Contributor mosestruong - Ubuntu brainstorm

Ubuntu QA:

Blog

Brainstorm

Package status

The Ubuntu community has contributed 22700 ideas, 138270 comments, 2629576 votes

All more
Projects more
Global categories more
- Accessibility
- Documentation
- Education
- Gaming
- Graphics
- Hardware support
- Installation
- Internet & Networking
- Look and Feel
- Marketing
- Multimedia
- Office
- Programming
- Quality
- Security
- Server
- System
- Usability
- Others

Idea sandbox

Popular ideas

Ideas in development

Implemented ideas

Contributor mosestruong

It would be nice if in Nautilus ...

Written by l3on the 6 Sep 09 at 20:50. Related project: Nautilus. Not an idea

It would be nice in Nautilus ...

... when you select some files:

the info in the Properties window:

was also included in the sidebar:

563
votes

124
votes

151
votes

Solution #3: Make the Nautilus-code more general so anything can go into the sidebar

Written by vhindriksen the 7 Sep 09 at 17:29.

See for example the side-bar of browsers like Firefox. So just let any code run in the sidebar (in a separate thread, so it can crash on its own). Output can be html or some specific xml, as long as perl-scripts can spit it out.

One serious problem with this solution is security, since the script is auto-run for each file/directory.

476
votes

224
votes

255
votes

Solution #6: Make this optional and configurable via preferences

Written by LaMeR the 10 Sep 09 at 20:18.

1. Let the user set which information to show in the sidebar.

Example options:
For one selected file show in sidebar:
- General info - similar for most file types (thumbnail, type, size,...)
- File specific info - different for most file types (for example video files info: duration, resolution, bitrate, codec,...)
- All info = General info + File specific info

For multiple selected files show in sidebar:
- General info - number of files, size of files,...
- File specific info - number of files for each selected type (6 images, 13 videos,...), size of files for each selected type,...
- All info = General info + File specific info

2. Make it possible for user to completely turn this feature off, because some users may have slower computers or want to use the sidebar for bookmarks etc.

49
votes

Solution #7: sidebar should automatically change to "info" after selecting a file

Written by haran_elessar the 14 Sep 09 at 02:37.

I think that extra information is neat but it would be better if the left pane just changed automatically when I click on a file instead of having to choose "information" from the menu to see the details. the way I see it, most people like having the places view and they would have to change it to "information" each time to see the info just by left-clicking.

Now if nautilus would automatically change from places to information whenever you select a file then I think that would be very handy.

Optional extra idea: how about if that sidebar would show some other info automatically. For example, say I'm selecting many files from a folder, wouldn't it be neat if the sidebar would show me small thumbnails of all the items I have currently selected? (similar to Picasa's tray) That way it serves as a double check that I'm actually copying all the files I need. Then after pasting the sidebar would just go back to show the "places"option.

As it is right now, not many people are using that drop down menu. It's sad since now that I tried it I see it can be useful and even more if the feature you are requesting is implemented. Still, I don't think I would be using "information" as my default sidebar BUT if it was automatic as I mentioned above I think everyone will benefit from it.

what do you think? can that be done?

[...]

edit: As some users have suggested this feature should be configurable, allowing users to turn it off completely, showing only certain kind of information and generally behaving like users want it to behave.

-26
votes

-104
votes

-59
votes

11
votes

Solution #11: Make sidebar more dynamic

Written by alx321 the 4 Oct 09 at 08:50.

Many ubuntu users don't even know that it is possible to change the content of the sidebar. The dropdown menu is very unhandy.
I propose to replace it by tabs with small icons and make it dynamic. That is, you can drag the tabs to another position, so that, for example, 'information' is shown beneath 'places' or in a second sidebar at the right or lower border.

8
votes

4
votes

Solution #13: Add nautilus option to show file info in a popup bubble

Written by oliver-joos the 5 Oct 09 at 14:23.

I propose to add a Nautilus option to pop up a light yellow bubble with basic file info when the mouse rests on an icon for a second or so.
The option should be very simple: on/off or none/brief/verbose. The file-type decides what info is shown.

Advantages:
* no mouse click needed
* no second/dynamic/changing pane needed
* plays well together with the audio preview feature
* the mouse is on one icon at a time => the info is clearly related

See the 22 comments or propose a solution (latest comment the 26 Jul 12 at 13:50) >>

Different icon for .deb packages

Written by mydoghasworms the 22 Jun 10 at 18:32. Global category: Look and Feel. In development

The icon for a .deb package looks too much like an ordinary archive.

Developer comments

Thanks for the suggestion. Making .deb icons more distinctive is a great idea.

I have reviewed the proposed solutions with Michael Vogt, our packaging expert. Solution #1 is straightforward, but we particularly like solutions #5 and #10, using a thumbnailer to show the application icon from inside each package.

Deb Thumbnailer is a good start, but it needs a bit of code cleanup. We will work with the Deb Thumbnailer developers to see if it can be made suitable for the default Ubuntu installation.

If you would like to help out, the next step is to get it packaged: http://launchpad.net/bugs/685851

— mpt

426
votes

-40
votes

Selected solution (#2): Allow package to optionally include its own icon

Written by mydoghasworms the 24 Jun 10 at 05:15.

It would be nice if every package could include its own icon, which would display in Nautilus, and if it has no icon, then some default must be displayed. This would be especially nice for third-party vendors providing software (e.g. Google, Skype, VirtualBox, etc.)
This is of course an optional specification, and if the package includes it, and the file manager supports it, it would show the icon.

