Tag Archives: Moblin

The MeeGo Comeths

Posted on by

A few weeks ago, Nokia and Intel announced a merger of the Maemo and Moblin projects. The resulting platform shocked the world. OK, it only shocked the communities (the third amigo) involved. Yours truly wanted to blog about it badly weeks ago. But I apparently never got the knack of timing and priorities just right. So please forgive my late entry into the fray, and accept this out-of-sync piece of journalism.

Sidenotes for the Technically Curious

I will touch on a few things that uninitiated might find confusing.  Without getting into the nitty-gritty technical details there are a few things some readers should know.  A stack or platform refers to a collection of software that works together to perform some larger task.  In the context of this article, a stack includes the operating system, libraries for supporting applications and system programs that provide services for user applications.  User applications being any program a user would use and identify on a device.  UI refers to the user interface for a program, namely all the buttons, sliders and widgets a user can interact with.  Packages refer to installable applications in the Linux world. A library is just a bit of code shared by different applications.

Aside from that glossary, please keep in mind that due to physical constraints small mobile devices will always be underpowered compared to full-scale computers.  There is just so much processor power and memory that can be crammed into a small handheld device before the device will use so much power that it will drain a conventional battery in minutes.  Batteries can only hold so much power without being physically bigger and heavier.  Thats why it is essential for anyone programming something like a cellphone to write code that is fast, efficient and doesn’t use too much memory.  On that note, programmers have the option of writing native or interpreted applications.  Native applications essentially instruct a device using machine code.  Interpreted applications need to first run on a special interpreter program that transform the code into machine code.  For that reason alone, the same code/algorithm running on native application will always be faster than an interpreted application.

The Birth of MeeGo

The MeeGo project comes from two very different but equally important platforms: Maemo and Moblin.  Intel created Moblin as a Linux stack optimized for netbooks.  Nokia created Maemo as a Linux stack optimized for pocket computers/handheld Internet tablets.  Aside from using Linux, a conventional GCC and Gtk technologies,  the Moblin and Maemo platforms are radically different.  Moblin targets Intel Atom (x86) powered netbooks.  Maemo runs on ARM powered mobile devices.  Moblin uses RPM package management.  Maemo stems from Debian and uses DEBs.  The UI is radically different for both platforms.  Moblin exists as a developer preview and doesn’t ship on any production run of netbooks (as far as I am aware).  Maemo runs on Nokia’s N700, N800, N810 Internet tablets and the N900 smartphone.  Oddly enough Nokia still considers Maemo a “developer” platform.  Or rather considered it a developing platform, up until the announcement of MeeGo.  Now Maemo is effectively a dead platform.

The Reaction

Quite naturally both the Maemo and Moblin communities were shocked at the announcement.  Considering how MeeGo is structured, the Moblin community just reacted with a surprised “Oh, thats nice.”.  The Maemo crowd didn’t take the news as well.  No wonder considering that many early adopters and developer paid a lot for the expensive must-have toy, the N900.  Mine own cost around 550 USD, but came out to over 600 USD after taxes, parcel forwarding to Canada (damn you CRTC, Bell , and Rogers for having a bizarre GSM network) and customs.  Many people were looking forward to a bright future with Maemo6.  Instead lots felt that Nokia pulled the rug from underneath their feet.  Needless to say a lot of dust and fuss got kicked up.

We Go into the Future

As soon as the announcement came out, the Nokians found themselves with a farther despondent and worried Maemo community.  The Maemo community felt that their expensive toys would  eventually become expensive paperweights.  However Nokia announced that the new MeeGo platform will work on Maemo devices, especially the N900.  No promises that MeeGo will get optimized to work smoothly on N900s, but a working platform is better than no platform.  So with the community calmed down but cautious, what are the next steps for the MeeGo project?  Well most of the Maemo community web resources are transitioning to MeeGo.  Maemo developers and packagers will need to learn how to build RPMs instead of DEBs for MeeGo.  The MeeGo folks chose to go with Moblin developer tools, simply because the tools are far more mature than Maemo’s.  Fortunately the majority of the Maemo community applications are open source, so in theory simply recompiling and repackaging the applications is all that is required to get Maemo apps running in MeeGo.  Also the MeeGo platform will use the Qt library heavily.  Since Qt is one of the most cross-platform toolkits/libraries out there, developers will benefit from learning a single library that will run on nearly ALL operating systems.

