subdirectory rather than your whole file system.
Let me second smag`s suggestion, nautilus on the one hand and classical file search (find, tracker) as well as Zeitgeist should integrate as much as possible, switching over should be just one click away.
Another thing I am curious, are there plans to port Zeitgeist to C/C++ or Vala as soon as the concept is polished? Python is nice for rapid development, but as Zeitgeist (and Gnome Shell, but that is another topic except it should also provide quick access to Nautilus, Zeitgeist, Tracker…) will become an important part of Gnome3 I feel it should not depend on a ressource intensive environment like a python interpretor.
Keep up th good work
Torben
I would love to follow up Zeitgeist news by twitter.
@Torben
I was thinking of taking part of the engine “the datasink” which is the core of Zeitgeist and writing it in C and creating a wrapper for python around it! The rest of the engine can then stay in python!
For Nautilus one can easy write plugin in C that uses the Zeitgeist engine. It is possible to ask for a journal of a specific folder and its child folders!
I will suggest the idea to other
@Ismael Olea
You can follow two of the devs here
http://twitter.com/seiflotfy and http://twitter.com/aantn
This sounds like it would be duplicating the work done by Nepomuk and Akonadi which already has a wealth of support and funding. I would prefer they were moved to the freedesktop project so that they can be used by any desktop, but they are based on C++.
So it’s like what dashboard promised?
@James
Not only! We will include a feature like it! But it is not the main purpose of it!
Please don’t have yet another NIH storage and query engine and use for example Tracker (or indeed, like what Chris said, Akonadi) to store and query your RDF stuff. And please consider reading about SPARQL and Nepomuk.
All of us, several teams working on metadata in general, are really heading and cooperating in that direction. The storage engine really sounds like a duplication of efforts. It would be great if instead projects would integrate with each other, to form a cohesive platform for metadata storage and querying.
If somehow Nepomuk or SPARQL isn’t suitable for Zeitgeist’s use-cases, then as a developer at the Tracker project I will personally make sure that we’ll investigate your use-cases.
Also please follow the activity of the Tracker project for the last months. We have effectively bootstrapped this project to a completely different level. This we started doing even before GUADEC last year (so before Federico’s speech already).
It’s true that a lot of that work is aimed at mobile devices. This of course doesn’t mean that we made it unsuitable for desktops.
Also note that several key people of the GNOME community are involved, they/we are taking the future of the Tracker project unprecedentedly serious.
Zeitgeist is a terrific project and it adds a lot of value to gnome desktop. However, considering where tracker is going these days, I would recomend you guys to use tracker as the backend (and push requirements to us), and focus on doing the automated tagging and other automatic property/data management in addition to being a great UI front-end.