-- UPDATE --
As per comment below, this is now a reality: https://launchpad.net/deb-thumbnailer

78
votes

-19
votes

Selected solution (#4): Solution #1 + mockup icon with a concept

Written by lvxferre the 1 Jul 10 at 19:56.

Mockup icon: http://img812.imageshack.us/img812/218/debinstpack.png

The mockup I made is horrible, I know - please, look at the concept behind it.

The icon is still a package, somewhat similar to the old icons. The Debian logo is still there (it's .deb after all), but it should be bigger and with a more contrasting color.

The package is open. Inside it's a floppy - while few of us still use floppy disks, most associate it with "install". The green arrow in the floppy hints "download".

So, an icon like this tells us: "You downloaded this file, There's a program inside. You can install the program if you want." Exactly what we expect from a .deb .

13
votes

-4
votes

Selected solution (#6): Have a different icon for archives

Written by Stancel_Spencer the 10 Jul 10 at 09:07.

That looks like a box. So it is perfect for software packages. The metaphor is great, it is like the package that comes in the box that you order and you open it up.

Archives, like zip files, have often been shown in other OS as a folder with a zipper on it. You could also use a filing cabinet for the icon. But my point is, if anything needs to change, it's the archive icons.

-13
votes

-16
votes

-31
votes

51
votes

See the 22 comments or propose a solution (latest comment the 12 Apr 12 at 07:54) >>

Less screensavers

Written by programad the 11 Mar 08 at 15:24. Global category: Look and Feel. Implemented

Actually, we have 5 wallpapers, 9 themes and 80+ SCREENSAVERS!!!

WHAT THE HELL!

Who need that amount of screensavers?

I think we should vote five to the default installation, and the other ones comes with a optional package?

Isn't gnome "keep it simple"?

For me just a "dim screen" is enough for default.

681
votes

36
votes

-7
votes

Selected solution (#3): do not remove any screen saver but improve them

Written by slsolaris the 18 Mar 09 at 03:52.

for example GLText should be editable, give as the option to set speed, improved graphics, etc... but do not remove them because a lot of them are awesome.
in the other hand some of them are ugly and u should remove:
Lavalite, sonar, spheremonics, circuit, deco, engine, FlipScreen3D, Floating Toasters (what the hell is that, LOL), GFlux, GLKnots...
please do it canonical, Linux for human Beings, computer needs to be personal.

-208
votes

178
votes

406
votes

712
votes

Selected solution (#7): place in a "screensavers-extra" package

Written by lee.jarratt the 28 Apr 09 at 20:34.

A good idea is to remove many of these screensavers (some are quite poor) and place them in a "screensavers-extra" package so users who do want to use a not-so-popular screensaver can just install the package.

An advantage for this is that it will cut down the size of the disk image by around 8Mb. That 8Mb for Screensavers could be used towards something more useful.

81
votes

Selected solution (#8): Offer Download & Install easy interface

Written by pumukyjunior the 6 May 09 at 14:34.

I think it could be interesting to mantain the most popular screenSavers by default, and let the user to search, explore, and to try another screensavers.

Maybe, these ones could be stored in a repository, and shown to users by a user-friendly interface.

I think something simililar is done in KDE 4.0 when you want to install a new Grub image.

Finally, I would keep the black screensaver by default, in order to try to safe Energy.

-208
votes

Selected solution (#9): Offer "Screensavers" option in Ubuntu installation.

Written by rmconard the 7 May 09 at 15:16.

During the install process the user should come to a section about the "look and feel" of Ubuntu. From this spot in the installation the user should be given the option of installing which screensavers he/she wants. You should have 3 options:

1) Install All Screensavers
2) Install No Screensavers
3) Install Some Screensavers (choose which ones)

This way, the ones the user doesn't choose don't even get put on the HD.

See the 42 comments or propose a solution (latest comment the 18 Jan 12 at 17:19) >>

GIMP is out from Lucid Lynx.

Written by hellmet the 20 Nov 09 at 18:49. Related project: GIMP Image Editor. Won't implement

GIMP is being taken off the default install due to "CD space" problems. However, it should remain easily accessible/installable.

-450
votes

-45
votes

146
votes

409
votes

550
votes

Solution #5: give a simple paint app

Written by yzarc the 21 Nov 09 at 14:37.

Very often I just wanna draw an arrow over a map and write "the party is here!". Or just crop a picture.

And it's a pain to do it in gimp. It's like operate war tank to kill an ant. Please give us a simple image editor, no layers, no vectors, no Filters, etc. Try to remember the first time you opened the gimp. What do I do with all these windows and tools?

341
votes

Solution #6: Add an applications section to the installer

Written by tehalynn the 21 Nov 09 at 23:31.

When someone installs Ubuntu, they are shown an applications page with the following options:

1. Install all default applications. Some will need to be downloaded.
2. Install only default applications that do not need to be downloaded.
3. Select which applications to install.

Option 1 will install applications that are not on the CD, such as GIMP, by downloading and installing them when an internet connection is made.

Option 2 will install only applications on the CD. It should tell the user how to get them later if they change their mind.

Option 3 shows a page listing applications that will be installed. The user can exclude applications they do not want, and include ones they do. Here's a mockup of the general idea:

Only major applications need to be mentioned here. This is not a replacement for Ubuntu Software Center/Synaptic.

-165
votes

-117
votes

Solution #8: first run

Written by nloewen the 23 Nov 09 at 05:02.

At the first boot of the system after install, a dialog box could pop up asking if the user would like some help setting up there system. It could go through users and what not as well as ask what the user would like to use there computer for. if they select something like photo editing or replacement for photo-shop, the program would select GIMP for install with a button to let them select a different program from the repository.

