‘Til All Are One

Carla Schroder from Linux Today repeats a question that I’ve heard asked many times:

“Here we go with another round of Linux Today reader comments. Let’s start off with an issue that has been on my mind: Vendors who boast of the their Linux-based devices, but they only support Windows and Mac clients. It’s a step in the right direction, but would supporting Linux clients be so difficult?”

There are two major mistakes that are often made in considering this question:

• that all “Linux” systems are the same
• that by using Linux in one place, it only makes sense that you support other “Linux” systems

We need to remember that the only thing most of these devices share with a desktop “Linux” system (or even with each other) is the kernel (i.e. the precise definition of “Linux”). The userland is different, and there’s a lot of their own proprietary stuff on it too. Even the hardware (such as CPU architecture) is often wildly different. I think people have grown to think it’s all the same since we call it all “Linux”, but it’s not.

Because of this practical conundrum (as totally distinct from any philosophical or other arguments), I have some sympathy for those who prefer to call the system we use on our desktop and server systems “GNU/Linux”.

Argue all you want about its accuracy, but the fact is that it is far more accurate than merely using the kernel name as nomenclature for the entire OS. It specifies a userland that with the kernel comprises a workable operating system. Come up with a better name if that makes you feel more comfortable.

This opens up a whole can of worms. If I’m an applications or device developer and I announce “Linux support”, what do I mean? Will it work on my mobile phone? On my television? Probably not. Chances are it refers to particular versions of particular distributions for a particular architecture.

If I produce a device that is based on “Linux”, what relation does that have to other “Linux” systems? None. It’s not just devices: another major culprit is Web services. Linux runs most of the Internet, but many online services are not compatible with desktop Linux systems.

The reasons for this are simple:

• correlation does not imply causation
• the small market size of desktop Linux users

The first point relates to what I said earlier, that there’s no connection between the use of Linux on servers and devices versus its use on desktop computers. The usefulness of Linux on servers and devices is firmly recognised in many sectors.

The same cannot be said for desktop systems, despite what we may wish. If it costs a developer more to support a tiny market, they are probably not going to do it. That’s just business. Companies that choose to support desktop Linux often do so for other reasons, such as to foster a developer/fan base or tap into a very specific set of users.

So everyone, I share your frustrations that many so-called “Linux”-based devices/services don’t interface with my computers, but I keep in mind the points made above.

LotD: NSW Police: Don’t use Windows for internet banking (iTnews)

The OpenJDK plug-in that comes with modern distros is usually very good at handling Java in Web pages, but some applets are just stubborn. Thankfully, Sun have finally (after over six years!) released a plug-in for x86_64 Web browsers.

I managed to get the JDK version working on Fedora 11 and CentOS 5.3. Here’s the process.

1. Firstly, download the JRE or JDK from Sun. You’ll need to get version 1.6 Update 12 or above. I got the RPM version.
2. Run the install script to extract the bundle. On the RPM version, this automatically installs it to your system if you run the script as root.
3. Execute this in a terminal:

   # ln -s /usr/java/default/jre/lib/amd64/libnpjp2.so /usr/lib64/mozilla/plugins

   This part took me a while to work out, as I was looking for a file called libjavaplugin_oji.so, the name of the x86_32 version.

4. Restart Firefox and type about:plugins in the location bar to check if the new plug-in has been accepted.
5. Enable the plug-in: Edit ? Preferences ? Content tab ? tick Enable Java
6. You can test your plug-in at java.com and javatester.org

Last month I proposed that the FOSS community create an integrated software installer for Windows and Mac OS that only included FOSS applications. If Google can make Google Pack, I opined, why can’t we make a FOSS Pack?

As I had expected, my idea was already realised, at least in part. WinLibre and MacLibre provide a menu of free/libre software packages for the user to choose from, and can automatically install them for you.

That’s a big step in the right direction, albeit not the beauty we have on GNU/Linux through tools like Add/Remove Applications and apt-url. It haven’t tried them (I rarely use Windows and I don’t have a Mac), but here’s what I think they need to truly shine (based on my last post on the subject):

• an updates management service, that automatically checks for available updates and installs them for you
• an ability to cleanly remove the software just as easily as it was installed
• a file system scanner that recommends FOSS software to install, based on the software and file types it finds on the hard drive

Just for a second, put yourself in the shoes of an average PC user. You use the software that came with your computer, plus perhaps some others that you downloaded, bought in a box or ‘borrowed’ from a friend. You’ve heard some good things about something called “open source”, but you haven’t the foggiest clue of where to get it or what applications to try. You aren’t a technical person, have limited time and even less patience. Ultimately, you’re looking for something that ‘just works’ and is either free (of cost) or clearly better than what you’re using now. Why make the effort otherwise? Honestly, you’d rather be down at the pub watching the cricket with your mates.