As a Non-Maemo/Non-Moblin User Why Should I Care?

Now most people who use netbooks or smartphones do not use either Maemo or Moblin.  No surprise since both platforms are currently targeted toward developers and early adopters.  Most netbook users use either Windows 7 or a Linux variant.  While most smartphone users use Symbian, Windows Mobile, Apple’s iPhone OS or Android.  Heck even Palm’s WebOS supports more devices than what MeeGo will support from the get-go.  So why should anyone care?  Well in the netbook category, Linux at least is consider a serious alternative to the Windows monopoly.  However the user experience on most Linux netbooks ranks from “decent” to “ho-hum”.  The Moblin (now MeeGo) project worked heavily on making the user experience on netbooks a lot,  a lot nicer.  I would even dare say, a sexy user experience.

On the smartphone side, iPhone is the game in town for good user experience.  Windows Mobile and Symbian are good but still lag behind.  Open sourcing Symbian hopefully will rejuvenate that platform.  Real competition comes in the form of Android (Google Java stack mixed with a rather unorthodox Linux setup) and WebOS.  However WebOS is counting on pure web (HTML & Javascript) applications which are limited in terms of rich user experience.  And Java is yet to be fast on ANY mobile platform.  Building on top of a full Java-Linux stack like Android does, will always be a performance hit.

This is were MeeGo can shine, by providing native Qt applications with rich user experiences.  Nokia and Intel want to work with other device manufacturers to build MeeGo powered solutions.  Solutions such as advanced in-car entertainment systems.  By building on a single platform, there will a lot less repeated work on building the necessary underlying platform.  Instead everyone can focus on building interesting new user experiences, and applications that link with systems in new, unimagined ways.  Of course, only time will tell if any of these blue sky dreams will become reality.  A MeeGo release for currently existing hardware such as netbooks or N900s, will be a definite step in the right direction.

Review of Ontario Linux Fest 2009

Posted on by

2

Two weekends ago I went to Ontario Linux Fest 2009, held here in chilly Toronto for the third year in a row. And for such a young conference, it was quite good.  In fact good enough, that longer review of the event is in order:

Morning Keynote – Changes to the GPL

The day started with me rushing out the door in the morning, to catch a morning bus.  I made a bit later, missing the first part of Bradley Kuhn‘s keynote.   I rushed in, grabbed my conference package, put on my name-tag and rushed to hear the second part of the keynote.  Brad discussed the changes between versions 2 and 3 of the GPL (GNU General Public License).  As someone who followed the licenses and uses the GPL in my current project (justCheckers), it was pretty interesting to hear about why behind the changes.  The original GPL2 was quite brief for a software license, but not quite as understandable as the GPL 3.  The GPL 3 helped simplify the license, made it international and got rid of some icky loopholes.  After the keynote, I did a quick swag run even buying a fleece from the Eclipse guys from Redhat.  (And no Nick, I wasn’t there just for the swag I just haven’t gotten around to contributing to Eclipse just yet.)  I then quickly popped quickly back in for the first track of sessions.

Session 1 – Enterprise Content Management (ECM)

Cheryl McKinnon from Nuxeo did a talk on ECM.  I can definitely relate to having to hunt down information in the mess of Office documents and e-mails at the office.  I can just imagine what happens at larger organizations that have less rigorous guidelines to handling documentation.  So I definitely see the need software that handles such data, especially in terms of productivity and maintaining documents for regulatory purposes.  I can’t understand why every organization doesn’t have at least a wiki.  Interesting note that the concepts I learned in a university course on information science, really applies here.  Managing the capture, collaboration, review, publish, archiving and search-ability of information is really a science on to itself.  No wonder open source vendors like Nuxeo, Alfresco and MindTouch that provide ECM solutions in a flexible and low-cost way are such a big hit for enterprises.  There also was a neat discussion on how Nuxeo can do both centralization (bring documents into itself) and management of meta-data (like the locations and “tags” related to scattered documentation).  Really neat.

