« May 2003 | Main | July 2003 »
GRIC is an interesting network of thousands of POPs all over the world. In Denmark, only TDC Internet offers access to the GRIC network, and they charge heftily for the privilege. During my recent trip to Paris I used Tempest Telecom. Their site looks shoddy but the service works.
Oracle has offered $6.3b for PeopleSoft. On their website, Oracle presents some pretty persuasive arguments for the merger. Oracle would basically buy PeopleSoft's customers, the maintenance revenue from existing PeopleSoft installations, as well as the opportunity to roll over PeopleSoft customers to Oracle 11.
PeopleSoft shareholders are probably better off with cash now. Even with a JD Edwards merger PeopleSoft will have a hard time competing with SAP or Oracle. Oracle has sued PeopleSoft's management and board
PeopleSoft customers may be getting screwed, but at least Oracle has pledged to extend the lifetime of existing products and to continue to release enhancements. They can continue to buy additional seats but no new licenses will be sold. A change from PeopleSoft to Oracle 11 is probably not much more expensive than the switch from PeopleSoft 7 to PeopleSoft 8.
In their ads in major periodicals they point out that customers from the purchase of Rdb (a SQL database for OpenVMS) from Digital 9 years ago are still happy, and the product is still maintained.
Philip Greenspun takes offense at PeopleSoft's management strong opposition to the offer:
Craig Conway, the CEO of PeopleSoft, however, states that "I could imagine no price nor combination of price and other conditions to recommend accepting the offer to our shareholders". I.e., if Oracle offered to give his shareholders a Dr. Evil-esque $100 billion he would turn it down because the acquisition would mean the end of his CEO job. Mr. Conway earned $17.6 million last year, during which revenue and profit declined and his shareholders got creamed.I agree that it is outrageous. Conway even went as far as to say:
If I said, I'll give you $10 for your dog and then I'll shoot it, what would you do? Would it make any difference if I offered you $20?"(I have this quote from Oracle's complaint against PeopleSoft and its management in Delaware chancery court.) The complaint makes sense. Whether or not an Oracle/PeopleSoft merger would be in the best interests of PeopleSoft shareholders (and to be reasonable that has also been questioned), it is clear that Craig Conway is not.
Update: I forgot to disclose that I am an Oracle shareholder.
Dave:
If you like using your aggregator to read RSS feeds, please find a way of saying that publicly. If you want mature steady leadership for the technology, find a way to say that too. If you don't want the pavement ripped up because a few competitors have fallen behind and want to create confusion until they can catch up, say so.
I use a news aggregator (that I help develop) to read RSS feeds. I want mature, steady leadership for the technology. I don't want the pavement ripped up because a few competitors have fallen behind and want to create confusion until they can catch up.
Dave Winer has shut down Scripting News after the flamewar on RSS, blogging APIs and Echo: The lack of support, even name-calling, from people who think of themselves as my friend, has got me thinking that maybe this isn't worth it. And the abuse from others, bordering on cruelty, is intolerable.
Naturally, the archives are still up so no links will break, but imho the blogging community will lose an important voice. Dave created XML-RPC and co-created SOAP, writing the most sane SOAP subset.
Update: It's up again.
The centralized news aggregator project that I've blogged earlier now has a name—Lucmo. It has a blog where you can read about what lucmos are.
Simon and I going to freeze version 0.2 on Saturday, July 5th, with a release planned the following Monday. Version 0.2 will include very significant changes compared to the current OpenACS release, changes that most news aggregator users will benefit from and which can also be applied to .LRN. Unfortunately version 0.2 will not include upgrade scripts. You can monitor the progress of 0.2 in our public bug tracker.
I just realized that most of the Economist Style Guide is online. Only the sections on American vs. British and the fact checker (with such interesting details as the proper English spellings of all the German bundesländer) are missing. The style guide opens, Clarity of writing usually follows clarity of thought. Wise words.
Users get the best news aggregator in the world, one that does not require users to install annoying software on the desktop (that invariably only works in Windows), makes reading RSS feeds much easier and gives pointers to related weblogs.
The service addresses two specific target groups:
Since My.UserLand was discontinued there has not been a good news aggregator with functionality similar to the best ones on the market (in my opinion Radio UserLand, AmphetaDesk and Python Desktop Server).
Some of the features that we intend to implement, such as collaborative filtering and relevance ordering, do not exist in any other news aggregators, to our knowledge.
Apart from UserLand products (Radio and Manila), there are no news aggregators that allow users to create public aggregators such as blogs.law.harvard.edu/aggregator. Users will also be able to create RSS feeds from their news aggregators.
First of all we provide a great service to the weblog-reading community because the news aggregators out there suck big time.
However, we can potentially make money from the service. It is possible to place Google ads on our service. We also have so many more advanced features lined up in the pipeline that we can create a premium version of the service.
A news aggregator places less load on our servers than services such as webmail or blogger, meaning that our service is cheaper to provide than many other free services.
Will we ultimately make serious money from the service? If we actually fulfill our goal of making the best news aggregator in the world, I think we can make money.
