Google is launching this week a beta version of Google Desktop search for Linux in a sign of encouragement by the search giant for Linux on the desktop. Google Desktop allows people to search the Web while also searching the full text of all the information on their computer, including Gmail and their Web search history. Because the index is stored locally on the computer, users can access Gmail and Web history while offline.
Ah, a good desktop search for Linux. About time.
Finally some competition for beagle and other indexers. And one which doesn’t use insanely amounts of memory and CPU.
doesn’t?! if it is anyhting like the windows or mac one (depending on your system configuration) i wouldn’t be suprised if it uses even more memory.
On my system it uses about 5 MB total. Thats with UI and indexer.
CPU usage at the least should be controlled by the scheduler. An appropriately ‘nice’ indexer should chomp up CPU when its there to be used and let it go again when anything else (especially anything interactive) wants a turn.
The trend towards having more processes hanging around in the background doing tasks like indexing combined with the growing sophistication of the Linux desktop should turn out to be an interesting proving ground for the work that has been going into the kernel’s scheduling algos of late.
Edited 2007-06-28 20:01 UTC
http://shots.osdir.com/slideshows/slideshow.php?release=844&slide=1
So far Google was able to xerox your windows hard disk to their servers. No they have complete access to linux geek’s disk also..
yeah yeah i know you can turn it off that feature to share across computers but how many from Joe userland do it…
Would be interesting to know steps taken by linux geeks to prevent Google from keeping your data and privacy being recorded…
Linux userland is alot dif from windows though…it’s REALLY easy to hide whats going on in the background in a windows environment. I’m sure the hounds are on it as we speek, sniffing holes through it, its much more difficult to “hide” stuff on a linux box. Not impossible, but significantly more difficult.
How many from ‘Joe Userland’ actually use Linux?
Increasingly many. (Depending on whose install-base numbers you look at)
Google Desktop vs Beagle(etc.) = Good Thing
Installed and tested this afternoon.
First impressions were the indexing was very slow, perhaps just cautious. It indexes your home and a lot of system-wide documentation, which is a good thing.
It’s great they’ve gone for a native UI … it’s just a shame it takes up so much screen space on a 1024×768 laptop screen! Hopefully that will be configurable in the future. It would also be nice if, as it’s a shaped window, it took advantage of a composited environment. Sadly not.
Searches are quick, however fonts etc cause a long list of suggestions to fill the screen. A picky person would also criticise the text layout and choice of colours…
I noticed this too; I intentionally kept my system idle and kept an eye on the performance monitor while it was indexing. There was some measurable cpu % useage, but not enough to even bump the processor to scale from the lowest frequency. This is pretty much the entire opposite of the abuse I’ve seen Beagle inflict on various systems.
However, me thinks that maybe they are being a little overly cautious. I’d prefer it to turbo if it can intelligently determine my system is idle, as long as it immediately let go when I started doing something. Short of that, I’d prefer errors on the side of caution.
This is very encouraging. Thanks Google.
It would be great to see Google take up the Nepomuk specification and provide an additional implementation for a truly semantic and searchable desktop.
Strigi is already in KDE4 and by this time next year, maybe sooner, should have a very large number of search plug-ins.
The Linux desktop just gets better and more interesting each single day.
Great, now where is the Linux PPC or SPARC version? Oh, I’m sorry there isn’t one, only x86 is supported.
That’s why I’m a supporter of Free Software and not just “Linux”. If this was Free Software it would run not only on Linux, but also on OpenSolaris, BSDs, on any hardware platform.
Thanks, but no thanks.
Google is worth $154 billion dollars. Every single one of those dollars was earned on the back of Linux servers. What has Google given back to the Linux community: Very close to zero. Every offering Google has for Linux is either a warmed over Windows application or a sorry after thought.
Google founders Larry Page and Sergey Brin pride themselves in being politically “progressive” and busy themselves with these multiple left wing causes but when all is said and done their philosophy is just like that of every other run of the mill dyed-in-the-wool scrooge: Take all you can and give nothing back.
I will take these Google offerings for Linux seriously the day they set up a $3 billion foundation to promote the use of Linux and the open source programmers that are about to make them trillionaires.
first of all, they aren’t required to give anything back at all. second, what is the summer of code? i think they do more than you are giving them credit for.
The current Linux indexers Beagle, tracker and Google Desktop, relies on batch indexing by a running process. While this is fine for initial indexing, it is slow and the index is outdated as soon as someone creates a new file or changes an old one.
In spotlight for MacOS on the other hand the index gets automatically updated as soon as a file gets changed. If you open a Spotlight search for “Foo” and go to a terminal and type ‘echo “foo” > test.txt’, the search window will update to include test.txt immediately.
The same should happen with Beagle and Tracker. As an added bonus, you wouldn’t have to have the indexer search through your file system so often to find updates.
This is not true. At least Beagle also index changes as soon as they appear. It is informed by the inotify facility of the linux kernel about filesystem changes.
I think that’s reasonably true for every search utility now outside of locate and find-utils. None of them are doing batch updates as part of a cron job anymore.