How would free software advocates best woo such a person into their camp? They aren’t going to immediately repartition their hard drive and use GNU/Linux exclusively. They would more likely be willing to try some free software on their existing OS, provided that the barrier was sufficiently low. If you’re lucky, that toe-dip will lead to deeper immersion in the world of FOSS, and hopefully also into some appreciation of the philosophy beyond the practical.

If this person has a knowledgeable friend or pays attention to certain information sources, they might get some ideas on what software to use. Applications like Firefox and OpenOffice.org are fairly popular choices these days, but what about less publicised treasures like the GIMP or ClamWin? Sure, there are Web sites that let you search for FOSS equivalents to proprietary applications, but these still require some effort:

1. Search for the application you want.
2. Go to the Web site for that application.
3. Find the download page and pull it down.
4. Run the installer.
5. To uninstall, use Windows’ Add/Remove Programs.

These steps need to be performed for each application you wish to install, so can become tiresome very quickly.

How could we simplify this process? What I propose is a software management application. Let’s for the sake of brevity call it FOSS Pack, named after the closest analogue I can think of, Google Pack. The process is intended to be as simple as possible for the end user:

1. The user downloads a single application (FOSS Pack) and installs it.
2. When they launch FOSS Pack, they can select from a menu of categorised FOSS applications to install, similar to how a GUI package manager front-end works on (GNU/)Linux.
3. The user selects the applications they want, and then they are downloaded and installed in batch.
4. Uninstallation should be as simple as installation, all within FOSS Pack.

Here’s the killer feature: FOSS Pack should be able to scan the user’s system for proprietary applications. These are identified based on an internal list, which also contains information on FOSS alternatives to those applications. Those alternatives are presented for easy download and install.

FOSS Pack contains descriptions of each application, so the user doesn’t have to visit another Web site to understand what they do (although a hyperlink should be provided as well). The option should exist to be able to select only from applications that have Linux versions, as a means of facilitating an OS transition. FOSS pack should also be able to automatically check for updates at regular intervals, and offer to install them when available.

I’m not expecting any of this to be as clean as a real package management system. FOSS Pack will likely have to execute the external installers. Perhaps in the future the applications authors could co-operate with FOSS Pack maintainers to deliver a more seamless experience.

It looks to me that a lot of the pieces to create FOSS Pack are already there, and as is often the case in the FOSS world all that’s required is to tie them together in an appropriate way.

LotD: 30 Things That Are the Same In Microsoft Word and in OpenOffice.org Writer

Two weeks ago, we had the Education Expo.

Here’s my report, as co-ordinator of the Linux Australia stand:

Education Expo
Sat 14 to Sun 15 June
Rosehill Racecourse, Sydney

The Education Expo is an annual trades show targeted towards the K-12 educational space. Visitors consist of families and educators. Linux Australia once again had a stand, with volunteers spreading the word about free and open source software.

As always, we were very successful. With each passing year, the level of awareness of FOSS noticeably improves. Whereas at previous shows we would spend much energy expounding the basic concepts of FOSS/Linux, this year most people had either heard of it or were already using FOSS products such as Firefox and OpenOffice.org.

One thing we did differently this year was place more focus on FOSS running on Windows. Our past efforts have been meet with some resistance, as installing a different operating system posed a barrier to entry that many would not surmount. We had plenty of copies of the OpenEducationDisc to distribute, in addition to Fedora, Ubuntu, Edubuntu and Mandriva.

The fact that the NSW Dept of Education is migrating over 40,000 PCs across the state to OpenOffice.org was a useful selling point as well.

Our marketing efforts have been improving with each event. Our message is becoming more refined, and our leaflets are becoming more relevant. On the technical side, FOSS is becoming easier and more accessible, with projects such the aforementioned OpenEducationDisc and Wubi leading the way.