-49
votes

-123
votes

Solution #10: Offer Multiple ISO downloads

Written by fix98-win the 24 Nov 09 at 09:58.

CD-Rs come in packs more often than not.

Offer for download in ISO format a two disc system with just Ubuntu on the first disc and extra applications on the second disc. This would possibly be in addition to the standard one disc system.

This way, users can save both discs for when they need to reinstall software that came with the system, but are offline.

Besides, using two discs frees up space for more software to be included with ubuntu by default.

ps: Don't forget to give users a choice of which applications to install from the second disc.

176
votes

242
votes

-285
votes

-197
votes

Solution #14: Remove every internet related application.

Written by Vladimir Hidalgo the 26 Nov 09 at 01:46.

This is a mix of previous solutions, but what I propose is to remove basically everything on "Internet" menu but Firefox.

Why?, because people that need those internet related apps do *have* internet to use them, therefore they can download & install them in an easy one-click install process like described in #1.

That is, to remove by default:

-Transmission
-Pidgin/Empathy/Ekiga
-Terminal Server Client
-Evolution
-Ubuntu One

I mean, common, those are needed only if you have Internet, so they can be pulled & installed on first use.

429
votes

Solution #15: Software Center should have a "recommended Software" section

Written by Umang the 26 Nov 09 at 12:59.

The software center should have a "Recommended Software" section where programs such as Thunderbird and GIMP are offered for downloading.

New users are flooded with too many packages to actually get a taste of the best FOSS can provide them. This section should be some sort of an exclusive section and only stable and very popular software should be included.

153
votes

-72
votes

Solution #18: Include KolourPaint instead of GIMP

Written by 8bit the 27 Nov 09 at 07:34.

The thing missing from Ubuntu is a simple, MS Paint-like program. These are nice in that they are easy to learn, run smoothly on low end system, lack clutter, have small install footprints, and open up seemingly instantly. KolourPaint is all of these:

Pros:

-The design is good. It doesn't feel cramped, it feels uniform and inviting, and most of the button icons are self-explainable. It basically looks like a prettier, shinier pre-Windows 7 MS Paint.
-It functions almost identically to MS Paint. A casual Windows user likely wouldn't be able to notice the difference, and the time it would take to transition would be minimal
-Start up time is fast
-All features that are in KolourPaint and not in MS Paint are pretty useful, and in no way get in the way of the MS Paint-like features.
-The undo history is huge. Where as MS Paint only allows 2 undos, I've gone back hundreds of actions and still haven't hit a wall
-Anything that's been changed from MS Paint has been changed for the better. Zoom is more powerful, the eraser has more brush sizes, the shape tools let you choose line thickness from within their tool box, and don't require you to switch to the line tool to choose line thickness.
-It's pretty stable. I had a weird hangup when using the zoom tool once, but I haven't been able to reproduce it since, and gPaint, on the other hand, crashes whenever I click on the 'fill' tool.

[...]

-As I mentioned, there are a couple of original tools- The connected lines tool, the color eraser tool, the freeform selection tool, but the most interesting addition is the Color Similarity tool which allows a user to set selection threshold for the fill tool. (and other tools, presumably) What this means is that you can have two, very similar colors, set this setting to something high, and then use the fill tool on one, and it will fill two. This is very useful, and is somewhat analogous to the magic wand tool/color select.
-Supports saving in a multitude of files.
-Supports editing in a multitude of files. (including XCF)
-Sort of has support for transparency (THIS IS AWESOME!)
-Only just over 2MB in size, as compared to GIMP's 14MB
-Everything is in one window. I quite like this.
-Free and open source. Uses the BSD license.
-It's being actively developed.

Cons:

-No layering support
-Very limited transparency support. While transparency itself is supported, translucency isn't, so you can't have an image with parts that are sort of transparent. It's either completely transparency, or completely opaque.
-No transparency for the tools. This seems obvious considering the above. You can't set transparency levels for, say, the paintbrush, for example.
-There's an uneven number of tools, meaning that there is a space where an icon could go which is empty. (Not a big deal at all.)
-The 'fill' tool isn't a paint bucket, so it may be confusing for a half a second, but the icon makes sense, and is more elegant than a paint bucket.
-The color picker icon looks like a safety pin to me, which was confusing at first.
-When you pull on the edge of the canvas to enlarge it, it doesn't resize in real time. Instead, it shows you a selection box sort of thing. (MS Paint does this in real time.) This is incredibly knit-picky. 99% of people wouldn't even notice this, I don't really care, some might even prefer it.
-Does not support saving in XCF format.
-It ate up my processor once, forcing me to restart my whole system, once. I've tried pretty hard, though, and I haven't been able to show the issue again. It's entirely possible it was another piece of software I was running. Other than that, it runs completely smooth.
-On a live CD 2MB can still be substantial, but it's definitely an improvement over 14.
-The color swatches bar takes up a little bit too much vertical space in my opinion. Why can't the color profile be on the right side, rather than above? There's all that extra horizontal space, afterall.

In conclusion, the cons are either extremely knit-picky, or exist in similar programs, such as gPaint and MS Paint, anyway. KolourPaint seems perfect for the live CD as a lightweight, Paint-like program. I highly recommend it.

GIMP, on the other hand, is just too large for Live CDs, too unweildly for casual users, and really easy to get for anyone who actually needs it.

