How Would You Refocus Linux Development?

buddyglass writes "The majority of Slashdot readers are no doubt appreciative of Linux in the general sense, but I suspect we all have some application or aspect of the platform that we wish were more stable, performant, feature-rich, etc. So my question is: if you were able to devote a 'significant' number of resources (read: high-quality developers) to a particular app or area of the kernel, and were able to set the focus for those resources (stability, performance, new features, etc.), what application or kernel area would you attempt to improve, and what would aspect you focus on improving?"

Make everything "Just Work"

[vonFinkelstien]

Find out all the things at take too many clicks, or require editing text files and make them "Just Work" in a simple and easy way.

Re:Three things.

[__aawkdb2598]

• Visual coherency and a refined GUI. Taste in UI's vary between people, but most linux GUIs that aren't very minimalist tend to suffer from wasted space.
• In interests of making linux more accessible, more configuration utilities that don't require specific knowledge and in-errant editing of configuration text files.

Re:Three things.

[pcnetworx1]

If I could refocus Linux Development... I would try to pool all the development into 1 distro to reduce duplication of so much effort.

The Hurd

[John Hasler]

if you were able to devote a 'significant' number of resources (read: high-quality developers) to a particular app or area of the kernel, and were able to set the focus for those resources (stability, performance, new features, etc.), what application or kernel area would you attempt to improve, and what would aspect you focus on improving?

I'd budget $1M/year for a minimum of five years for full-time work on the Hurd. No, it isn't Linux but it is an alternative kernel with interesting features that is sadly stagnating.

WINE

[jstomel]

Admit it, wine sucks and there are lots of programs that will never be ported. I want wine to be integrated and almost invisible, like the Classic interface in OSX.

The Elektra Project

[Thaidog]

http://elektra.g4ii.com/Main_Page [g4ii.com]

I think it's at least worth trying such an implementation. Ok... now bring on the "It Windows again" haters...

Re:Might I Suggest...

[jeevesbond]

This topic is 12 minutes old and three post have already suggested we bury the command line; part of what makes Linux so fast, flexible and customizable is access to virtually every setting from a text editor.

Whilst I totally agree with what follows after the semi-colon in this sentence am not so sure about the part prior to it. All we're seeing is that people do not want to be forced into changing settings--am assuming, on their desktop machines--using the command line. This does not mean we should 'bury' the command line, or stop using text files to hold settings! In fact you've made my point for me:

If you are looking for a completely GUI drive *nix I would say OS X is your best bet (yes, I know you can use the CLI in OS X, but you never have to unless you so desire).

Aye, there's the rub! The user should be able to choose between a GUI configuration interface or editing a text file: everyone's a winner! Also a GUI should be able to read/write text configuration files whilst handling seperate user changes to those files gracefully.

In fact I'd spend a lot of the money on getting everyone (or as many projects as I could) to agree to a configuration file format that could easily be interpreted by an application. A one-size-fits-all library could be written to get the settings from file into memory and back again, then it would just be a matter of organising that data into a front-end that's meaningful for the user. The real joy is that with a standard file format, and library to support it, a rudimentary GUI for a new application could be created in minutes.

This is not something that needs to be changed, instead change your mindset that this is not Windows.

This is a very conservative viewpoint, why can things not change? Why can't we have the best of both worlds, with both GUI configuration tools and text files?

Re:Might I Suggest...

[oatworm]

Things I've had to hit up the command line for and wouldn't put up with from Windows or OS X on my Ubuntu Feisty Fawn installation:

Installation - I have a Compaq Presario V6000-series laptop with an AMD chipset. Apparently, it has some BIOS issues that require special flags in the initial boot loader screen on the Ubuntu install CD. To find these flags, I had to hit the Ubuntu forums; there was absolutely no way that I could've known which flags to plug in unless I had prior experience with such issues. Ideally, Ubuntu should have figured out that I might have a troubled AMD chipset and provided the correct flags for me. This could probably fall under "hardware support". Alternatively, a quick GUI that listed what advanced options were available to me would've been helpful, too. Heck, even BSD's boot loader will give you a basic list of commands to muck around with, and it's text-based. That would've been more helpful than Ubuntu's CD.
Wireless - I know wireless is a big issue on Linux, especially with Broadcom devices like the one in my laptop. That said, getting wireless working eventually required the use of ndiswrapper, and, if there's a clean GUI for that on Ubuntu (not saying there isn't - didn't look all that hard), I didn't see it recommended. Again, this is probably as much a "hardware support" issue as anything; ideally, the drivers would've already been in Linux and we wouldn't have these problems.