Our Web presence is improving, too. It’s far easier to point a newbie to just one easy-to-remember URL instead of confusing them with a list. In addition, I built an education portal for Linux Australia just in time for the expo.

There were at least two other stands that were FOSS-friendly. In fact, one of the largest stands were demonstrating their Web-based software product on about ten computers, all of which were running Ubuntu. Other stands expressed real interest when approached.

Other highlights of our presence included:

• OLPC XO laptops (from OLPC Australia)
• Intel Classmate PCs (from Mandriva Australia)
• ASUS Eee PCs
• laptops showing Edubuntu

Rodger Dean has some photos of the event.

A big thanks to everyone who helped at the stand:

• Ashley Lynn
• Ashley Maher
• Brendan Puckeridge
• David Andresen
• Gloria Arnold
• Harrison Conlin
• John Arnold
• Megha Kanth
• Pia Waugh
• Rodger Dean
• Vicki Burke

A special thank you goes to Melissa Draper, who was instrumental in ensuring the success of the stand in more ways than one.

LotD:  Insurance company bets health on open source — I’m quite heavily involved in this project, so needless to say I’m proud of what we’ve achieved

Those who remember my ancient quest for the perfect IRC solution might be interested in these posts by Aaron Toponce explaining how to couple a remote irssi session with GUI notification. I’m still quite happy with my current Bip + Xchat combination, but I’ve always lusted after the 1337ness of irssi. Icecap looks intriguing, but my first instinct tells me that their solution is over-engineered.

Note: If you see duplicated words in the above post, I am aware of them. Wordpress is doing something funny and I can’t figure out what it is. When I get the time I’ll upgrade to 2.5.

LotD: Ubuntu theme for Symbian S60v3 (works on my Nokia N95)

Simon and Lindsay: EasyTAG is indeed a useful tool for tagging many music files at once. While EasyTAG does automate a lot of the work, it is still quite a laborious process. This really grinds when you’re trying to manage a large music collection. What if your tagger worked more like your ears and brain — it just listened to the music and worked out what song was playing?

Enter Picard, stage left.

Picard ‘listens’ to your music and ascertains an audio fingerprint of each track. Using this information, along with more traditional data such as existing filenames and tags, it consults various online sources to deduce the details of the track and populate the metadata fields. I’ve found the results to be amazingly accurate. Sometimes it finds multiple matches, and it can occasionally get confused if the same track is available on different albums (e.g. a single, an original album and a ‘best-of’ compilation). If you have some idea of what the track is, you can lend Picard a hand by manually adding a more useful filename or some tags. This is where EasyTAG works well with Picard, since Picard isn’t geared towards manual tag editing. Still, it’s bloody impressive nonetheless.

As an album-based tagger, Picard behaves somewhat differently from file-based taggers like EasyTAG. It can take some getting used to, and it might be less accurate for people who prefer to collect single songs and not entire albums. If you’re like me and do compile full albums, it can do clever things like ascertain that you have the ‘White Album’ (or part of it) if it sees ‘While My Guitar Gently Weeps’ as well as ‘Revolution 9’. The developers have recognised that the UI does need some love, but once you’re used to it it isn’t too bad.

Picard is a mass-tagger, so drag a whole stack of music files onto it and watch it do its work. It’ll try and group your music into albums. To correct allocations, drag their entries to arrange them in the way you please (or drag them away if nothing is suitable). Depending on how esoteric your music tastes are, you should find that most tracks are handled fairly accurately. If you sign up for a MusicBrainz account, you can submit your changes for others to benefit.

Addendum: If you’re using Ubuntu, don’t forget to install libtunepimp5-mp3 for MP3 support.

LotD: Excellent speech by Nicholas Negroponte on One Laptop Per Child. I would especially recommend that the naysayers listen to it.

George Orwell’s classic allegory, Animal Farm, presents many perspectives on human behaviour and society. One of these is how people can be led and manipulated through the control of information. In the story, the Seven Commandments formed a de facto constitution for the Animalistic society. Since only a handful of animals could read, the rest were dependent upon what they were told was written. Gradually, the writing was cunningly altered to the benefit of the pigs above all other animals, and the populace was taught to not trust their recollections of what was written in the past.

What made this subversion possible was the inability of most animals to read. The two animals that could read (aside from the pigs) chose not to do anything about what they saw. Amongst other things, the right to access and read information is an important cornerstone of democracy.

