Comment 11 for bug 425258

Seif: Thanks for your kind words.

Everyone: Okay now that I'm home (wrote the previous message when checking mails from the phone, so that I don't forget to answer :P), here comes what I wanted to say.

As I understand it, the point for "origin" is to know where a particular event came from. So, eg. if you are on a website and there's and you click on a link to download a PDF, the resulting DOWNLOAD_EVENT would have the page where the link was as origin (as opposed to the online location of the PDF, which is an information that should go into Tracker). If I am on page1.com/links and there I click on page2.com/foo, the resulting VISIT_EVENT with subject page2.com/foo would have origin page1.com/links, if I read a mail and there I right click on an e-mail address and add it to my contacts list (in case we ever log that) would give an event where the origin is the e-mail you were reading, etc. Like this I think events make sense and are useful, even though I'm not sure they are worth the mess (which is why I wouldn't disagree with removing them entirely).

About just storing the domain name or the document, what's the point? It's just duplication of information. For filtering we should support full text searches (in URI and Title), using the origin (which for this is also named wrongly, btw, it should be called "root" rather than "origin") is rather useless; why would someone want to look for items in /home? What I want is to look at items in /home/rainct/Documents/Foobar/ containing "lala" in the filename.

Cheers!