Session 2 – Ubuntu Moblin & Netbook Remixes

Jorge Castro from Canonical of Ubuntu community fame presented the Ubuntu Moblin and Netbook Remixes.  As with any live demo, the demo technology co-operated with Jorge like any demoed technology or self-conscious prima donna.  Still the Ubuntu Moblin Remix looks gorgeous.  It sports a nice, simple, elegant and understandable user interface.  However this remix qualifies as a tech preview more than something to hit the mass consumer.  Still I think people will be impressed when they change over from Windows to something like Moblin.  It looks a far bit of effort was put into making refactoring the UIs to fit a smaller, wider screen.  The Netbook Remix looks great too, and if you are reading this and running Ubuntu 9.10… switch over to the Netbook UI, you’ll love it even on your main system.

Lunch Chat 1 – Free Software and Its Impact on the Future of the Software Industry

I skipped out on the last session of the morning to pester Brad Kuhn with a few questions.  It started with an innocent question about his opinion about certain companies planning on moving from the GPL to the Apache licenses.  Brad figured out that I had read one of Matt Asay’s blogs on the topic, and went on a mini-rant.  Apparently Matt is one of Brad’s “nemesis” (not that they don’t get along rather their views on software freedom are quite different).  Brad envisions a future where software stops being a “big box” industry of packaged software products to more of a lawyery/consultancy profession.  I have to agree with him on that, and I don’t mind such a future being a consultant by trade.  However I did point out the difficulty one runs into with consumer-level products or projects.  Basically how does one sufficiently fund a project like Inkscape?  One way would be to sell a proprietary product, which defeats the purpose of going open source.  Another way would be to offer a service.  The heart of the problem is how to finance a regular development and enhancements (like usability) on a regular basis… and still get someone to pay for it willingly.  It isn’t an easy problem to solve… A problem that I plan on working on in my semi-stealth project… But Brad tried to convince it wasn’t as big of a deal as I think it is.

Lunch Chat 2 – Decoupling User Interfaces from the Application Backends

Feeling now famished for lack of a breakfast, I went out with Scott from the GTALug for pizza.  Scott is into user interface design both on a hardware and software level.  We both agreed that user interfaces should be loosely coupled to their backends.  In enterprise web application design the concept of different views for different users and environments comes to mind.  Scott introduced me to the concept of a framework that allows for completely decoupling the user interface with the underlying application, called Metisse.  It allows for building UIs at runtime using a widget palette or toolkit… brilliant!  That would let developers concentrate on what they are good at: application development and design.  And HCI/usability/graphics designer concentrate on what they are good at: design kick-ass usable user interfaces.

Session 3 – Building Business Applications Using SugarCRM

Right after lunch, I went to a more practical session by John Mertic, a developer at SugarCRM.  The session was naturally about developing applications using SugarCRM.  It looks a very nice application, and CRMs apparently can solve a lot of problems centred around customers.  Even just using SugarCRM to manage a client’s organizational contacts would save a lot of my time at work.  However since I work in Java and not PHP, thats not really an option.  But it looks like a neat application, the latest version SugarCRM 5 being all built in PHP and using the YUI (used to use ExtJS).  There is a vibrant community forge and marketplace around the application.  Neat.  Also neat that John just published an Apress book called: The Definitive Guide to SugarCRM.  See a pattern, yet? 😉

Session 4 – World Domination, Documentation and Ponies

OK, I admit it as a writer masquerading as a software developer,  (Or is it the other way around?) I actually enjoy writing documentation.  Some would even say, I enjoy writing documentation too much.  So I jumped at the opportunity of hear Emma Jane Hogbin talk about world domination through good documentation.  Actually I was disappointed by the lack of talk on world domination.  But there was much talk of ponies… oh and documentation.  I swear I’ve never heard anyone comparing documentation to wearing high heels.  I guess the analogy of it being sexy and painful makes sense.  But then again I’ve never (or do I ever plan to) worn high heels, so I’ll trust the ladies on this one.  Still it was an awesome and enlightening talk.  I’ve never consider all the various audiences (devs, users, marketing) and types of documentation.  I must agree with Emma’s statement, that if you are doing a lot of work documenting how to use your program… maybe reconsidering how the UI is done is in order.  The concept of automating documentation and using an XML source is new to me.  But I will have to look into technologies like pywebdoc, Mallard and DocBook.  Emma went through the Capture, Organize, Translate, Output, Review and Revise cycle, using the Status.net documentation efforts.  (Emma’s project on writing open source documentation.)