Pros:

-The design is good. It doesn't feel cramped, it feels uniform and inviting, and most of the button icons are self-explainable. It basically looks like a prettier, shinier pre-Windows 7 MS Paint.
-It functions almost identically to MS Paint. A casual Windows user likely wouldn't be able to notice the difference, and the time it would take to transition would be minimal
-Start up time is fast
-All features that are in KolourPaint and not in MS Paint are pretty useful, and in no way get in the way of the MS Paint-like features.
-The undo history is huge. Where as MS Paint only allows 2 undos, I've gone back hundreds of actions and still haven't hit a wall
-Anything that's been changed from MS Paint has been changed for the better. Zoom is more powerful, the eraser has more brush sizes, the shape tools let you choose line thickness from within their tool box, and don't require you to switch to the line tool to choose line thickness.
-It's pretty stable. I had a weird hangup when using the zoom tool once, but I haven't been able to reproduce it since, and gPaint, on the other hand, crashes whenever I click on the 'fill' tool.
-As I mentioned, there are a couple of original tools- The connected lines tool, the color eraser tool, the freeform selection tool, but the most interesting addition is the Color Similarity tool which allows a user to set selection threshold for the fill tool. (and other tools, presumably) What this means is that you can have two, very similar colors, set this setting to something high, and then use the fill tool on one, and it will fill two. This is very useful, and is somewhat analogous to the magic wand tool/color select.
-Supports saving in a multitude of files.
-Supports editing in a multitude of files. (including XCF)
-Sort of has support for transparency (THIS IS AWESOME!)
-Only just over 2MB in size, as compared to GIMP's 14MB
-Everything is in one window. I quite like this.
-Free and open source. Uses the BSD license.
-It's being actively developed.

Cons:

-No layering support
-Very limited transparency support. While transparency itself is supported, translucency isn't, so you can't have an image with parts that are sort of transparent. It's either completely transparency, or completely opaque.
-No transparency for the tools. This seems obvious considering the above. You can't set transparency levels for, say, the paintbrush, for example.
-There's an uneven number of tools, meaning that there is a space where an icon could go which is empty. (Not a big deal at all.)
-The 'fill' tool isn't a paint bucket, so it may be confusing for a half a second, but the icon makes sense, and is more elegant than a paint bucket.
-The color picker icon looks like a safety pin to me, which was confusing at first.
-When you pull on the edge of the canvas to enlarge it, it doesn't resize in real time. Instead, it shows you a selection box sort of thing. (MS Paint does this in real time.) This is incredibly knit-picky. 99% of people wouldn't even notice this, I don't really care, some might even prefer it.
-Does not support saving in XCF format.
-It ate up my processor once, forcing me to restart my whole system, once. I've tried pretty hard, though, and I haven't been able to show the issue again. It's entirely possible it was another piece of software I was running. Other than that, it runs completely smooth.
-On a live CD 2MB can still be substantial, but it's definitely an improvement over 14.
-The color swatches bar takes up a little bit too much vertical space in my opinion. Why can't the color profile be on the right side, rather than above? There's all that extra horizontal space, afterall.

In conclusion, the cons are either extremely knit-picky, or exist in similar programs, such as gPaint and MS Paint, anyway. KolourPaint seems perfect for the live CD as a lightweight, Paint-like program. I highly recommend it.

GIMP, on the other hand, is just too large for Live CDs, too unweildly for casual users, and really easy to get for anyone who actually needs it.

127
votes

Solution #19: Ubuntu App Collections

Written by 8bit the 27 Nov 09 at 07:41.

Mozilla has a section of their Firefox add-ons site dedicated to Add-on Collections. These are collections of add-ons handpicked by developers for different uses, (Web developer, family, etc..) which can be downloaded and installed with a single click. Implement a similar section for full applications in the AppCenter. Have one for artists, one for web developers, one for gamers, etc...

-36
votes

Solution #20: Solution #10 Revised.

Written by fix98-win the 29 Nov 09 at 09:39.

I think it might have been the double ISO thing that ruined it. Lemme try again.

Offer Ubuntu for download one one disk. This will save download time, and installation time.

Forget the second disk. I was trying to imply the second disk would be optional, or even nonexistent.

Why would anybody want to backup their software anyway? (sarcasm)

24
votes

16
votes

Solution #22: Like Solution #19, but have it where any user can suggest applications

Written by aubre the 30 Nov 09 at 18:12.

Allow users to submit lists of applications that they like to have on their machines, have them assigned a "collection name" and then give users the option of either browsing the lists or searching for the list name (perhaps they already know the list they want to install).

An example is perhaps an academic lab manager has certain applications they install on each machine. They make a list, submit it to Ubuntu, and then as they install new machines they can simply input the name and all the packages they need will be pre-selected for them. Even better, an academic lab manager on the other side of the world can browse the list, find this pre-built list, and either use it unchanged or adapt it as a starting point of their own list.

There are text/CLI ways to do these things but it would be nice if it could be incorporated into the graphical install interface.

26
votes

-26
votes

-151
votes

53
votes

Solution #26: Gimp Light (a easy and not so messy Version of Gimp)

Written by account the 5 Dec 09 at 13:54.

The newest Gimp is a single-window-Application. The source is open. So I propose to take that source, delete all what is not really necessary and create by this way a simpler Version of Gimp. I call it Gimp-Light and I think it's a great idea. So peoples can accustoms the feel of Gimp, but without the messy feel of the normal Version of Gimp. If they needs more, they can install the regular Gimp. The Gimp-Light shouldn't have many functions - only basic ones for painting etc. This will reduce the weight of this application and can be shipped also in the CD-Version of Ubuntu without problems. A paint-application can stay also in the whole amount of the packages! What you think about my idea!?

-47
votes

-11
votes

Solution #28: Remove extra stuffs or compress less used with paq

Written by DrG the 7 Dec 09 at 08:36.