This is where open file formats come in. As our lives become increasingly defined by electronic records, there needs to be a way for independent viewing and auditing. Paper is easily read, but computer files require software to decypher them. Imagine if you needed special (and expensive) glasses just to read the letter that you yourself wrote only a few years ago.

There has been a fair amount of discussion in the press regarding the OpenDocument and the so-called ‘Open’ XML formats. The primary focus of this reporting thus far has been on the political and technical facets. This is slowly changing, as the importance of long-term data preservation and freedom of information become apparent to ordinary folk.

The BBC has published a report on the problem, and discusses how the UK National Archives are attempting to deal with it. Alas, it appears that they have opted for a short-sighted approach, relying on virtualisation of older operating systems and applications, through a direct partnership with Microsoft. With this approach, the format decoders/viewers (not to mention the operating system and software performing the virtualisation itself) remain closed in source and specification, and one must deal with a cumbersome virtual machine just to view a document.

Where is the guarantee that files can be read hundreds of years from now, just as we can do today with paper documents such as the historic Magna Carta? How does this partnership benefit me, an ordinary citizen who might wish to view ten- (or even two-) year-old public documents that are only available in a proprietary electronic format?

It’s both sad and frustrating to see that history is yet again repeating itself. Whilst the contents of the Domesday Book can still be read nearly 1000 years after completion, the digital BBC Domesday Project was rendered virtually unreadable a mere 16 years later.

Thankfully, there are efforts to create an infrastructure for long-term preservation and management of digital documents. To start with, there are open formats such as OpenDocument and PDF. The Australian National Archives have long been supporters of OpenDocument, to the extent that they are standardising upon it. Putting their money where their mouths are, they are building a completely open source (GPL, no less) data managment system that anybody can use or improve to suit their needs. Michael Carden gave a great talk [Ogg video] at this year’s linux.conf.au about this technology, known as Xena [PDF]. Whilst their UK counterparts seem to have forgotten that access to data is not just a privilege for those able to make exclusive agreements with purveyors of lock-in technologies, the Australian National Archives have been striving to ensure that nobody is left out of the digital revolution.

Four legs good, two legs… better? Let’s prevent this subversion from happening.

LotD:  Mexican ‘world’s richest person’

Are you receiving syntax errors when running a shell script on Ubuntu 6.10 (Edgy Eft)? If so, you might want to check what shell you are using. While testing scripts for the LCA A/V Team a few months ago, I discovered that the Ubuntu developers decided to symlink /bin/sh to dash, for faster and less memory-intensive script execution. For POSIX-compliant scripts, this isn’t a problem at all. However, there are many third-party scripts which call /bin/sh but use shell-specific (typically bash) syntax. They can be fixed by altering their first line to call the most appropriate shell, for example:

#!/usr/bin/env bash

env makes it possible to call bash, wherever it may lie. While my Ubuntu and Gentoo systems have a /bin/bash, there are other distributions which have /usr/bin/bash or /usr/local/bin.bash.

While you can and should fix your own scripts to operate in this way, constantly mending others’ mistakes can become tiresome. You can return your /bin/sh to point to bash with the following command:

$ sudo dpkg-reconfigure dash

When asked if you wish to install dash as /bin/sh, tell it to go to hell

Note that bash does use more memory, but on a modern desktop machine the difference is negligible. This change will not affect the default login shell, since that is already bash.

LotD:  Obese Aussies get big ambulances

There has been a lot of debate lately about the issue of binary drivers in Linux. Most of these discussions centre around the nVidia graphics driver. What many of these debates under-emphasise or even overlook is the importance of the Nouveau project. Nouveau is an effort intended to go beyond the 2D capabilities exposed by the nv driver, by providing full support for 3D. The end result will be a 100% free driver on Linux-based systems for nVidia graphics hardware. This is all without the assistance of nVidia, so much work needs to be done to interpret how the hardware behaves and write the required supporting code.

To support these valiant efforts, a pledge has been established to raise at least $10000 USD towards Nouveau’s development. This target has been successfully met with a month to spare, but don’t let that stop you from contributing. Even if you would like to see nVidia’s driver included in GNU/Linux distributions, I’m sure that most of you will agree that it is preferable for us as a community to have our own free driver. All it costs to play a part is a meagre $10 USD. How much is your freedom worth?

LotD: LCA2007 presenters should consider incorporating Gems into their presentations.