Session 5 – The Look at Introducing FLOSS into Education

Karlie Robinson of On-Disk did a talk about her experiences with connecting the OLPC, Fedora and RIT together.

Afternoon Keynote – How Linux is Like Music

The conference wrapped up with Joe ‘Zonker’ Brockmeier‘s afternoon keynote.  Zonker is an incredible speaker, and he also threw plush penguins at the crowd which instantly made the keynote awesome.  Alas I didn’t have any thing worthy to add to the conversation so I didn’t get another plush penguin. 🙁  But I’m getting ahead of myself…  Zonker talked about how we should move away from the tired analogy of choosing operating systems like cars.  Rather he suggested how choice of OSes and Linuxes should be compared to bands and musicians:

  • Fedora -> Frank Zappa
  • SuSE -> The Who
  • Ubuntu -> Duran Duran (Hurrah, I mean huh?!?)
  • Gentoo -> Write your own sungs in your own built studio (Damn right!  Sorry, ecstatic ex-Gentooer here.)

A bit zany, but with a nickname of Zonker… 😛 Ok, I’ll stop teasing.  But he did go into the problem of widespread Linux adoption: better self-marketing, pre-installed PCs (we are getting there slowly), market shift to the cloud and mobile (at least we have that one covered to a degree) and  the in-fighting that the F/OSS community seem to love (welcome to the new face of anarchy and things to come?).

The Pre-After Party

At this point most of the participants went out to either go home or wait out the time for the reception.  Having come alone, I felt like I’d love to mingle at the afterparty but I needed to do something while I waited.  Fortunately, the KDE 4 guys (Troy Unrau, Shawn Starr and Eugene Trounev) were quite welcoming and let me tag along with them.  Troy was pulled into the speaker’s dinner so I didn’t get to talk to him much.  But Shawn, Eugene and I chatted for a bit, grabbed a quick bite to eat at Burger King and came back to chill at the venue.  Shawn works on developing plasma, while Eugene makes graphics for KDE Games.  Actually if you count myself and Troy in, we each represent a different aspect of the KDE  project: Shawn – developers, Eugene – artists, Troy – marketing and myself – the humble user/wannabe contributer.

When Eugene found out about my semi-stealth project, he suggested I should join the KDE Games and look into the Gluon project.  I’m planning on getting involved, once I’ve hacked around with Qt, something I started on recently.

The After Party

The conference wrapped up with a reception sponsored by Google.  Actually I got a good amount of useful swag from Google, and the free booze was much appreciated.  At the reception I got to mingle with the presenters, organizers and participants of the conference.  I got to discuss the finer points of revision control systems (CVS, Subversion and Bazaar) with a Bazaar/Launchpad developer, Aaron Bentley from Canonical.  Also got to watch Brad Kuhn and Jorge Castro get into a heated but amicable debate about Mono, software patents and why Canonical should stop releasing and supporting proprietary software.  And I got to chill and talk about Apache Qpid (a reliable message queueing system, think something like IBM’s MQ) with Redhat’s Rajith Attapattu.

I left the reception really, really late.  And I got home around 2 in the morning, which just goes to show how awesome of a conference OLF2009 was.  After the conference I felt a lot smarter, enlightened and inspired to contribute back to the Linux and libre software community.  I’m almost sure I’ll be at OLF2010 whenever that comes around.

Links to what others had to say about OLF2009:

  • Troy Unrau [http://troy-at-kde.livejournal.com/23041.html]
  • Eugene Trounev [http://my.opera.com/it-s/blog/show.dml/4455292]