FOSDEM 2026: Desktop Search Interview

It could be a really massive step to have technology and instruments like this I believe. What's everyone's opinion on utilizing this know-how as a substitute of a totally new one? KDE can ship an entire surroundings of applications that benefit from this know-how from day 1. these functions will make all of the difference IMO. Do not get me improper, I haven't got an issue with Postgres, I simply hate to see KDE use the identical argument in opposition to something that has been leveled towards us. It's licensed underneath Apache license, Ansi SQL, portable (pure java), has a small footprint and is supposedly very easy to use. I'm stunned too, sqlite is very small and is in public area license. It might be cool if the backend was developed utilizing Qt4, so it would be a pleasant small dependency that other tasks may use. Time. And i mean that in a few methods - we want time to complete fleshing out the concepts, time to let the stuff mature inside of KDE and properly, the couple of us working on it could use more time for such.

I did the original mockup prior to the KDE conference this summer time with SQLite and the performance for the type of queries that we're doing was so unhealthy that it simply isn't an option. So KDE is going to choose a database that's LGPL because of licensing restrictions from utilizing a straight GPL'ed database when we use a GPL'ed toolkit to do Everything associated to Qt? About 5% of the time, it will be much less efficient but they won't know that going in, so they'll use it anyway. What is going to take a superb while will be porting over applications to make use of it where appropriate. I say.. good luck? That a few of the information is not online doesn't change that categorization is a great way to prepare it. I feel this is a good choice. Think horsepower. OS/2 had a useful skill in the Workplace Shell the place you would have a folder containing documents, and when you opened or closed the folder, the documents could be opened or closed, ie. After all, that could be found simply by the fact that I do know the name, but there are some things that are not so simply tracked by identify, have no searchable contents, and are hardly ever used however necessary.

Reiser and RDBMS's are like X Window: low level applied sciences that gives a way to accomplish the task (with various levels of success); things like what this interview is about are like Qt: a layer that makes software improvement leveraging the potentialities of the platforms possible. The middleware, what knowledge is archived and listed, and how contexts and patterns are matched is the key. I need something that acknowledges contexts of activity. Perhaps I'm missing something here, however what I need to begin with is that I've a listing with a bunch of HTML recordsdata in it and i wish to be in a position to look them for content material identical to Google searches the web. I've began playing with reiser4 some time in the past to jot down a plugin to enumerate modified recordsdata fast and reliably. Does your plugin have a webpage? E.g. I would have /residence/bryan, and mark it recursive and hidden As root I might add /usr/share/music (all my oggs and so on., shared between users), and mark it as recursive and visible to all. For each directory have a sqlite DB in the basis of that listing referred to as .dirindex.sqlite or something. SQLite additionally locks the whole database on write, which simply isn't acceptable in a software used often by multiple processes.

But I do wish it could all be unbiased of the particular database used. Then the tasks that I work on, ie taxes, planning a trip, researching a selected subject, work duties equivalent to proposals and product analysis, and so on. For example, in june of last yr I was researching travel in europe since my daughter was travelling there, and that i wanted to figure out how one can her get somewhere. I found fascinating sites, some helpful emails got here in, including correspondance with my daughter. The one model for this that I've found to be useful is that of categorization. I used to be referring to storing solely the categorization data in the filesystem, not the remainder, and that just being a mirroring. Another vast font of ineffective info thrown in your face. Nat Freidman at RealWorld Linux last year speak on something comparable simply to repeat Scott right here: Beagle and this venture really aren't the same form of thing. A standard Unix type filesystem is only one factor that you may make utilizing this kind of system, but way more would be potential. No joins, far much less memory, and far more speed! Not to mention that a rich filesystem plus a complicated code rights system (a la jail) will result in a very safe, highly effective and stable system - far more abuse tolerant and flexible than has ever been generally obtainable ("Open all of the attachments you need - they run effective, something they change rolls back, and they can't send Spam").