Slashdot Mirror
Eazel's Nautilus Preview 1 Released
menthos writes: "Eazel released the first preview release of Nautilus, the new GNOME file manager in development, at the LinuxWorld Expo. You can Download Preview Release 1."
← Back to Stories (view on slashdot.org)
PLEASE do not scream about the bloat (20MB download).
The only reason for this, is that most of the
libraries are alpha, and not present on ANY
production systems, hell.. some of the libraries
included could very well f* up your productionsystem good, if installed regularily. This means that the Nautilus-package (for it to be easy to install) has to include almost everything.
Of course this is going to mean a 20MB download.
Think about the size most apps would be if totally
statically linked. Nautilus-preview is not far from that.
In addition. It now downloads and installs
the FULL mozilla-app, while it when finished will
probably only use the mozembed-package.
All this means that the whole thing will move from
a 27MB download (with mozilla) to an about 9-10MB
download when it is all finished.
This is NOT too shabby for an application that is
graphical shell, web-browser etc.
Furthermore. A lot of people are bitching about it not being revolutionary.
What are you guys expecting?
This application will be able to embed about every
functionality on the desktop through Bonobo when
finished. Right now, there are not that many bonobo-components. But to get a glimpse of what
Nautilus will mean when finished, take a look
at the "music"-view when looking at a mp3-directory. This way of viewing directories is
just the beginning, and is TOTALLY different from
the way any graphical shell has operated on Linux,
and even on Windows earlier (I cannot speak for Macs).
Nautilus is not finished by far. But if you look
closely you may get a glimpse of what it may all
look like when finished.
The concept of different views for different kinds of people is excellent. And if you try to use
the "Novice"-view, you might understand how easy
Gnome may be to use come Gnome 2.0.
Its interesting that you bring this up. I don't want to go into a lengthy response, but the Nautilus architecture is in place to totally remove the necessity of a conventional filesystem for users who don't want to manipulate it. We won't drive that to the interface for 1.0, but post-1.0 almost all the underlying stuff is in place and will be tied in.
:-) The index files are pretty small (10 megs or so) and of course optional if you don't want this feature.
Medusa, which was developed at Eazel and will be part of GNOME 1.4 is a disk cataloger and search tool similar to slocate - except that it indexes far more than just filename. It takes about 30 minutes to scan a normal to large disk, and of course isn't going to be doing this while you're working
So what does this all mean? This means that arbitrary, complex searches take a couple seconds to run. Medusa is *also* interfaced in through our virtual filesystem. So, the term I like to use for it is remarkably similar to what you quote... Medusa *is* a multi-key semantically queried virtual filessytem. And yes, post-1.0 this will be tied into virtual folders that are actually "searches" that will live update as you change the disk, etc etc. And it will be *fast*. That's only one example of all the things Nautilus architecture is letting us do...
-seth (seth@eazel.com)
Linux is great. I like it. X is pretty fscked and makes it hard to do some things. X is great because its so networkable, etc, but X is definitely showing its age. This sometimes limits what we can do. There are things we'd like to do but couldn't because of Linux too, but all in all Linux (do to its draw on Unix) is a pretty good architecture, and at least its really stable :-)
I strongly doubt that you've actually used Nautilus, because it doesn't work like explorer. In fact, I thought we'd be flamed to death for it working so much like the Macintosh, but a bunch of lam3rz saw the sidebar and rather than actually looking at Nautilus decided that it looked like Windows Explorer. Guess what? We aren't exactly running explorer next to our development stations. And I'm not hearing people bitch about other free software products like Evolution, KWord, AbiWord, and KDevelop (not to dis them, they're great products - if somethings good by all means copy the good stuff!) that rip off Microsoft products interfaces DIRECTLY - to the menu organization level in some cases.
So guess what? Its not windows explorer. People don't think of mobile phones as computers, and hopefully Nautilus will be intuitive and physically consistent too. Explorer does a lot of things right, and has some good underlying architecture, but they just don't get the final result right. Its really missing in the "feels right" department. And Nautilus isn't. Even this really preliminary release, slow as it is, should give you a feeling for how cool Nautilus is going to be.
And its not a simple file browser. File browser is stupid and limiting. I'm too tired to really go into why its not, but basically Bonobo means that Nautilus can do a lot more to interface you with your documents, data, code, etc etc.
Besides all this, when you actually use Nautilus it looks nothing like Windows explorer. Nautilus has a lot of great eyecandy. Our sidebar is actually fucking useful too, not just wasted space (though I often prefer to turn my sidebar off, all that being possible and configurable of course). The sidebar allows you to access meta-view components like a tree browser, a drag board, annotations, man/info browser, etc etc (all componentized, so its easy for anyone to add new ones, we just have some sample ones).
Try Nautilus. Those who have just looked at screenshots...Fuck off. Come by the booth at LWE and I'll talk about the architecture. It rocks.