Remove some of the

Games totalling > 3.0 MB
Backgrounds 9.5 MB ( replace high detailed ones with less detailed ( ie. more compressible to the sqashfs ) ,but eye candy ones.
Themes 1.4 MB - ( remove some ugly ones )
Sounds 3.1 MB - these can be replaced by symbolic format sound ( which takes very little space )
Screen Savers

And such similar items
( choose only a few of these items by voting )
------
or
compress the less used one with paq8px
Example - in karmic , when Shoes.jpg was compressed with paq806 -6 the size was reduced to 802.3 kb from 1016 kb.
( lzma , which is used in the Karmic CD squashfs is not able to reduce the size )

For more informations on compression
http://www.maximumcompression.com/data/summary_sf.php

[...]

----

To make space for GIMP or GIMPShop

19
votes

38
votes

-6
votes

-27
votes

Solution #32: Develop www.phofiltre-lx.org an include in in Lucid Lynx

Written by walo the 12 Dec 09 at 16:56.

¿Haven't you used photofiltre fow windows?

This is one of the best free and useful image tools i have ever used.
Photo filtre LX (http://www.photofiltre-lx.org/) was created to have this simple but really useful aplication under GNU/Linux. I think this develop needs to be released as soon as posible in order to replace GIMP or to have other SIMPLE image manipulation program.

It's going to ve developed using C++ and imagemagic library

42
votes

-10
votes

9
votes

Solution #35: Use a more efficient compression algorithm

Written by dant1612 the 18 Dec 09 at 02:50.

use the xz utils package to compress the files instead, the author claims that it produces archives up to 30% smaller than gzip (although decompression speed would probably be sacrificed)

http://tukaani.org/xz/

(Its still in beta, but I've noticed quite a few GNU programs have started providing source packages in the .xz format)

5
votes

4
votes

Solution #37: Let the user compose their own CD from selectable components

Written by 10110111 the 7 Mar 10 at 13:08.

Make a program which would let the user select the needed packages (or use defaults like official Ubuntu compilation, Kubuntu, etc.), which then would be fetched from the repos and packed into a bootable CD/DVD image.
And leave the official CD as its maintainers want (including removal of some packages like GIMP).

Edit: hmm... it's sorta dup of #30...

See the 30 comments or propose a solution (latest comment the 13 Dec 11 at 07:11) >>

Import any device drivers written for Windows for Linux

Written by balaji the 28 Oct 08 at 00:57. Global category: Hardware support. New

The most difficult part is when you get a new piece of hardware and really love it because it has all these cool features and find out later on that it will work only on Windows. Nothing can be worse than that, especially if you have done away with Windows. This means there is a perennial fear in the minds of a Linux user that the day his computer contracts a brand new hardware for which the Linux community has not gotten down to write drivers, he will not be able to use his computer on it. Obviously MS does not even care if a new hardware product is released, because all hardware vendors give closed source drivers that can run with Windows.

Hardware vendors have their own reasons to keep their source code closed - we cannot force them to release the code. Often source code of drivers reveals information about their hardware that may be critical for them to compete. Obviously therefore, they don't agree with the OSS philosophy. But just because he has written his driver for Windows, should you have to change your operating system. You are now forced to choose between letting the hardware take over you and you give up Linux for Windows or, letting the OS take over you and forget about the $120 you spent on the fresh piece of hardware that you needed so badly and instead choose to continue on Linux (just because you are crazy about OSS). The second option is a sacrifice - and I don't think Linux users should have to pay the price for using the OS.

Legally, a person that buys a piece of hardware (say some scanner from some obscure manufacturer) has chosen the hardware. The hardware vendor cannot force him to choose the Operating System on his computer. But when a hardware vendor writes drivers only for Windows, he essentially forces Windows down the throat of his customers whether they like it or not. (This in itself should be a violation of the customer's rights.)

So how do we fix this? Let's see why manufacturers have a problem writing drivers for Linux. People often say that these are the reasons why Linux suffers for drivers.
1. There is no way he can figure which distro of Linux he should support.
2. He has no clue in what format Linux accepts drivers. There is no standard driver format like Windows has .dll or .drv.
3. Noone pays them to write Linux drivers. This probably is least of the problems. I don't think Microsoft pays hardware vendors to write drivers for Windows.

[....]

17
votes

Solution #1: Import Windows drivers into Linux

Written by balaji the 28 Oct 08 at 00:57.

Import the driver written for Windows into Linux. This fixes a lot of problems. Often vendors provide an installation disk with the driver written for Windows (because the manufacturer prefers to write in a known standardized format and also, perhaps he wants to keep his source code confidential). And Linux should simply import it for our needs.

This could be a hard task depending on how they packaged their driver. Was it contained in .dll/drv/inf files or does it come in some other format? Whatever be the format, they will for sure use some known (standard) format that will surely run in Windows. In other words, we should be familiar with this:

http://msdn.microsoft.com/en-us/library/ms809956.aspx

If we do that, we shouldn't need to rewrite drivers for every piece of hardware.

This is what the above will potentially lead to: Given any piece of hardware (even if it has drivers written only for Windows) all you have to do is run a piece of software that will extract the machine code of the Windows driver (which was shipped with the installation disk) and automatically build a Linux driver from that. Alternatively, we should provide a driver development kit (to all vendors) that works perfectly with the same code used to make an Windows driver. They can surely recompile their driver for Linux and ship the closed-source driver to you.

[...]

If the Linux community implements this idea (or at least makes a DDK like Microsoft - a slightly inferior idea to simply importing drivers), the hardware compatibility issue which people often have in Linux will completely disappear. At least your hardware vendor can never say: "It works fine in Windows, though."