Other wish-list items:
Drivers - Look, I get the Linux driver model and why some people find it great. It's wonderful that, when I plug a USB mouse into my laptop, it immediately works and doesn't play the "10-Mississippi" game that Windows XP plays, even on initial login. I understand that's because all of the drivers are pre-compiled and included into the kernel. Great. That said, if I decide to run the most up-to-date video drivers for a Linux-based computer and I don't hit up their respective repositories (something Ubuntu is getting better about, by the way), why must I recompile my drivers each and every time I install a kernel update? I don't mean a technical "why" - I know WHY, technically, and I even get the philosophical WHY, in that I understand it has to do with keeping open-source halfway open. That said, from a practical standpoint, I think that the first distro that comes up with a halfway decent driver API to link against is going to be the first desktop Linux distro that becomes ready for prime-time, because, from that point on, if someone decides they want bleeding-edge drivers, they won't be punished with repeated driver recompilations after every kernel update.
GUI - Don't get too carried away on having every single function in the GUI. Microsoft already does that and it's a royal pain in the ass. In fact, though I like some of the functions of Powershell, I think it's another fine example of Microsoft trying to do too much with their widgets. Sometimes less is more. If I have to keep hitting the '?' key to figure out which one of the 8,000+ verb-noun commands I want to use, that product is going to be less powerful than the one with only 500 commands to deal with that I actually have a halfway decent chance of remembering. That said, I'd still like to be able to do basic administration on a Linux machine without hitting up the command line. I'm okay if mastering administration requires shell-scripting skills, but let's keep the basic stuff... basic.
Regression Testing - The latest kernel update for Feisty Fawn ruins ACPI, making it where my laptop doesn't know whether it's on battery or not. More than anything, I'd like to throw more resources at preventing these sorts of shenanigans - heck, let's do this for Windows, too.
VIM configuration GUI - But only because I'm feeling ironic. :-)

A bit of biz, a bit of fun

[dorath]

First, I'd like to see something that really takes on Outlook/Exchange.

Second, I'd like to see something along the lines of OpenX: a competitor for DirectX. Sure openGL competes with Direct3d, and there's openAL for DirectSound... But why not bundle up a bunch of open goodness the way DirectX packages together graphics, sound, networking, and input? Wrap it all up in a nice tight IDE and what've you got? Something that'll write games for Linux, Mac, Windows, Wii, and PS3. Compare that to DirectX, which does Windows and xBox, and frankly the only hurting party is the xBox. Sure, sure, it's more complicated than that, but damnit, that's what I want to see.

Re:Three things.

[number11]

Ubuntu wins on all of those points except 5.

Unfortunately, it still has a way to go on #1 (have it autodetect all my hardware and just WORK). I handed out a few Ubuntu CDs.

On the first (a cheap Acer laptop) it complained about the wireless chip. No real options for the user, other than to get a new laptop. But it did work.

On the second, it attempted to make a W2K machine dual-boot, and (going with all the defaults) instead made it no-boot (Win BSOD). The user then attempted to reinstall Windows, which blew away everything. Now, I'll grant that the user made things worse, but what would you expect them to do? And how would they know?

On the third, I attempted to test what had happened on the second, by just grabbing a random used hard drive (containing some version of Windows) and installing Ubuntu on it, accepting all default choices. Ubuntu choked. (Possible causes: multiple partitions, fragmentation, Win swap file in the middle of the area Ubuntu wanted.)

So I reformatted the drive and did a clean W2K install. Ubuntu installed over that very nicely.

So, my impression is that Ubuntu is not really ready for the general public to install, not on a Win machine that's been in use. At the very least, not until it knows how to deal with fragmentation, Win swap files in inconvenient places, and the like. I'd even settle for a message like "Fool, this HD is hopelessly fragmented, fix that and move the swap file, before Ubuntu can help you." Bonus points for instructions on how to do those things. Double bonus points for creating a batch file that will run on reboot and do those things for you.

Don't get me wrong, I think Ubuntu is the best of the lot for the naive user. But the install isn't ready for prime time yet.

