Linux Desktop

Just A Little Easier, Please – Networks

The Linux desktop has progressed quite a bit in recent years. Ubuntu, out of the box, mainly just worked. I was able to quickly set up a nice desktop for my girlfriend using it. It was really nice to see the little things all fit together. For example, when we took a pic with her new camera and plugged it into the USB port, a dialog popped up asking us if we want to import the pictures. Now obviously, I expected that, and it’s not like that was invented in Linux, but it’s one of those nice touches that just makes life a little bit easier. Still, there is so much room for improvement.

I’ve been trying to think about what specifically I find annoying in day to day usage of Linux. I think my biggest gripe right now is how much of a pain it is just to move my laptop from network to network. I have five networks I tend to use. The first is my home network through my wireless router. The second is also the home network, but wired, through another router. The third is the network at work. The fourth is the wireless network back at my parents’ house, and the fifth is the wireless network at my grandparents’ house, which is next to my parents’ house.

Now, I don’t visit my parents that often, and I don’t connect to the wired portion of my network often either. However, I do switch twice a day between the wireless network at home and the wired network at work. And every time I do, I have to switch network interfaces, re-enable/disable the proxy servers, and change my Gaim account configurations (port numbers for going through the proxies, accounts I wish to auto-login, proxy settings, etc.). It’s just enough of a chore where I think to myself, “Ugh, must do this again.”

When I plug in a Windows computer into a wired network, a little bubble pops up saying that the Ethernet is connected, and it (usually) tries to configure my network settings. It doesn’t always work right, but hey, it’s an effort. It would be nice if we had such a thing in Linux. Maybe we do and I just don’t know about it, but if that’s the case, then we need an easier way for users to discover it and to configure it. Basically, when I plug in a network cable, I want my wifi connection to go down, my wired connection to go up, and a DHCP server to be scanned for. Now the wired-only portion of my network at home doesn’t use DHCP (yet), so it’d be nice for some kind of auto-discovery magic to happen, but really I should be using DHCP here anyway.

The little network selector in my panel is a nice start, really. It’s been buggy here, but it mostly works. However, it’d be much nicer if I could also configure proxy servers for each interface and network. Not tied to that applet, mind you. It would have to be a layer below it somewhere. When I change networks on the command line, the same magic should happen.

The Gaim auto-reconfiguration could happen via a Gaim plugin, which I’m very tempted to write. I don’t know what the easiest way would be to determine when a network changes on an interface, and when interfaces change. Perhaps some kind of D-BUS layer somewhere could intelligently broadcast this information in an easy-to-use form.

I don’t imagine a lot of this would be difficult to develop, and it’s largely a matter of putting the small pieces together (once written). It would certainly make this one aspect of my daily usage a lot easier. I’m sure I’m not the only one frustrated by this. I don’t know if there is work going on in this area or not, but hopefully someone will get the motivation to hack on a piece or two.

Just A Little Easier, Please – Networks Read More »

Various Updates

I don’t blog as often as I should anymore. So, here’s another general update on stuff.

I released 0.4 of the Desktop Notifications Spec tonight. No, people, it does not allow XML + CSS or XSLT or XEMBED or sound or anything else. It does support enough to be useful, without going overboard. I think it’s getting to be a very good spec, and I’m about ready to call it done, after a few more (sane) discussions.

Galago .NET bindings are coming along nicely. Galago# itself works rather well, and I’m working on GalagoGtk# now. The one problem is that Gtk# supports Gtk 2.2, not 2.4, and I have some 2.4-only widgets in libgalago-gtk. So, I’m going to have to provide 2.2 alternatives. *sigh* Once that’s all done, GalagoGtk# should be ready as well, and the Beagle and Tomboy guys can have some fun with them.

I must say, Tomboy is very cool. It Just Works (TM) and I’m already finding it very handy. Kudos, guys! Now if only I could auto-link #123456 to Bugzilla entries at work. Hmm…

Various Updates Read More »

Desktop notifications spec announced

Mike Hearn and I just announced the first officially public draft of the Desktop Notifications Specification we’ve been working on. It has a few rough edges, but is there for the most part. The specification was announced on the freedesktop.org xdg-list. We’re looking for feedback and some implementations. Currently, we have a few projects who have announced their intention to use the specification, hopefully with more coming in the future.