Of course, none of the extracted drivers should directly affect the kernel. Instead, they should only be part of peripheral hardware drivers that are called by the kernel in a protected manner.

10
votes

Solution #2: Start 'Ubuntu certified' logos to make Linux compatibility popular

Written by balaji the 17 Jan 09 at 16:21.

The real problem however is in all fairness, the fact that hardware vendors often make products that use non-standard protocols. Reverse engineering these protocols is a daunting task.

One way to prevent this is to make open protocols popular by rewarding those that use open protocols supported by Linux. So for example if an HP printer carries a 'Compatible with Ubuntu Linux' logo, it will do a load of good:
1. It will attract customers that have only Ubuntu to choose such products that carry these stickers. Obviously, other hardware vendors will see the stronger competition and will try to get these logos themselves too. To do this, obviously these vendors will be forced to use open standards and protocols and favor them.
2. Canonical could charge a small fee to test the hardware on their platform and provide the vendor the 'Compatible with Ubuntu Linux' logo. This is one other source of income for Canonical too.
3. Customers will become aware of Ubuntu Linux. Awareness will make Linux a popular alternative to other OSes. Right now, we are not able to cross the barrier because the general public is unaware of Linux.
4. If however a vendor still chooses to have a closed standard protocol for his device, the competition will force him to write his own Linux drivers and package them in apt, yum or rpm packages or all of them. In fact such vendors are encouraged to provide a CD that will work on Linux and install their product and their driver for the OS. Such vendors that provide their drivers through CDs should be given a different sticker - 'Can be installed on Ubuntu Linux'. The logo should be different to make it clear that these products do not in fact support standard protocols but do provide packages that will install in Linux and you can still use it in the mention distro of Linux.

[...]

This sort of campaign will really help Linux build a strong base for hardware compatibility. Hardware will either be automatically compatible with the OS or can be installed using the vendor's CD.

The real problem however is in all fairness, the fact that hardware vendors often make products that use non-standard protocols. Reverse engineering these protocols is a daunting task.

One way to prevent this is to make open protocols popular by rewarding those that use open protocols supported by Linux. So for example if an HP printer carries a 'Compatible with Ubuntu Linux' logo, it will do a load of good:
1. It will attract customers that have only Ubuntu to choose such products that carry these stickers. Obviously, other hardware vendors will see the stronger competition and will try to get these logos themselves too. To do this, obviously these vendors will be forced to use open standards and protocols and favor them.
2. Canonical could charge a small fee to test the hardware on their platform and provide the vendor the 'Compatible with Ubuntu Linux' logo. This is one other source of income for Canonical too.
3. Customers will become aware of Ubuntu Linux. Awareness will make Linux a popular alternative to other OSes. Right now, we are not able to cross the barrier because the general public is unaware of Linux.
4. If however a vendor still chooses to have a closed standard protocol for his device, the competition will force him to write his own Linux drivers and package them in apt, yum or rpm packages or all of them. In fact such vendors are encouraged to provide a CD that will work on Linux and install their product and their driver for the OS. Such vendors that provide their drivers through CDs should be given a different sticker - 'Can be installed on Ubuntu Linux'. The logo should be different to make it clear that these products do not in fact support standard protocols but do provide packages that will install in Linux and you can still use it in the mention distro of Linux.

This sort of campaign will really help Linux build a strong base for hardware compatibility. Hardware will either be automatically compatible with the OS or can be installed using the vendor's CD.

-1
votes

Solution #3: Generic Printer driver for Windows

Written by fernandoc1 the 5 Jun 09 at 22:24.

The way I thought to solve this problem is to implement a windows driver that could receive the jobs and send them to a specific printer attached on windows.
In another words, the idea is to have a driver in windows that is shared in the network and it can direct jobs to a real printer.
I thought that it could be implemented like a PDF converter, and the pdf file can be submitted to a real printer to print the pdf generated by the "generic printer driver".
It couldn't be so hard to implement, since that we already have a printer driver that prints in a PDF file and the only thing that is missing is an option to make windows get this PDF generated and print it with a printer attached on it in consequence of a command from a Linux machine.

9
votes

1
votes

Solution #5: Windows Drivers

Written by ct2193 the 30 Oct 09 at 20:12.

For many reasons, native drivers exist. Substituting a Windows driver is unlikely to be a final or long-term solution. As such, IF such support was enabled, there must be a way to reliably run the installers. WINE could be an answer to this portion.

I am more interested in having Linux "interpret" a Windows driver .INF file in order to "understand" how to use a device.

If you have ever used a decompiler, you may have observed the technical ability to then alter and recompile the modified code, though decompiled code is frequently very complex, undocumented, and much less human readable that the original source code that was used to compile the executable that you just decompiled. In short, I think it's a huge step in the right direction being able to drop a Windows driver into some sort of Linux Driver Assistant utility and have it automatically (or as automated as possible) decompile, rearrange, and recompile a Linux compatible driver. This scenario would let the end user have access to otherwise Windows only hardware, while a native driver remains absent. This would also let companies perform the same steps, but then debug the driver as to package an "official" Linux driver more swiftly.

It would be the hardware / driver equivalent of WINE, but recompiled into something that wouldn't have the overhead of JIT compiling every time.

1
votes

Solution #6: Create a Wrapper for Windows device drivers

Written by cando the 6 Feb 10 at 16:43.

Why reinvent the wheel twice? And - looking at the othe proposed solutions - there is one issue not mentioned yet: Legal stuff! Before using any provided software, one have to agree to NOT reverse engeneer or hack or change the software provided.

So there is only one "legal" solution: use the de-facto industrial standard set by Micro$oft for devie drivers.

