Contributor henk0775

Currently we have a number of repositories maintained for ubuntu or other linux distros. This is good in the way that users don't usually have to google for the software they need. However some of the software doesn't make it into the repositories, and there is no way for an application developer to reach the users other than put their software on the web and hope that users will find it through google. Another drawback of this approach is that there are requests to put software into the repositories. And yet another drawback is that it takes a little more time for a newer version of the software to make it into the repository. All of these problems can (at least partially) be solved by using a distributed repository approach. In this case there can be a central directory which will only need to contain some metainformation: where the packages are, how to download them, their checksum, and maybe if they are secure (read: maintained by canonical)

there are other advantages to using this approach, however there also are some security issues that need to be solved before implementing this idea. One is that since anyone can publish to the central directory, an attacker can possibly publish a modified version of software which was developed by another author and already gained some popularity (e.g. Pidgin), that contains a back door.

there are some ideas like http://brainstorm.ubuntu.com/idea/4901/ that are, as you can see here, already implemented - to make http links work automatically in Brainstorm.

But it would be also useful to have the keywords "Idea #XXXXX" interpreted as a link:
Then I could write just "Idea #10660" or "idea 10660"
and it would automatically appear as "Idea #10660"
without the <a href="...">foo</a> tags.

Same could be done with "bug #XXXX"
I like cross - linking.

http://launchpad.net and http://bugs.kde.org does this fine.

Why not use Splashy for a better looking graphical boot? It supports animations, transparency, it's very easy to make new themes (it's a xml with images)...

Splashy complete features from http://splashy.alioth.debian.org
- Require zero kernel patches/full functionality in user-space
- Boot/halt/reboot/runlevel-switch support
- Progressbar support (with optional border)
- Verbose mode (with F2/ESC keys)
- Configuration file in XML
- Cope with any video-mode resolution/size
- Cope with 8, 16, and 24 bit framebuffers
- Alpha channel (transparency) support
- Video mode detection
- Initramfs support
- TrueType2 fonts support
- Lots of image/animation file formats supported: jpg, png, gif, mpg, swf
- Low dependencies and code in C to best perform
- Full LSB support
- Multiple themes support
- Really easy to create new themes
- X detection on exit
- Smooth progressbar movement
- Animations support
- Fade in/out effects
- Totally configurable

Premise: I know that there is two projects, unionfs and aufs, and neither is complete, after all this is a brainstorm...

I think that it will increase the read performance, mounting a ramdisk (a little rd, because etc is small), because there is no heads wandering. And the data will be written anyway in the hd. And a unionfs also allows a snapshot of this important directory.

Consider an early in the boot unionfs to shorter the boot time (but this must be verificated, I admit)

ATM - at least in KDE, I don't know how it is in GNOME, if there is a popup dialog with the possibilities Yes,No,Cancel, you just have the Shortcut Keys "Alt+[key]".
[key] in that case is the underlined letter.

I think it would be good to have additionally the possibility to answer with 'y' and 'n'

or - at least in dialogs where there is no other keyboard interaction possible - let me press JUST the underlined letter, without the Alt key, to speed up my workflow.

This of course should be systemwide, additionally to the normal hot keys, and standardized.
e.g.
y -> yes
n -> no
s -> save
ESC -> Cancel (already works in most of the dialogs)

When I create a new file (for example, on my Desktop),
it is automatically world readable, that is its permissions are (output of ls -l)
-rw-r--r-- for a file or
drwxr-xr-x for a directory.
Every user has his own group by default, so the disturbing thing is the file is

world readable.

I suggest setting a safer umask (not the default 0022), to prevent others from reading private files.

I suggest to use mouse doubleclick at empty space. It must be user-selectable as follow:
do nothing;
create empty text file;
create new folder;
make a sticky note (for desktop)
or so.

(added) Why dislike this? Please drop a line. By default it's never doubleclicks. Only option.

in addition to http://brainstorm.ubuntu.com/idea/10299/

what about a /home on a stick option. if you plug in the USB stick then it auto boots your desktop enviroment on whatever buntu... even a Live CD..

..lightweight Mobilebuntu...

/nomad

what it is: a dir like /nomad/*username*/home and it holds a copy of all the users documents and configs

the way it works
Load a live CD or a Ubuntu until before the login, then plug in the /nomad stick, Ubuntu sees /nomad and it knows now i must log in with this users config (for me this would be like, no wallpaper, no eyecandy fx in gnome, a tuned firefox, eventually a pidgin config, etc.)