I am not certain that good news aggregator providers should have a ethos as such, apart from providing a fabulous service, but here are some ideas:
Large parts of our software will be open source, at least all of the basic news aggregator package, multiple aggregators, feed caching, etc. It is quite possible that we can make all of our software open source since the value of our service explicitly lies in the fact that you don't have to install anything. The value of collaborative filtering in finding weblogs is in the data that our users create, not the software per se.
Our goal is to make reading weblogs easier and more enjoyable. We should jointly write a weblog that documents the service's development and provides perspective on issues related to weblogs, syndication and news aggregators. Such a weblog would create a personal voice for the service.
We should participate actively on the syndication mailing list and in the development work for RSS.
We want to be what Blogger is to weblogs and what fastmail.fm is to email clients. We absolutely abhor the 3-pane user interface adopted by FeedReader, NewsGator and Oddpost, among others.
We need to determine which of these features are truly useful, and which are
The service would be based on the existing OpenACS news-aggregator package developed by Simon. The vast majority of the new features discussed here would be GPL'ed and hopefully incorporated into OpenACS.
GPL is not a threat to our business model, should we ever succeed in making money from the service, because the very point of this service is to spare users the burden of having to install their own news aggregator on their desktop. The value of collaborative filtering and recommendations for weblogs depends on the number of users on the centralized news aggregator service. Such filtering would be equally useful on OpenACS or .LRN sites that have a community of weblog readers, which is distinct from the "general" weblogging/news community that would use our service.
DaveNet: BBC has released RSS feeds for almost every news index page on their site. This is amazing and should set an example for other major news sites. The BBC is non-commercial and government funded. Their archive is open and free to all who want access, anywhere in the world. Paid for by British taxpayers, it's a dividend to a former colony, for borrowing their language.
Now they only need to resurrect their Ogg Vorbis streams.
News aggregators should themselves have RSS feeds, something that the blogs.law.harvard.edu aggregator does not even have. This would allow users to subscribe to each others' feeds and also public feeds like the Harvard ones. Of course, an intelligent news aggregator would weed out duplicate postings.
It appears that our friend Simon Carstensen was the one who wrote the OpenACS news aggregator. I just discussed some of the ideas from the Lafayette Project (Simon was also at reboot and heard Meg Hourihan's talk), and we agreed to meet and discuss whether it is possible to create a similar service very quickly using OpenACS, Simon's news aggregator and some additional code for RSS feed caching, multiple news aggregators per user and collaborative filtering. Kind of like My.UserLand on steroids.
I have now put Google AdSense ads on my site.
Meg Hourihan's talk at reboot was about the project that she and Nick Denton are doing, The Lafayette Project (which we are lead to believe is just a codename).
My impression is that the First Stage is a web-based news aggregator, with a Radio-style view where all the postings from different weblogs are collected in a single page. AmphetaDesk uses the same layout. This is a Good Thing. The news aggregators out there really suck from an end-user perspective. Many of them use the horrible Outlook-style 3-pane layout — one is even integrated with Outlook (!).
Sure, AmphetaDesk is very nice, but it requires users to install a program. Is this really necessary? I don't believe that there is any reason why news aggregators have to run on the user's desktop, at least for the vast majority of users. A good web-based service would be sufficient for most uses (my news aggregator is a copy of Python Desktop Server running on my server). The load on the RSS feeds would be reduced because each news aggregator service only needs to fetch them once.
All in all, the Lafayette Project sounds like a good idea even without the collaborative weblog recommendations features that Meg Hourihan also talked about.
reboot6 is the best I have ever attended. The speakers were fabulous and the message was inspiring. I especially liked the fact that Thomas did a mini Q&A and peptalk after each speech. That gave reboot6 a personal voice, something the previous reboots have lacked. It was as if the event was created around a central message, and each speaker only treated a portion of that message.
OpenACS 4.6.3 was released a couple of weeks ago. I've just checked out the latest oacs-4-6 branch, and it definitely qualifies as the Best OpenACS Release Ever(tm). Among the really great stuff is the one-page package instantiation and mounting, Tcl callbacks for install, instantiate and mount, and great new packages like the news aggregator.
From Cory Doctorow's site.
You can see O'Reilly's book cover specifications on their rights site, apparently for translation publisher use.
I visited the website of Comendo, one of the reboot startup award contestants.
Tim O'Reilly is up now. I still don't have enough power. There are power plugs in the ceiling, but I am not tall enough, and Tomas simply refuses to jump up from a chair to plug my ThinkPad in.
Unforfunately I won't be able to blog Marc Canter (or Dan Gillmor) because I have run out of power.
Thomas: This is the interface session, and the toughest talk because it is right after lunch. Jason Fried is from 37signals.
Jason is using Keynote!
37signals is a small web development interface shop that also does usability stuff, in Chicago, Illinois. Not flash, no awards — judges and magazines don't care about this kind of thing, but average people care. In interface design it's the little small things that people take for granted which make a difference.
Context and Perspective: Help people relate. Jason is showing a picture of a Pentax camera from their website which is supposedly the world's smallest. However, there is nothing to compare the size with. Finally, he finds a picture of the camera in an Altoid (mint) box, and it fits inside. Putting it in a box of mints really gets the point across that it is a really small camera.