What does this mean to linux? A device driver wrapper needs to be built to allow machine code from a device manufacturer to comunicate to the OS kernel and vice versa.

Using that approach there will be immediate support for any hardware out there for all intel / amd based linux hardware.

I mean a kind of "wine" emulator for windows hardware drivers - so linux coud be enabled not just to run native windows code on the user layer - but support any hardware supported by windows with linux software on the user level.

Fortunately the Windows device driver model is well defined and specified / published so every manufacturer can develop device drivers for all kind of hardware to a standard interface definiton.

So why not build a plug to consume data of this interface and wrap this in a generic linux kernel driver. If there is a genuine linux kernel module - fine, than lets use it directly - if not - the wrapper can do 99% of the work.

[...]

And the quality of the drivers is almost guaranteed due to the "MS Certification Programm" for free!

2
votes

See the 12 comments or propose a solution (latest comment the 30 Nov 11 at 22:52) >>

Open Office: support for more than ~64000 rows, advanced mathematics

Written by billdotson the 4 Oct 09 at 18:50. Related project: OpenOffice.org Spreadsheet. Not an idea

Excel 2007 allows for more than ~64000 rows in its spreadsheets. Many people who use spreadsheets need more rows than that. For example, my father is an accountant and said he couldn't use it because of this limitation.

Also, Open Office spreadsheet needs to be able to do advanced things like detailed statistical analysis and other advanced mathematical operations, pivot tables, and anything else that might be needed by someone at work.

248
votes

136
votes

-7
votes

See the 8 comments or propose a solution (latest comment the 5 Oct 11 at 15:45) >>

Ubuntu Developer Network

Written by JReagan1990 the 17 Feb 09 at 16:19. Related project: ubuntu.com. Implemented

Just like Microsoft has MSDN, Sun has SDN, why not have a UDN? :)

We have MOTU, but still the wiki pages are scattered and not exactly the best guides on how to begin programming. Beginners would find it troubling on how to begin. Creating a developer community around Ubuntu would be extremely beneficial.

446
votes

Selected solution (#1): CREATE UDN!

Written by JReagan1990 the 17 Feb 09 at 16:19.

A page could be created on the main ubuntu.com site, and would include sections for "beginners", "experienced developers", etc. This page would have all the resources (or links to wiki pages) to get started developing applications for Ubuntu. Let's face it -- there is no easy guide to creating a simple package for beginning programmers, or guides that might be suitable for such a task. The page could also embed videos from the Ubuntu Developer's channel, and even host the Ubuntu Developer News (the weekly news for ubuntu devs).

It would certainly help improve the image of Ubuntu's development process and show that it is very active. The site would also make it more visible to mainstream developers (game companies, etc.), and it would also help to create a standard around Ubuntu.

Update - added a mockup (screenshot) over on my blog with a more complete writeup:

http://jonreagan.wordpress.com/2009/02/17/ubuntu-developer-network/

You can read the official plan here:

http://docs.google.com/Doc?id=dfdw2zz6_155gdk646df

[...]

You can also read the blueprint here:

https://blueprints.launchpad.net/ubuntu/+spec/ubuntu-developer-network

-171
votes

210
votes

47
votes

Selected solution (#4): Make a learn on your own style course

Written by asdlkf the 2 Mar 09 at 14:53.

While many people that develop for linux/ubuntu have formal education in programming, etc...

We can not expect users who do not know how to make a hello world program to contribute to bug fixing or creation of new content.

Far more work needs to be put into creating an organized, prerequisites-basted, comprehensive, and motivating site/videos/etc. Call it ubuntu classroom or something. Maybe it could be part of edubuntu or just in some way related.

EVERY topic should be touched on and there needs to be quality control and granularity to the subjects included.

A very tree-like structure should be used.

"How to program for ubuntu"
"How to Program for ubuntu in C"
"How to Program in C"
"For loops in C"
"If-then control structures in C"
"binary tree searching in C"
"Basics of data structures"

[...]

"Basics of data types"
"How to define data types"
"How to program for ubuntu in php"
"How to program for ubuntu in ..."
"How to port programs to ubuntu from Windows"
"How to port programs to ubuntu from windows in C"
"How to package your program for use with ubuntu"
"How to compile the program for use in a package"
"How to compile the program"
"How to configure an IDE"
"How to obtain/select an IDE"
"How to locate include'ed source packages"
"Settings to use for generic distribution"
"Where to submit completed packages"
"How to get help with your program(ing)"
"Where you can go to find help"
"Where you can go to find help online"
"Where you can go to find help IRL"
"Where you can go to find help in class"
"How to decipher compile errors"
"How to troubleshoot logic errors"
"etc... etc... etc..."
"etc... etc... etc..."

Obviously this list could be 1000's of lines long. But the basic structure should be:
Item
-List of items which [Item] is dependant on
-Information on where/what/how to use [Item]
-List of items which depend on [Item]

This way, when people link to a tutorial "oh, just go here and see how to program in C", they are also linking to all of the tutorial information someone would need to know how to program in C reguardless of prerequisite knowledge.

By providing a complete framework for getting started at ANY point, anyone with any level of educational background could say "I want to learn how to make a program in C and package it for use in ubuntu". Then find how to program in C. Then read how to use data types/control structures/GUI's/etc..

It might take them a few days/weeks/whatever to read up to speed, but in the end, they could do it, for free, open.

-- Chris

While many people that develop for linux/ubuntu have formal education in programming, etc...

We can not expect users who do not know how to make a hello world program to contribute to bug fixing or creation of new content.