Areas of focus

[Anonymous Coward]

Slanted towards gnome, but nonetheless.
1. OpenOffice base: Make it as good as ms-access, as easy to export to the web, and as versatile of a reporting tool.
2. Get Out of tree drivers in tree, especially in the video area. LIRC should be in mainline (Or something that handles all of it's drivers) same for spca5xx.
3. Video and voice IM chatting (There's no reason for this to be missing, as all of the support is coded, it's only integration work).
4. Flash reverse engineering: This is especially critical for 64 bit and ppc.
5. Webkit everywhere. Gnome uses 3 html libraries, kde still doesn't use it. From what I understand embedding mozilla is a pain, as you have to send everything beyond the simple stuff with xpcom. Everyone seems to think webkit takes care of it. I'd just be happy with 3 less libraries.
6. DVD/VCD playing in totem for unencrypted discs (With menu support) Right now I have to install xine for dvds only. The elements are in gstreamer (though not maintained) This is just another duplication.
7. Integrate openmoko with linux as much as windows mobile is with xp (Playlist syncing, drive mounting, contact/dates/email/bookmark syncing)
8. Remove Bonobo from gnome (From what I see nobody uses it as it was intended and there are 3 people who understand it, and with this status it does the opposite of what it's supposed to).
9. iDVD: Make qdvdauthor or dvdstyler as easy as idvd.
10. garageband: make jokosher or the beast as easy as garageband.
11: Easy: Give me different wallpapers on different desktops.
12: Make the gnome dock cool again, give me avant window navigator or gimmie by default.
13: Integrate im/irc/email/contacts. Make it work. Right now telepathy has 80% coverage of protocols, and crashes for me so I can't even try the integration w/evolution.

Re:Three things.

[arivanov]

All of these are not that bad already. Granted they can be better, but it is a case of diminishing returns. To make them better you have to invest a lot for a little return. Also, as Linux is used more and more in embedded devices serving this purpose many of them will improve anyway without throwing effort at it.

Also, IMO Linux actually does the job for an average desktop more or less all right already. I am judging by the number of complaints I get from my significant half which is a Mac user when using any of the machines at home. They have been diminishing towards 0 lately.

Where Linux/OSS sucks royally is business use. There are plenty of areas there which will deliver Linux to a bigger audience with much less investment than any of the UI improvements and support for fancy hardware which linux supports better then windows anyway. The last 3 pieces of really obscure hardware I threw it at it just worked.

• Linux/OSS needs good diagram drawing program support including basic visio import/export (without the executable extensions and VB integration). Dia has significantly improved over the years and it currently approaches Visio as far as features. Some of course are fundamentally different like python integration instead of VB integration, etc. It is still way behind visio on ease of use for the basic UI (when it comes to advanced stuff they both suck bricks through a thin straw sidewise). Also, dia integration into openoffice is inexistent. What made visio the de-facto corporate diagram standard is the integration into office. Dia has none. It does not even have a suitable export which can be imported into openoffice as vector graphics. While at it - openoffice support for internal graphics drawn using impress and/or writer. These leave a lot to be desired as well.
• OSS needs good client-server project manager. There is no need to go for a standalone UI project management software (it is not the route everyone else in the market is is going anyway). Server-client does best for things like this. There is a number of hacked together serverside projects which are desperately asking for a non-Web or Ajax UI. Here OSS can also jump straight ahead of the MSFT in the game if it integrates it to issue tracking and CRM where OSS is way ahead of the game (MSFT still does not have it and is least likely to).
• Contact management integration and contact management in all PIM apps must improve before Joe Average can use it. At the moment most of linux PIM has finally reached some form of useable state and linux itself can sync to most devices like phones and PDAs (if everything else fails Bluetooth usually works). Still, setting it up and getting it going is often outright painful. There are also annoying glitches all over the place.
• Correct timezone processing in calendaring applications must improve. Both Evolution and Sunbird suck at this royally. One forward or import export and the TZ of your calendar gets completely and utterly screwed. Granted MSFT has this f*** up as well so OSS apps are not alone here (dunno when will the developers learn to use universal time for all time storage).

Any one of these will

Tibetan characters

[jjohn_h]

Let the keyboard behave the same in text mode and X11.

Let the default keymap be the same in text mode and X11.

Allow all those combos (ctrl-home etc.) to be identifiable by default in text mode.

Get a new maintainer for keyboard/keymap issues.

Let Andrew Brouwer be happy with Tibetan characters in Unicode.

Re:Three things.

[bateleur]

Emacs is a great editor (pretend I said "Vi" if that's your thing) and has some things that are nice for coding, but I'll take a true IDE in a second.

Not only that, but speaking as someone who loves Emacs, *nix shells and so on... I still run Windows as my desktop OS of choice because Emacs runs fine on Windows, Cygwin deals with my command line needs and (crucially) all the proprietary apps and weird hardware I own just work.