In order to implement my idea ( http://brainstorm.ubuntu.com/idea/8377/ ) about a document-centric system, I thought about a diagram which represents a tipical situation. When you download a ogg, you hope that that file should put in a folder where you have your collection. When you download a video, you wish the same. If you use a program that can fetch video from youtube, you want this behaviour.

Tipically, I drop my downloads in the desktop folder, and after a while, I distribute those files in a proper folder.
Why not to create a fuse filesystem that distribute downloads for you. You can create many classes of folder, like music, videos (there is yet in a fresh new user, to simplify our life) according to the file types, or according a project.

I've got a 1 Gb USB storage with an 800 Mb file on it.

When I want to copy an updated version of this file (same name), Nautilus warn me that the operation can't be done cause there is not enought space on the disk.

So I have to manualy erase the old file, empty the trashcan, and finaly copy the new file.


Therefore I suggest that Nautilus check files name when coping or moving operations BEFORE complaining about having not enought space left. (and then simply ask if we want to replace/rename/cancel)

Problem -- Many generic requests such as "Make Ubuntu faster!" are being submitted as ideas. Making Ubuntu faster is definitely something that everyone wants so everyone votes up but we still don't have any ideas on how to make Ubuntu faster.

Solution -- Split ideas into "Problems" and "Ideas". Problems could be submitted separately from ideas but ideas need to be linked to a problem.

In the "Make Ubuntu faster" example, the submitter would enter a problem such as "Ubuntu boots slowly". The idea could be "Don't start any non-interactive processes such as Apache until the user interactive ones have completely loaded".

Voting would be done on the ideas and problems. This way we could see which problems are most important by vote count.

Idea voting would work a little differently. You could only vote for one idea of each problem but have the option to move you vote between ideas for the particular problem.

Gnome-session-properties now is still the same since gnome 2 was launched. But now we have many panel replacement, like bar-panel, awn, gimmie and forgive me if I forgot your favourite. A lot! Now we have many other useful applications, like gnome-do, tomboy, zim, etc.
But is rather difficult today, for a newbie, add a program at the start-up. I suggest a drag-and-drop system of selection of these small apps.

Include forkbomb protection in ubuntu by default. May use limits.conf for this. (limit for number of processes per user)

You have this protection if you get this when executing a forkbomb:
bash: fork: Resource temporarily unavailable

The xsane frontend is the most feature-complete gui scanning program, but the interface is complicated and difficult to use.

There are a ton of buttons for adjusting contrast and gamma and resolution etc. that most users don't need. And there's the histogram window, which is just plain confusing

Would it be possible for a simplified version of it to be made with the more complex options visible only under an advanced tab? Just do something to make it less scary-looking.

File-roller has 46 critical bugs open, is (I think) orphan and has an inconsistent interface (contexual menu view and open with). It could be replaced by fuse or by a nautilus script.
Remove file-roller is secure, because is only a front-end to tar, unrar, etc...

At least a small set of css2 and 3.
The advantages are convergence, standardization and (with a small set) simplicity.

There should be no reason to include the /usr/games environment path on a server installation. This should not be in the $PATH by default, and serves no purpose other then to weaken security. Please correct me if I'm wrong in this.
I did check debian etch installation $PATH variable, and it does not have /usr/games in it.

Moving this to Server rather then security.

Symphony OS is an Ubuntu-derived distro with a unique desktop environment. It's still in beta stage, but with some help from the Ubuntu project, it could become very usable.

Here are some links on it:
http://www.symphonyos.com/
http://www.symphonyos.com/wiki/index.php/Main_Page
http://en.wikipedia.org/wiki/Symphony_OS

I propose that Symphony OS become an official Ubunutu distro alongside Xubuntu, Kubuntu, and Edubuntu, and all it's packages become installable from Ubuntu repos, so that users of the other desktops can simply do a "apt-get install mezzo-desktop" to easily install it.

With all the artists Ubuntu has, and Grub not going anywhere soon, could grub at least get a splash screen by default? It's much more appealing than the simple menu that exists now.

Michelangelo said that his art was actually "to take out marble chips and free the statue".
There are many old and dead apps in official repository.
For example scrollkeeper project is dead since 2003.

A brutal solution is to delete the package.
A clever solution is to finger each possible substitute, work on deleting deprecated dependencies and reuse of code, if possible.