Far more work needs to be put into creating an organized, prerequisites-basted, comprehensive, and motivating site/videos/etc. Call it ubuntu classroom or something. Maybe it could be part of edubuntu or just in some way related.

EVERY topic should be touched on and there needs to be quality control and granularity to the subjects included.

A very tree-like structure should be used.

"How to program for ubuntu"
 "How to Program for ubuntu in C"
  "How to Program in C"
   "For loops in C"
   "If-then control structures in C"
    "binary tree searching in C"
     "Basics of data structures"
      "Basics of data types"
       "How to define data types"
 "How to program for ubuntu in php"
 "How to program for ubuntu in ..."
  "How to port programs to ubuntu from Windows"
  "How to port programs to ubuntu from windows in C"
 "How to package your program for use with ubuntu"
  "How to compile the program for use in a package"
   "How to compile the program"
    "How to configure an IDE"
     "How to obtain/select an IDE"
    "How to locate include'ed source packages"
   "Settings to use for generic distribution"
  "Where to submit completed packages"
 "How to get help with your program(ing)"
  "Where you can go to find help"
   "Where you can go to find help online"
   "Where you can go to find help IRL"
   "Where you can go to find help in class"
  "How to decipher compile errors"
  "How to troubleshoot logic errors"
  "etc... etc... etc..."
 "etc... etc... etc..."

Obviously this list could be 1000's of lines long. But the basic structure should be:
Item
 -List of items which [Item] is dependant on
 -Information on where/what/how to use [Item]
 -List of items which depend on [Item]

This way, when people link to a tutorial "oh, just go here and see how to program in C", they are also linking to all of the tutorial information someone would need to know how to program in C reguardless of prerequisite knowledge.

By providing a complete framework for getting started at ANY point, anyone with any level of educational background could say "I want to learn how to make a program in C and package it for use in ubuntu".  Then find how to program in C. Then read how to use data types/control structures/GUI's/etc..

It might take them a few days/weeks/whatever to read up to speed, but in the end, they could do it, for free, open.

-- Chris

25
votes

6
votes

Selected solution (#6): Like Solution #1 and #4 but also with mentors

Written by trumbun the 22 Apr 09 at 15:01.

Sometimes new programmer would like to help but it is difficult to find a project to help in it.

Their should be a page with projects that Ubuntu developers help in.

A subscription page for new programmers that wish to help with their knowledge (like a CV)

When a project is assigned to the programmer he/she can talk to a mentor using voice chat to help out.

See the 14 comments or propose a solution (latest comment the 9 Jul 11 at 16:51) >>

Offer to take a picture for the user account through webcam

Written by gregbenson314 the 21 Feb 11 at 22:03. Related project: Gnome. New

When you go to system --> preferences --> about me, there is no option to take a picture with the webcam to use as the user account picture

247
votes

Solution #1: Add a "take photo now" option

Written by gregbenson314 the 21 Feb 11 at 22:03.

When you go to change the user account picture (system--> Preferences--> about me) and click to change the icon, a nautilus file browser pops up, and allows you to choose a picture.

I think it would be good to have a "Take a picture now" option, that uses the webcam to take a picture. If the user likes it, it could be used as the account picture. If the user doesn't like it, another picture could be taken (discarding all previous attempts) until the user is happy. They could cancel at any time.

See the 3 comments or propose a solution (latest comment the 22 Mar 11 at 03:24) >>

Synchronize desktop wallpaper with xsplash background

Written by phrizek the 17 Nov 09 at 17:51. Global category: Look and Feel. New

Once you change your wallpaper on on your desktop, the overall look and feel of ubuntu becomes inconsistent when it concerns the xsplash background and the user's desktop. This leaves the user to find hacks and workarounds to make their desktop theme consistent.

397
votes

Solution #1: Make a checkbox to easily synchronize desktop wallpaper with xsplash background

Written by phrizek the 17 Nov 09 at 17:51.

Take a look at this script here:
http://www.gnome-look.org/content/show.php/desktop+background+as+xsplash+%2B+gd m?content=114984&PHPSESSID=183f5f0adfa276aa6310278b4caa0979

It enables you to change the xsplash background to automatically match your current desktop wallpaper. It would be fantastic if Lucid Lynx had a similar feature installed by default. There could be a checkbox in Appearance Preferences under the background tab that allows you to quickly enable or disable this feature. Having such a feature would not only make ubuntu feel more personal, but would give a more consistent desktop appearance as well, from splash to desktop.

This feature could be disabled by default and left up to the user to enable.

111
votes

Solution #2: Make another tab where the user can select a xsplash background

Written by Peter-Alexander the 24 Nov 09 at 01:20.

The comments are good and I like the idea of changing the splash screen so the easiest would be a new feature:
The user can customize the splash screen.

Benefits:
- the user would know the image he has set.
- the contents of the tab can be greyed out and need to be unlocked which is only possible via permissions and can be locked by root for multiuser systems.

-39
votes

46
votes

29
votes

4
votes

See the 8 comments or propose a solution (latest comment the 18 Jan 11 at 22:22) >>

Mouse clicks with touchpad should be enabled by default in Ubuntu 9.10

Written by Herakleitoszefesu the 3 Oct 09 at 03:48. Related project: Gnome. Implemented

Because it's not easy to use Ubuntu (Netbook Remix) 9.10 Beta on a laptop (no mouse plugged in) with this option disabled - I can't perform (emulate) left mouse button click by double tapping the touchpad area.

520
votes

165
votes

-128
votes

-117
votes

-74
votes

-42
votes

See the 8 comments or propose a solution (latest comment the 27 Oct 09 at 06:10) >>

1 2 Next >>

Submit your idea