Back when I was a student I ran a dual boot machine and would try running Windows content using Wine and the like. Now? I just don't have the time to wrestle with all the problems that arise. I still run Linux on my server and it's great. But on the desktop I haven't even bothered to install Linux on my newer machine.

Re:Make everything "Just Work"

[Athaulf]

I'm hoping I won't have to contact the developers because I'm hoping to create a user supported repository, so anyone could make a configuration utility for their favorite programs and put them up for everyone else to use. I wondered why there didn't seem to be one out there already, but maybe I should have looked harder. Also, what do you mean by "dumb down" the user interface? Considering the fact that there isn't a user interface for configuring many server programs like samba, apache, and vsftpd, I'm not sure this is possible.

Re:The Hurd

[Slashcrap]

I'd budget $1M/year for a minimum of five years for full-time work on the Hurd. No, it isn't Linux but it is an alternative kernel with interesting features that is sadly stagnating.

No, it doesn't have interesting features. If it had interesting features it would not be stagnating and lacking in developers. OpenBSD is an alternative kernel with interesting features and that is why it has enough developers and support to be usable.

There are no end of unfinished OS projects with a couple of developers that move to a completely different kernel every six months and will never, ever be finished. I don't see how the Hurd stands out from them. I foresee your $5M funding nothing except a huge amount of meetings, committee decisions, politics and rewrites. You'd be better off selecting a random microkernel based OS project on Sourceforge and giving it to them. They might actually produce something.

I don't want to discourage anyone from working on the Hurd if they find it interesting, but giving them $5M and expecting something useful to come out is just ridiculous.

wine and installation

[Cyko_01]

as a linux noob I found that installing programs that were not listed in the default repositories to be very difficult. I think an MSI type installer would be extremely useful. I was so happy to discover autopackage and I would LOVE to see it more widely used for non-repo apps.

The other thing that I think needs improvement is WINE and game compatibility. Lets face it, commercial games are not being developed for linux. we need a way to make them work as well as they do on windows so that "linux has no good games" can no longer be used as an excuse not to switch. we should be able to reply "linux runs almost all the games that windows does"

Re:Three things.

[Eli Gottlieb]

While we're at it, how about a configuration and build system that doesn't require me to learn a medium-sized macro language just to distribute a package that will compile and run on other people's machines? Maybe we can be really nice to ourselves and even invent such a system that doesn't care what language I coded in!

Re:UI needs a lot more consistency

[DamnStupidElf]

Why is it that all the developers seem to be able to code to a standard API - but they can't even come close to agreement on the way a program is operated? Maybe it's time to create a UI standard for Linux apps?

Probably because there are no universal UI guidelines that fit in every case, and no one knows of the "best" way to design a UI. There are almost certainly better UI models that will be found in the future, and part of open source development is trying to discover those models. Modal dialogs with "Yes, No, Cancel" may not be part of the optimal solution (it's hard to imagine them being the best for any task, really), so exploration into other representations and interaction models is a good thing.

For one thing, modal dialogs are essentially points in the program flow where the user is forced to make a decision. In many cases, the decision can be made automatically or simply removed from the program flow if the correct UI model is used. All the dialogs warning about saving documents before closing programs are probably the worst thing to happen in UIs since their creation. The proper solution is to save and version everything the user does, and never lose work (to the maximum extent possible) when the application is closed or crashes. That requires support from the filesystem and a UI design that emphasizes the versioned nature of files and data. Forcing users to "log on" and "log off" and closing all their applications and opening them up again is another genuine UI mistake that's a holdover from the days when memory was so tiny that overlays were popular and it was simply a necessity of operating a computer. There is no reason that modern operating systems should not present a persistent interface to the user. Hibernation is the closest thing to persistence that's currently available, with suspension a close second, but neither are truly persistent. They still rely on the old model of individually managing each piece of data.

Better hardware support..

[Gnodab]

Better laptop support. I've tried three different distributions on my new Pavillion Laptop, and not one has worked without massive tweaking and command line editing...not exactly a great incentive to switch.

Better wireless support...

actually you know what? Better hardware support all around. Make it as easy as Windows to set up. also a universal install file (like .exe on windows)

Outlook Calendaring Replacement

[egarland]

Linux needs to be able to dislodge Microsoft Office in order to become a viable option for large-scale replacement of Windows in the corporate environment. Open Office can handle most of the document creation work and Firefox/Thunderbird do a great job at handling web browsing and email.

What is needed now is something that can integrate with an Exchange server's calendaring and also integrate with a robust open-source calendering server system to replace Outlook's calendaring functionality with an open, standards based system. We need to embrace and extend Outlook.