The protocol for the desktop notifications uses D-BUS. It’s meant to be desktop-neutral, and does not define presentation. An initial GNOME notification server is undergoing development, as well as a reference implementation library. These are not yet available, however, as Mike is without an Internet connection and has not been able to submit his latest changes. These will be updated as soon as possible.

Desktop notifications spec announced Read More »

Package management, autopackage, and “The user interface vision”

I just finished reading Mike Hearn’s article The user interface vision. It’s an interesting read on his views of what the UI for package management can evolve into. I like the ideas, although some things need fleshing out more (as he also states).

If we really want to get Linux on the desktop, this is an area that needs to be tackled. Sure, Gentoo’s method works great for Gentoo. Debian’s method for Debian… The thing is, the average person who buys a computer for school or entertainment or whatever shouldn’t have to care what distribution they’re using. I’ve seen this many times in Gaim. A person will come in asking how to install Gaim, saying he downloaded “the file” from our site, but can’t run it. We then have to explain package management, the commands required, and tailor it to the distribution they’re using. A better solution is needed, and it’s about time we figured something out.

About a year ago, various people who were/are involved in package management (Mike Hearn from autopackage, Jeff Johnson from rpm, Alain Penders from Gentoo, myself from GNUpdate, and many others got together to discuss package management and a way to standardize on meta-data (such naming conventions, dependencies, etc.). We had some good discussions, but in the end, not a whole lot came from it. Although disappointing, it wasn’t unexpected. The differences between various package systems is not only technical, it’s political in some cases.

So far, Mike’s autopackage is the first possible solution that has impressed me. While it has a ways to go, it’s providing an interesting distro-neutral form of packaging, and appears to be well thought-out. I’ll be supporting autopackage in the first release of the Galago software.

I guess only time will tell. A lot of effort will be needed, as well as cooperation, if we’re going to solve this problem.

Package management, autopackage, and “The user interface vision” Read More »

The Future of the Linux Desktop?, Part I

It seems there are a lot of interesting things being developed lately. I’ve been reading about GNOME Storage, which is a really cool concept. I haven’t played with it yet, or looked into the code, but I think that if done right, it can really add to the desktop experience. Perhaps it could be used as a feed for Galago.

Linux as a desktop system has, in my opinion, a lot of potential. Though I wouldn’t recommend it at this point to most people I know, I do hope for a day when I can. The reason why it has so much potential is that anybody can modify the source of any program to provide the integration, assuming there’s not an API and plugin architecture that would do the job. Of course, we all know this, but think about it. Work like Dashboard or Gnome Storage just wouldn’t be all that doable under a closed environment without a lot of collaboration and licensing between Microsoft and any other companies. Unless of course there’s an API available there, but MS APIs only go so far.

There’s a lot that can still be done. My goal with Galago is to be able to automagically retrieve information related to the task at hand on almost any supporting program, in a desktop-neutral fashion, and to have an indicator of a person’s status wherever you see his/her name, e-mail address, or other contact information. But this is just one small part.

I want to see a desktop where when I open my report for school, I can instantly see on the side of my screen or somewhere icons for the other files related to this report, URLs I accessed to find the information, and the status of any people on my buddy list I worked on the report with. I’d prefer not to see this in a big white window on the side, like Dashboard does and Galago will do, but something more integrated into the environment. How, I’m not sure yet.

Something else that may be neat to see in the future is a revision control feature built somewhere in the backend. I believe Gnome Storage has this in their plans, and I hope it ends up being a part of the desktop experience. I’ve had clients and family members complain that they accidentally deleted or overwrote a file they were working on, and wanted me to recover it. If this was built right into the system, their documents would be safe.

Drag-and-drop install/uninstall of applications is another feature I’d love to see. This was something that Mike Hearn brought up the other day. If a person could download a single package and drag it to their desktop to install, and then drag it back to the trash to uninstall, it would simplify package management for the user dramatically. Of course, this requires that we finally get packaging sorted out and standardized, but this isn’t likely to happen any time soon. However, his autopackage project does provide an interesting form of package management that I wouldn’t mind seeing take off. The rest could be abstracted through GNUpdate.

At this point, I’d like to bring up how cool the stuff Robert Love is working on. It’s one of the upcoming things that are really exciting me.

I had more I was going to bring up, but it slipped my mind, so I think I’ll end this post….. now.

The Future of the Linux Desktop?, Part I Read More »

Scroll to Top