Older blog entries for sisob (starting at number 88)

I built a new sane-backends rpm from Andy's Source RPM which includes Micheal's Patch. Now my scanner works a charm :) Thanks guys.

Eugenia: Good point, I'll try to get to looking at that other scanning program. The general idea is to write a gnome scanning library, with my UI being a frontend and with other applications being able to use the lib too. I hope we can make it possible to make this library not totally dependent on Sane if that is the right thing to do.

Fast User Switching
Every now and then someone mentions that it would be nice to have fast user switching in Gnome. The fact is that we already have it, it just isn't gooey yet. Right now you can run `gdmflexiserver` and use that to log in as a new user. Then you can use ALT+F* to change from one user to another. If you want security you can turn on screen locking.

So how to make this nice and gooey? We could add an "Actions -> Switch User" menu item that would run gdmflexiserver. Then we would need to make gdm aware of users currently logged in, and make it switch to an already open session before creating a new one.


Eugenia: The idea is that the scanner program would be the epiphany not the galeon of scanning. I'm afraid that the target audience is not photographers and advanced users, at least not at the start anyway.

Andy: I'm also having some trouble with sane:

[sisob@sisob sisob]$ scanimage
scanimage: open of device gt68xx:/dev/usb/scanner0 failed: Error during device I/O
[sisob@sisob sisob]$ dmesg
usb-uhci.c: interrupt, status 2, frame# 1655
scanner.c: ioctl_scanner: control_msg returned -75
usb_control/bulk_msg: timeout
scanner.c: ioctl_scanner: control_msg returned -110
usb_control/bulk_msg: timeout
scanner.c: ioctl_scanner: control_msg returned -110
usb_control/bulk_msg: timeout
scanner.c: ioctl_scanner: control_msg returned -110

The scanner does work on other Linux PCs with the same hardware, and it did work once or twice on this PC, but then it stopped. I've even tried resetting my BIOS, but I still get his. Even with multiple distos the same thing ...

re: Eugenia

Eugenia: Sorry, should have been more clear, the scanning program doesn't exist yet. It just a Ui design. The idea was that it would be a frontend to libsane, so it would have what ever features are avaliable with xsane, but with a much prettier interface. Ross said he was thinking about putting some code behind it.

Programming Languages

I know that "every progamming language sucks, each in it's own special way". But I have to say that since I've been learning about object orientation in java it's struck me how UGLY object orientation is in C. Appart from anything else C/Gtk+ is just really hard to read, which to me is a signifigant barrier to entry to hacking.

I was delighted recently to discover how similar C# is to Java. I mean check out gtk#'s hello world, it's just sexy:

	using Gtk;
	using GtkSharp;
	using System;

	class Hello {

		static void Main()
			Application.Init ();

    			Button btn = new Button ("Hello World");
    			btn.Clicked += new EventHandler (hello);

			Window window = new Window ("helloworld");
			window.DeleteEvent += new DeleteEventHandler (delete_event);
			window.Add (btn);
			window.ShowAll ();

			Application.Run ();

		static void delete_event (object obj, DeleteEventArgs args)
			    Application.Quit ();

		static void hello (object obj, EventArgs args)
			Console.WriteLine("Hello World");
			Application.Quit ();

That was taken from the mono handbook. There are some hideous C/Gtk+ hello worlds online too. I always knew I hated trying to hack C/Gtk+, but now I know that there is something better.

I should be a pretty decent Java coder by the end of this (college) year, maybe I'll pickup some C# coding this summer if I'm not working. F-Spot looks very promising and would be great if combined with Original for web publishing, and a Gnome scanning program.

James: ahh, I couldn't see a changelog entry in bonzai but I had the changes locally so I assumed that they had been there for a while. Obviously the changes just hadn't synced to anon-cvs yet.

Nice work on the wiki page...

re: Havoc

Havoc: Seems that then freedesktop X has been in jhbuild for a while now....

<code>jhbuild build xserver</code>

Gnome does need direction

This is something that the KDE boys do better than us. We really need both per-release goals, as well as goals for the 2.x series.

I know that stuff only get's done when there is someone to code, but having a list will make it easier for people.

I also think that we need a methadology for starting new programs. Something closer to the way it's done in "Designing from both side of the screen". ie. The UI and functionality would be designed first on the lists and then the coding could begin..

Now to Haloween...... *witches cackle*

Gnome, the Universe & Everything

I have to say that I'm surprised(and maybe a bit worried) that it was Seth that brought up this whole lack of motavation/direction thing. Why? Because he is/you are one of the biggest injectors of new ideas into the Gnome project. Storage and SystemServices are two of the most promising new projects that I've heard of in the last few months, maybe even this year. On a purely practical level I'd try and get these two into the world - people's interest and maybe even contribution tends to be highly motavating.

As for where Gnome is going, this is how I see it (and none of your damned metaphors either ;) ):

Gnome 2.x Is about filling holes and making what we have at the desktop level really usable without any major changes. This is a pretty big list, but some of the areas are pretty interesting. This includes Multimedia(gstreamer, totem, rhythmbox etc..), Networking(network browsing, desktop sharing, file sharing), Desktop usability (menus, mime system, nautilus, control center), Digital Imaging (as yet we have very little really but there is some movement, need photo managment and editing, scanning, ocr), Internet (IM, IRC, Web Browsing, E-Mail: lots done, lots still to do), Printing, System Configuration, etc.....

