Since Alex Gabriel the maintainer of Mayanna is on vacation we got stuck with some issues! So while he is gone we started working on the old GNOME Zeitgeist again. The development is great. Me more or less absent because of work, Mr RainCT joined the developers during the week and almost finished the DBUS bindings. The code is cleaner. We have a daemon running and we are looking forward to write several UI’s that communicate with the daemon! So whoever feels like writing in a UI in MONO, C, Vala or Javascript DBUS is there for you.
Here is a shor t snippet of our TODO List for the next couple of days:
TODO:
- Backend:
+ Add on Tracker support.
+ Extend DB to support location of usage
+ Auto Tagging e.g: tasks, locations, and relations.
+ Finish the D-Bus interface. (We need to insert items via DBUS , maybe patch up the datasink to have an insert in string tuples)
+ Reduce the amount of “changed” signals send.
+ Fix the Firefox DB detection (see comment in code).
+ Is RecentlyUsedDocuments the best place for text/plain files?- Frontend:
+ Separate it from the backend and let it only use D-Bus for communication.
+ Do not show an item twice one beneath the other.
+ Use a generic CellRenderer to display stars for bookmarking instead of checkboxes!
+ create a bar displaying the related files to currently opened files (inspired by Gimmie)
Is it really needed to have yet another daemon running in the background ?
That will just be useful to kill battery faster.
ok suggest how you want to log the activity
RecentlyUsed from GTK does not really only saves the last usage of a data not the history of it
maybe the guys from gtk3 can look at it
Seif,
If you are using tracker, you should not need another daemon
If tracker does not do something important that you need then please let us know
It is intended for tracker to have all metadata so it can be queried via sparql. This data will include transient stuff like who is online as well as semi-persistent stuff like audit data (who edited file when). Nokia defintely wants tracker to do stuff like show us the most played songs or view when a song was played previously (history)
thanks
jamie
Jamie,
I agree with you totally. But since Tracker is not yet a included in GNOME it is hard to depend on it! I heard 0.7 will be proposed though!
Here is the plan.
The Zeitgeist version for Tracker will run with Tracker as it’s only dataprovider. We will need to know every time a file has been used or edited. in other terms we need a history of timestamps and type of usage (edited, read, opened, saved) for every file.
Zeitgeist will then have to be able to manipulate the RDF representations of each item from Tracker to set in relationships between files etc..
I use Tracker myself and love it and lately I developed more interest in using it as my only Dataprovider. Natan Yellin the co-maintainer will take care of “using Tracker as our main backend”!
Keep up the good job,
Cheers Seif
Seif: Tracker is not yet included in GNOME partly because Zeitgeist is or has been planning to implement their own storage. If Zeitgeist starts using Tracker, we’ll easily get Tracker into GNOME. Our team is also fully prepared to commit to the requirements of inclusion in GNOME.
Notes:
- With Tracker you can already manipulate the RDF of each item using SPARQL UPDATE support.
- I implemented transient storage today (to store the online status of a contact, for example)
- We have some plans to make it easy to turn-off the indexer (note that the indexer and the query & RDF storage service are two different processes). Turning Tracker into just a SPARQL implementation and a Nepomuk based RDF storage capability. This is still useful because using SPARQL’s UPDATE support we can allow you (Zeitgeist, among others) to store Nepomuk based metadata.
- We plan to support custom ontologies, allowing you to add and remove metadata schemas like Nepomuk. Although we strongly recommend, for interopability, to stick to Nepomuk as much as possible.
And finally
- Indeed, like what Jamie says, if there’s something Zeitgeist needs: we are willing to cooperate closely with you and your team to ensure that we’ll solve the query and storage aspect of your use-cases.
Social bookmarking is one of the fundamental internet marketing techniques and a fantastic promotional instrument to gear up your online business. But I think that social bookmarking is a process you just cannot skip if you desire your business to actually start moving, also using a nice automatic bookmarking software is unquestionably the most important and cost effective way to do it.