fosdem.org

Michael Meeks will give a talk about "LibreOffice: on-line and in your pocket" at FOSDEM 2012.

Could you briefly introduce yourself?

I'm Michael Meeks: Christian, Hacker, Husband, Irritation, and more. I work for SUSE who pay me to make Free Software on the desktop even better for our customers. At the moment, that means working on LibreOffice most of my time, though I've enjoyed working with some great hackers on other projects over the years, mainly GNOME and openSUSE.

What will your talk be about, exactly?

LibreOffice - and how we're developing to simultaneously address the two popular, competing visions of the world of today: 'Apps' that run on your phone/tablet in your oversize pocket, and 'Web Apps' that run on part of someone else's computer. Tackling just one of these is hard, to try to do both at once is perhaps crazy. Having said that, we take an interesting approach to provide both of these re-using the same code-base, please do come and hear about it.

Beyond that of course, I'd love to sneak some of the great things we've done in the last year into the talk, some of the new features in 3.5 and some graphs of our trajectory - the usual stuff.

I'm also planning to speak on copyright assignment to corporations in the Legal Issues DevRoom (which looks like an exciting addition this year), and to encourage people to get involved with some Easy Hacks in the LibreOffice DevRoom.

What do you hope to accomplish by giving this talk? What do you expect?

My primary goal is to encourage more great Free Software hackers to join in with us. By some lucky co-incidence I expect large numbers of these to be at FOSDEM. I'm hoping that people will be inspired to help us out, and make a significant blow for software freedom on the desktop by doing some small (or even big) improvements to LibreOffice. With such a huge code-base and feature-set, we are permanantly in an "every little helps" mode - so we like to encourage even those who only think they have a little to offer to get stuck in. Everyone no matter how unskilled today can do something useful and grow in doing so.

In your effort to offer alternative (mobile and web) user interfaces, I guess you have to decouple the current 'fat client' user interface from the underlying code. Is this a big hurdle or is it not so bad?

Great question. Luckily LibreOffice has been a cross-platform application for some decades, so there is already a reasonably tangled cross-platform abstraction there. For example on Linux we can create binaries that target all of gtk2, gtk3, kde3, kde4 or raw X, choosing at run-time which to use. Adding implementations of that interface for Android and the web is a fun task of course, but much of the underlying infrastructure is in place. The web front-end leverages our port gtk3 support, and the 'broadway' backend that allows all (well behaved) gtk3 applications to render into an HTML5 canvas inside your browser. The most painful piece of my port of LibreOffice to gtk3 to enable this was that 'well behaved' detail. The legacy rendering support we inherited is/was a quick and dirty graunching of the raw X rendering with some gtk fluff, it is rather cleaner now.

In what ways do these efforts to offer alternative user interfaces help the normal fat client office experience?

This is really the joy of sharing 95%+ of the code between platforms, much of the performance optimisation necessary to make LibreOffice work well on the web for multiple concurrent clients will help accelerate the single user case too. Similarly, optimisations for handheld devices to reduce memory footprint should benefit all users. LibreOffice's user interface has long been in need of some radical improvement and I hope that wider interest and adapting to the needs of smaller, touch enabled screens will have knock-on ease of use benefits for the desktop over time.

When will LibreOffice Online or the mobile app be ready for general use?

That is really the hundred million dollar question. I suspect it depends on what exactly you want. I'm optimistic that we can get something useful on-line for the end of 2012 or early 2013, but that of course depends on community involvement and help. We will be shipping our gtk3 support in 'experimental' mode for people to try out for 3.5.0 which should be out just after FOSDEM. The Android port, I suspect, will start with a high-fidelity document viewer, with that evolving over time to allow editing. Of course, there is some unpleasant proprietary Apple tablet and phone operating system that we may want to re-target some similar work towards too.

As always the punch line is that everything goes quicker if more people help out, so that we can bring these expertly guesstimated dates a lot nearer; tell your friends.

Last year Jos van den Oever talked at FOSDEM about the WebODF project which lets you view and edit ODF documents in a web browser. How does LibreOffice Online compare with WebODF?

I love Jos's enthusiasm and initiative. My personal feeling is that doing the last 30% of a WYSIWYG web office suite takes 95% of the time - you see, the first bits (that the browser supports natively) are really rather easy. After that, you get to write a huge mound of JavaScript, which you distribute and run under a JIT in several not-very-compatible browsers. Of course, we can always abandon WYSIWYG, perhaps printing onto pages is really truly dead, but I'm not so sure that crisp, repeatable paginated layout is really truly gone for good yet. That's particularly true for drawings, posters, marketing etc.

Also there are really two historically large companies whose experience are quite relevant on this. One was Corel - whose "re-write our entire office suite in Java" strategy seemed to bomb vigorously, though the prototypes running in the browser were quite an impressive Java technology demonstration for 1997. So, I'd prefer not to have to bet the farm on re-writing eight million lines of C/C++/Java/Python into several more million lines of (type unsafe) Javascript.

The second was Micropro; once the biggest software company in the US, whose 'WordStar' ruled the word processor market in the early 1980s. Unfortunately they chose to do a parallel, from-scratch re-write of their product (to compete with IBM's DisplayWrite), using a different technology and then aggressively competing with themselves while loosing in the marketplace. There are clearly some real risks and temptations here.

So, I'd really prefer to focus on a single product, with substantial source-code re-use for different segments. Perhaps that's a result of lots of SUSE presentations on common code-bases. I try to persuade others on this topic too, so far seemingly successfully.

From the beginning, you wanted to lower the barrier for new contributors to LibreOffice, e.g. by offering 'easy hacks' on the website. Has this been a successful initiative?

That is actually quite hard to measure, it is hard to disentangle word-of-mouth advocacy: telling people that LibreOffice is a fun place to hack, from other factors like Easy Hacks, the superabundance of low hanging fruit, the burgeoning user-base, the copy-left license, not Reply-To: mangling on developer mailing lists and any number of other topics that might be relevant. I did try to categorise responses by keywords found on the mailing list like 'easy hack', and I managed to produce some rather easy to mis-interpret numbers. Bjoern Michaelsen did some great work migrating the easy hacks to bugzilla, so hopefully we can start to graph the rate at which they are closed over time. Certainly, we have many people in our community today, whose first contribution was an easy hack - hopefully that helps people to push on the door, find it extremely open, and thus to fall into the project. If you're in doubt - try one and see how it tastes here.

Have you enjoyed previous FOSDEM editions?

Emphatically yes ! FOSDEM is an incredible conference - and better an important meeting place for the best European Free Software developers. If I could go to only one European conference per year it would be FOSDEM.

Every year, I enjoy meeting old friends, make new ones, and talk until extremely late into the night / morning. I find that meeting people in person, and understanding the rich texture of their character, passion and friendship helps me to understand their E-mail more helpfully later. Free Software is built out of people, and getting to know them better is crucial. FOSDEM is also a great place for people to get involved for the first time with Free Software, I'd love newbies to come and meet up with the teams I care about most: LibreOffice, openSUSE, GNOME, reprap, whatever takes your fancy - please come, learn, and find your place to be most effective, as we change the world together.

This interview is licensed under a Creative Commons Attribution 2.0 Belgium License.