The list goes on, and IMHO they are all areas where we have opportunities to excell and "be the best". But also IMHO we need to have a lot of these holes filled before we can really see a lot of popularity within the hacker comminity...

Gnome 3.x This is when the fun starts. As I see it we will have new technologies like storage to blow the user away, and Gnome will get closer than ever to the system with help from HAL, SystemServices and friends. Mono will be nice and mature making C# development for Gnome a breeze, opening up Gnome development to the seas of Uni students who have been tought Java/C#.

This might also be the age where we see freedesktop really come into it's own, and maybe we will see more code sharing between projects..

... I don't know if we really need a "grand plan" at this stage :/ there are so many un-grand things to be done... (eg. Making gnome-vfs reliable and useful for things like menu editing, ftp, ssh, smb, nfs)

Re: Computing facuilities

I'm told I'm in an "enviable situation" with respect to the computing facuilities in my Uni so I thought I'd elaborate on what we have...

Maths Network
8 FreeBSD and 4 (unsupported) Redhat 8 X servers connected to 3 or 4 labs of thin clients. FreeBSD Machines run Gnome 2.2 and Mozilla 1.2. Redhat machines are running Gnome 2.0. Because you can log into any of the servers with the same home dir, and because they have different versions of Gnome on different Unixes, one's Gnome config tends to get pretty wrecked.

I'm hoping that they will upgrade to Fedora and update to Gnome2.4 onthe BSD machines over Christmas.

This is a single Solaris Machine that is run by one of the student societies. Any student can get an account and access is via ssh. Gnome is Sun's version 2.0. Most software is generally pretty out of date, but this server has un-proxied access to the web so it's pretty useful.

CS Computers
There are 3 computer rooms reserved for the Use of the Computer science department. Half of one of these rooms (about 20 machines max) run Mandrake 9.1. The rest run Windows XP Pro. The XP machines have mozilla, and a windows X server so once can run Gnome etc.. off netsoc or the maths department.

Not bad overall I guess :) God bless Ireland
General College Computers
Windows XP Pro accross the board. No X servers or mozilla on most.

Linux in the College
Linux is pretty popular here. All Maths and Science students are required to use the Maths network, and CS students tend to learn about it from eachother and from netsoc. It's not uncommon to be told the wonders of Linux from an excentric maths lecturer in the middle of a Calculus lecture.

College OSS experiences

Over the last few weeks I've slowly had to actually _use_ computers for things other than OSS related hacking, and I've learned a few things.

One things is that the linux on campus is running Gnome 2.2 or 2.0. The 2.0 machines will be upgraded to 2.4 hopefully around Christmas. If we're lucky the 2.2 machines will get upgraded to 2.6 by the end of next summer. That means that we are looking at an 1 year delay getting machines upgraded after the Gnome release is made. If that trend continues I'll be lucky to be able to run Gnome 2.6 in college before Christmas 2004 :(

Realising this I now have a stronger feeling of urgency about getting Gnome 2.5 into a decent shape. In particular network browsing needs a lot of love. I have ssh(and to a lesser extent samba and nfs) access to 4 different sets of personal files space on different parts of the college and I'd really like to be able to browse these visually in Nautilus. Right now I'm using konqueror in college and gnome-vfs-sftp at home.

Mandrake 9.1
The CS machines at Uni are running this at Uni and I REALLY don't like the Gnome desktop for the same reason that I never used the Gnome that came with redhat 9. And the reason may sound stupid - but it's impossible to go back to standard Gnome icons for the application menu. I like to use the standard Gnome themes for all my desktops for consistency's sake, but I can't do this with Mandrake. Appart from just that, the menus are just a mess and there are too many ways to do the same thing.

But I guess that is a pretty common things with distros nowadays. I think that the days of kitchen sink distros may be numbered ;) Well no, they'll be around for ever, but more and more we will see things like XD2 take precedence: A coherent desktop where everything has a function and a place.

Defaults Wh%&e

Some people don't get why I'm a defaults wh%&e. A lot of Linux users are happy to spend time getting things the way they like it, and a lot of them have unusual ways of working - either way defaults are of no interest to them.

For me it's about pride. When I show people linux/Gnome for the first time they generally know that I'm in some way involved with it's development, so I want it to look good, and be easy to use, so that I can be proud of what I am showing.

Gnome Wallpaper Chooser

Dobey has Proposed his new background chooser but jdub
doesn't like it
and seems to think that it's designed from a theme designer's point of view, and not a user's.

<img alt="Screenshot-Gnome-wallpaper-properties.png" src="http://sisob.tuxfamily.org/archives/Screenshot-Gnome-wallpaper-properties-thumb.png" width="250" height="280" border="0" />

Above is what I'd mostly like to see the theme chooser look like. This is dobey's code but Gimped to have the images in a 3x3 grid.

The basic principles behind this, that I feel make it better than the current background chooser are:

1. The user should have to use the file selector to find wallpapers. The wallpapers. that came with their distro/desktop as well as wallpapers they have added should be easily found and used.

2. Each wallpaper tends to make sense only as either tiled, centered, streched or scaled. The user should only have to choose which one each image is once, and the system should make a decent guess at what it should be. Images shipped with the distro/desktop should be set to the mode that makes sense for them by default.

3. The user/distributer should be able to select a background color scheme to go with translucent wallpapers.

It was also said that background choosing is not important enough to warrant much magic, but I would disagree. Every dialog counts...

Note: I wrote about this extensively a few months back.

79 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!