-seth (seth@eazel.com)
Nautilus could probably be 3x faster or more when its optimized (before release). We have major algorthimic slow-downs like n^2 algorithms still in the code, but they'll all go away (and its fairly easy). We're almost to the "Feature complete" milestone, and then we'll be in hardcore performance and bugfixing mode before release. Pavel Cisler, who wrote a lot of the file stuff on BeOS (which is *FAST*, like faster than ls!) is working on Nautilus and intends to give it serious speed boosts.
-seth (seth@eazel.com)
If you've actually tried the preview and you have any kind of feedback about Nautilus, whether bug reports, feature requests, wishlist items, etc, let us know!
You can reach the Nautilus developers on or #nautilus on irc.gnome.org. Or you can post bug reports or feature requests directly into bugzilla.eazel.com.
This is our first preview release so it is our first change to get feedback from the community. We want to know about everything you guys think can be done better.
I'm also goint to look over these comments and see if I can bring problems that users are having to the team's attention.
- Maciej Stachowiak
P.S. I wonder if this will get moderated up - actually being a developer of a product doesn't seem to count as very Informative these days!
I'm following in the footsteps of an earlier poster in saying that I'm disappointed to see Apple and NeXT's best and brightest come up with... a file browser. I'm just as disappointed as I was five years ago when I signed up to be one of the first fifteen-hundred BeBox developers, after I discovered what their idea of "revolutionizing" the operating system was.
To quote Alan Cooper, from About Face: The Essentials of User Interface Design:
Fundamentally, I'm a bit tired of hearing about how everyone's "revolutionizing" everything, when they're really not. Look: revolution and revolutioniz e both imply "sudden, radical, or complete change". The American colonies didn't fight the Revolutionary War to install a local king. The French Revolution wasn't so they could hire a newer, prettier cake-eater.
The file system, fundamentally, is an implementation detail. It's an artifact of how "things have always been done". It's a drag on doing real, substantive improvement to the way computers work for people. There are millions of people out there who have never used a computer, and have yet to learn. They don't need to learn what a filesystem is, or to navigate it. They need to be able to find and use the information and tools that are important to them, period.
If we truly want to revolutionize the user interface, the user experience, etc., then we really need to start with a more fundamental re-thinking of how things work. Some of the ideas Miguel de Icaza outlined in his Let's Make Unix Not Suck talk/paper are a good starting place. The universal presence of an ORB, lots of small tools cooperating via the ORB interactively, are all good kicks in the pants of the Unix mindset. But, fundamentally, that's nothing more than what Redmond is doing with COM*, etc. There's more work to be done. There's ripping out the filesystem as a mechanism for data storage and retrieval, and replacing it with a dynamic semantic network, allowing information storage and retrieval (don't confuse data and information). There's moving away from skins and into real, powerful, direct-manipulation user interfaces. For those of you that remember OS/2 and IBM's System Object Model, there was some very, very powerful technology underneath all of that! Hell, you still can't reliably drag a document over top of the printer and have it "do the right thing" in Windows or Linux like you could in OS/2. And that was CORBA all over the place, too, so there was plenty of room for those services to make their way out over the network.
Don't even get me started on package management and installation management, or system administration. Suffice it to say that very little of our technology is designed to help us achieve our goals. It's a lot of work, but this community has boundless energy, and the opportunity and environment to do things that are truly revolutionary. We revolutionized the development model, now let's revolutionize the technology.
Does anyone else get the idea when looking at Eazel's logo that the poor penguin is about to be squashed under the weight of the precipitously balanced puzzle piece? I hope it's not an allegory for Eazel's products' stability or performance.
"If one is really a superior person, the fact is likely to leak out without too much assistance" -- John Andrew Holmes
Those in the Linux camp are fine with revolutionizing interface, as long as it doesn't interfere with legacy (read: awful) programs. As a result no useful interface work gets done except for half-assed attempts to emulate Mac and Windows. But you can skin everything!!! Don't get me wrong: I use Linux and the CLI 99% of the time and have been doing so for two years now. But whenever I boot MacOS into a window on my machine I get all nostalgic for the days when I could navigate to any file, anywhere, in seconds, using only the keyboard. That's because Apple has worked out reasonable and memorable keyboard shortcuts, like:
Thus if I have three JPG images in a directory named:
- FSMHSusNM131_N2.jpg
- FSMusNM131_G8.jpg
- FSMHSusNM132_N5.jpg
(and I do,) then I can select any one of these using at most two keypresses and two arrowkeys. On Linux, If I were to do anything to the third file, I would have to type FSMH(tab)2(tab) while squinting to see what was different about the filenames.Or use a graphical filemanager. But like most Linux wanna-be-cool software, the GUI software for Linux provides all of the look, none of the keyboard shortcuts.
So it's useless for ANYONE who wants to get work done quickly.
True, CLI has its advantages. But for me the speedups only come when I'm so frustrated with the limitations that I start scripting my own solutions. Which I could just as easily do on a GUI machine.
Usability is not a foreign concept, people... why do so few people get it?
I have a positive modifier on Troll. When I mod someone Troll their karma should go UP!