Solaris 10 is out today, and I am exceedingly happy to be working for Sun Microsystems. In case this did not come out in my previous post, I am being paid to work on the open source BlogEd. I don't think there are any companies out there as enlightened as Sun :-)

Open Source is sustainable the same way capitalism is: because a vast number of players find some (larger or smaller) interest in it.

I will speak from my experience, as that is what I can speak of with most confidence. I started using Open Source software because I like to use software that I know will be mine to work on all my life. I don't want my knowledge to end up in some proprietary black hole, where one day all the knowledge I have accumulated is suddenly closed to me because I run out of money, or the firm that invested in it initially goes bankrupt or sells out to some predatory competitor.

I also used Open Source software because it is a great learning experience. What better way to learn than to see how others, sometimes the best minds in the field, have build their work. Why do the best minds spend their time this way? Each has their own reason of course, but a very important one certainly lies in the paragraph above: because they too don't want their creations to be lost in some black hole never to be found again. Like the builders of cathedrals, of the greatest stories such as the Illiad or the Odysey, they want their work to last forever.

As a result I also use Open Source software because it really works. I had been using Linux since 1997 daily, and for the last year have been using Apple's OSX, which is build upon a very solid Open Source foundation.

Open source software gave me so much, that I recently wanted to contribute back to the community. As I have narrated in this blog, I started helping out a little with the Apache foundation, and since I wanted to learn more about blogging and wikis, partly because I was trying to help my brothers with their web site, I started getting involved with BlogEd. This turned out to be interesting for me in many ways:

• not only was I helping out my brothers (not very successfully in the end it turned out -- but nothing is final there),
• but it was also a great opportunity to learn about the Semantic Web: RDF, OWL, N3, and friends
• to learn XML Remote Procedure Calls
• meet a lot of interesting people (who would have had nothing to tell me had I not been around to help out)
• learn xhtml, css, XQuery, and other xml tools in greater detail
• improve my web site with this blog and the techniques learned above (just compare my site now with how it looked)
• And mostly good networking: what better way to find an interesting job, than to show openly how you can do something you are good at, and help others out with their problems.

  How good is networking in such a way in the open? Well pretty DAMN good! Instead of spending all my time writing CVs in MS Word format for head hunters that search for keywords in resumes without knowing their meaning, who then give you a job that has absolutely nothing to do with your skills, working often for some company where you end up being a pawn in some game nobody understands; instead of that I ended up with a job at SUN Microsystems, being paid for the Open Source work I was doing for free.

  So now why would SUN Microsystems pay me for doing work that they know I love to do? You may want to ask them. They have their own reasons like I have mine. Perhaps they reckon that someone who does work they like, under a licence that gives them a long term interest in the work they produce (such as the BSD licence I am working with in BlogEd) means that they are more likely to produce work that will try to make the right long term decisions. Perhaps it also save them on management costs. I don't know. Everyone has their reasons.

  There is no one reason why Open Source works. That is why it is sustainable. Sustainable the way religions are. IE: really long term.

Electronic voting without verifiability should not be allowed under any circumstances.

What the 2000 US elections showed was just how important verifiability is, even though it also showed the limits of simple paper voting: it was never designed for political machines that were so well oiled that they could create elections that are a few hundred or thousand votes close, where any mistake is close enough to making the difference in the campaign.

But is electronic voting completely impossible? I don't know, but I think it would need the following properties to be an improvement on the current system:

• Like the current system it should be anonymous
• It should be completely open: every voter should be able to verify that his vote is correctly recorded, and that the sum total is correct

So how could something like this work?

My first thought was something along these lines:

1. The voter goes to the polling office, votes, and receives a paper on which a digital signature is placed that uniquely identifies the paper he has in his hand and the way he voted. The vote cannot be altered without invalidating the signature.
2. When everybody has voted, theses signatures, with the corresponding votes, are all placed in one large tarfile and placed online.
3. Everyone can then download the tarfile, or copy their friends tar file (it could be placed on a p2p network) and
   • verify that the string on their paper with their vote is in the tar bundle
   • verify that all the votes add up the way the official count says it does.
4. In case of problem they can use their paper to proove that their vote was not counted or badly counted.

This would allow everyone to verify that their vote was counted, yet nobody would know how anyone else voted (without having access to their balot).

This last sentence reveals the problem with this system: it would encourage the buying of votes, as anyone could use their ticket to proove how they voted. Even though the current US system allows the buying of votes to take place, since it allows voting by mail, this should not be encouraged. We should add the following restriction on a good voting system:

• It should not enable the buying of votes

I wonder if it is technically feasible to devise a system which allows fully open vote verification, without also allowing vote buying...

Something to read up on. I found some interesting looking links.

The Financial Times recently published an article by Professor Richard Epstein Why Open Source is Unsustainable that is a nice re working of much of the FUD spread about open source.

This article is debunked:

• In the response by Professor James Boyle, following the article.
• In a more detailed legal analysis published on Groklaw: The Sustainable GPL.

This is really quite amazing. The phone and the camera are merging completely. This phone has a 5MP camera for making films and videos. The backside of the phone is visible here:

This is the phone to take on a bicycling tour!

But it is oddly enough, like so many other phones, completely missing out on the video conferencing market. All that is needed is to redesign the phone with a swivel screen so that you can take a picture of yourself while you are seeing a picture of the person you are speaking to on the screen. I would also love a phone like this to plug into my 17" Apple PowerBook so that I can use the camera in the phone for my iChat sessions. One thing less to carry around.

It would of course also be great if it had built in GPS, so that one could annotate one's photos with RDF. And of course on a bicycling tour GPS combined with mapping functionality would be indispensable.

So close to perfection!

It is been about time to add a new post. And what better way than to refer to über-blogger Danny Ayers, himself. :-)

Recently I have been wondering what the atom format gives us that RSS1.0 does not, or could not with a simple extension. (note: that e-mail was written a much too quickly, and would have done with a lot more careful crafting) I don't claim to have come to any conclusion yet, but it does seem to me to be an important question to ask when one develops something, anything new: how does it improve on what came before? And could one not get the same result simply by extending the old in some simple way? After all is that not exactly how even a widely understood to be inferior technology such as the intel processor has managed to capture such market share?

One area where clearly Atom does have something interesting to bring to the table is the RESTfull protocol. But the conversation on that mailing list does not seem to be advancing very fast. And I also wonder if one could not do RSS1.0 with that protocol...

There may be good answers to these questions, and Danny Ayers has a few himself. It is clear that there has to be a very big improvement for all the work to be justified. This should stimulate people to perhaps take some more original and far reaching ideas seriously, I hope.

I've had a little flue this week, perhaps due to too many late nights in Galway.

Nevermind. It has allowed me to catch up a little on some of the many interesting talks that I listened to there, search through the numerous foaf files with swi-prolog's RDF libraries, build a list of people I had met there...
I really look forward to the release of the foaf viewer's on show there, and to a the excellent graphical RDF browser being worked on by Benja Fallenstein.

I just received and immediately devoured Ben Hammersley's excellent "Content Syndication with RSS" [1]. I wish I had bought that before. It is a really good overview of the history of the RSS formats, and in particular is a great introduction to RSS1.0, which is a really powerful format.

I am going to have to study that carefully in order to work out how Atom is hoping to improve over RSS1.0.

[1] ISBN 0-596-00383-8, O'Reilly