We have had the chance to do an e-terview with Michael Meeks who is a leading OO.o developer and currently working with Novell. Read more as we ask him ten questions about Microsoft, The OLPC, Religion and Wikipedia.
Q. Michael Meeks, your name are associated with Gnome, OpenOffice, Ximian and Novell and you are also a seasoned hacker and code debugger. That is in addition to the role of fathering a six-month old baby. How do you exactly manage your time?
I'm not sure I manage my time that well; sifting the important from the urgent and doing the right thing takes half of each day, then sometimes I have a chance to hack, which is what I love.
Thankfully soon I hope to hand some of the management responsibility for Novell's large, and growing OpenOffice team over to a friend, and do more hacking again.
My Gnome contribution these days is mostly doing small maintenance tweaks to ORBit2 or bonobo code. In general for productivity though I've found reading mail just once per day, and planning your time in advance, are the keys here.
You are one of the lead OOo developer and you work at Novell, shouldn't you be working at Sun instead?
Well, Novell is working hard to try to grow community involvement in the OO.o project, reduce barriers to entry and helping to nurture external hackers. This takes a lot of effort and buckets of patience. Hopefully we're getting to a world where it doesn't matter where you work, or for whom, you can enjoy hacking on OO.o.
Having come from a somewhat anarchic Free software background, it's interesting to see the cultural differences in a big organisation, and to encourage change from outside; still a lot to learn.
Of course, you have to give tribute to the huge engineering effort Sun puts into OO.o, and the great result that achieves, and understand that they have compromised a lot in adapting their ways of working to accomodate outsiders.
One great recent example is the switch to three monthly time-based releases; this enables us to deliver incrementally higher quality versions every three months, including new (well tested) features. I think it's been successful in encouraging community developers, by letting them see their work get used by people.
The price delta between OOo and MS Office is three to five times that of Linux and Windows, so what aren't more people (and businesses) switching to OOo?
A very good question - we should ask them. Businesses of course want support, and Novell, Sun and others can offer that on both Windows and Linux. For individuals it's really a matter of marketing, many people I meet casually simply don't know about OO.o. Finally of course for some small minority of users there is a feature/function gap that we are working on closing.
Novell's work on Calc here is very helpful, and improving interoperability with Pivot tables, and VBA macros gives substantial improvements.
As a dedicated Christian, how does your faith influence your coding and your working life?
Let me tell you I constantly fail... constantly. Some people think that Christians think they are "good people", but this can't be further from the truth; it is (clearly) not the well that need a doctor but the sin sick. So - basically I'm not much good as a Christian, but Jesus Christ loved me enough to die for me, so I try to be faithful too.
Sometimes, I think the theology of vocation gets mislaid - it's a really good thing to use your gifts to serve Him doing what you love. There is no need to wander round speaking sonorously, wearing silly robes.
Your Wikipedia entry says that you have "contributed a lot of time to decreasing program load time". Do you have any metrics on that and what applications are actually concerned?
I've no idea who wrote that; seems a little performance focused. My interest in performance of course comes from OpenOffice which is the largest single Free software C++ application available. As such we uncover a lot of problems across the system.
Initially I focused on warm start time, where performance was dominated by ELF linking; over 50% of OO.o warm start (on a modern machine) could be attributed simply to run-time dynamic linking.
This resulted in a lot of research and several very promising potential optimisations, some of which have found their way into the platform in
recent binutils / glibc version, others of which need more work, and others which were rejected out of hand. Of course, Amdahl tells us that we can get no more than 50% faster by optimising linking.
Other recent work has been trying to encourage the kernel people to sort out the I/O story on Linux, I/O is about 80% of the cold start time, which
looks like a more fruitful area of work. Clearly there is still lots of exciting and fruitful work to be done with improving performance.
What are your thoughts on the One Laptop Per Child project? Is there any particular reason why Novell did not take part in it?
No thoughts in particular. I know some of the RedHat people working on it and they are sharp cookies. I'm of course interested in getting OpenOffice.org slimmed down so it'll fit onto it (and your mobile phone), but perhaps that's a pipe-dream.
As for the public policy issues of where to spend money in the third world: mosquito nets vs. education is absolutely inscrutable from my
perspective. Overall, though it's clearly a great win for Linux and the Free software desktop.
Let's talk about Microsoft. If you were to choose one project at Microsoft to open source, which one would you choose and why?
Wow - what a question. Firstly, I'd want them to do more than open-source their software, and make it available under a true Free software style license. As to what project, hard to say - perhaps Office, Access, Visio or Exchange - really it's a long time since I used Windows desktop apps in anger for anything but interoperability work.
You talked about the Linux Desktop at LinuxWorld 2006. How well did Linux fared on the desktop in 2006 and what are your predictions for 2007?
Predictions for 2007 - steady growth, increasing numbers of compelling business deployment case studies, lots more media interest as Vista is launched, continual growth in capabilities and innovation.
There is so much that starts to become possible when we have a capable platform to innovate on, without artificial barriers - I'm excited by the future for the Linux Desktop, ultimately the Novell message: 'Linux is Ready' is just true - we have a great product, now we have to encourage people to try it out, see what they can achieve with it, relax, and enjoy the freedom it gives them.
Novell has come a long way since the Netware years. At last count, the company listed 43 OS projects on its OS page. How has OS influenced the business strategy of Novell?
Really, I joined Novell as part of the Ximian acquisition, so I don't have a particularly insightful historical take on this. What encourages me about Novell though is that we can do things that people want that no-one else can do.
If you want to start migrating towards Linux on the desktop inside your organisation, then if you have Novell ZenWorks to manage your Windows systems, it will already manage the Linux systems too. If you use eDirectory or Groupwise, again you get a very smooth transition to Linux.
Novell is unique in its huge experience of managing, securing and providing products for heterogeneous environments.
What open source technology has impressed you the most in the past few months?
I think Xgl/compiz was pretty impressive; one of the fruits of Novell's deep investment in creating new desktop technology. I'm also pretty pleased with Mono - the improvements there are not individually earth-shaking, but the technology is really very, very useful. Overall though perhaps the most impressive is Beagle (written with Mono).
Lots of information pours onto my desktop, and of course I know where most of it is - but for those few percent of cases where I stupidly lose things, the tool is actually invaluable. A simple concept it's true, but vital in a tight corner.
Thanks Michael for your valuable time.
You can read the daily diary of Michael Meeks at http://www.gnome.org/~michael