Another example of things where size matters: cellphones. On one screenshot, all the phones are photographed at the same size. The design makes it impossible to compare sizes, but often only give dimensions in millimeters. This is an example of failing to create context and perspective.
Another example: Kicksology.net. On this site, the weights of a pair of shoes (ex. 19.3 ounces) is compared to the weight of a can of soda (13 ounces).
Setting Expectations: Help people know what's next. Jason's parents double-click everything because they are scared of what is going to happen. Amazon have small text messages under their buttons which offer a clear message about that is going to happen.
Wells Fargo, a 37signals customer, built an app to handle loan applications. They increased the rate of customers completing the process with a text, around 100 words, which outlines what the user will expect and how long each step will take.
Example: Meridian IQ case, where the keywords in a sentence are subtly highlighted in yellow.
Contingency Design: Help people get back on track. Contingency design is design for when things go wrong. On the web, things go wrong all the time. This includes things like: Error messages, pages not found, form mistakes, confusing wording, missing plug-ins, bad search results. Jason is showing example from the Contingency Design Hall of Shame.
Blogging is very problematic at reboot6 because of the lack of power. I currently have 30 min of juice left.
Megnut is up. Evan and Meg co-founded Pyra, originally to create a project management tool, but that died because they built Blogger in the meantime. Lately Meg has spent more time writing and evangelizing weblogs.
The title is "A talk about weblogs". Meg describes a weblog as a web page that is chronologically ordered, consisting of small chunks of timestamped hypertext that is timestamped. Anatomy of a weblog post: Title, post body, links, and finally timestamp and permalink. The timestamp creates an expectation of frequently updated content. The permalink is important: Before permalinks it was not possible to refer to specific posts on others' weblogs. Permalinks enable distributed discussions.
For Meg, weblogs are how we should have been publishing on the Web from the beginning. It just took longer to make it writable for everyone than we first imagined. Pyra wanted to make it easy to write weblogs. Now Meg has founded a company to make it easier to read weblogs. TypePad is an example of the ever more powerful tools that make it easier to create weblogs.
If you have written a book review of a book that you really like, you can post a review on Amazon that you can share with the users there, but you also want to post it on your weblog. It's a lot of trouble to post this twice. It should be possible to tell Amazon that they should point to the review on my weblog.
There is a gap between all the people publishing weblogs and the general population gaining access to weblogs. In terms of social software, RSS readers make it easy to read different weblogs. However, if there are 10 million RSS readers hitting RSS feeds every 10 minutes, they are interacting with all this weblog content in an almost one-to-one manner.
The Movable Type folks created TrackBack, which creates explicit connections between weblog postings, but this is also a very technical idea that a broader audience would not be able to understand.
The Lafayette Project's goal is to make it easier for the general public to read weblogs. It's not weblog search. When you find a weblog that you like, you usually like the person behind it. You trust them to find things that you find interesting. TLP creates connections between weblogs. A beta should be out in July. The first version will just be an RSS reader for the web. On top of that they will build the recommendation engine.
Over 50% of weblogs are not in English. When we have all this content that is not English, we are having links between weblogs that are not in the same language. Non-English weblogs can link to English-language weblogs, but it does not often happen the other way around. TLP wants to create non-English interfaces to the tool.
There might be a weblog backlash because it has become so popular that everyone will start hating it. Meg no longer believes that weblogs will be just a fad. More interesting content is being written on weblogs than ever before.
Thomas: Scott is a founder of iTraffic, and is basically to blame for a lot of the banner advertising that you see on the Internet. However, he is not a bad guy because he is creating an exciting new business that allows people to connect.
Scott was watching the webcast of Douglas Rushkoff's speech at reboot4. He kept on quoting, "Find the Others". The Internet was about people and connecting them together. This led, among other things, to meetup. Scott has had a life-long ambivalent relationship to advertising. When he set up in the Internet advertising business, he tried to do the Right Thing for advertising on the Internet, and try to avoid Madison Avenue. Later he sold iTraffic to Omnicom, the Madison Avenue evil incarnate.
Scott also read "Bowling Alone", a book by Robert Putnam, who argues that people don't really know their local neighbourhood anymore. In the past 40 years, people have lost touch with their communities. There has been a decline of lobal community but at the same time the rise of the Internet which creates new communities centered around topics of interest. The proselytizers of the Internet say that the Internet makes geography irrelevant. How should the Internet do something about this decline of community? They locked themselves in a room for 3 months and built meetup.com to create physical communities. Meetup has brought people together. Meetups happen all around the world.
Meetup is being used in new ways that the creators did not imagine. Some people believe that Meetup is the ultimate political activism tool. The Howard Dean For President campaign organized 270 meetups through meetup.com. Metcalf's Law can be questioned by the results of enablers such as meetup.com.
Who pays for Meetup? Local establishments that want to be listed pay to be recommended. Users that want extra features pay. Organizations such as the Howard Dean Campaign pay for special services.
The big trend:
Before 1950 People used to gather in groups.
Around 1960 people would watch TV in groups.
From 1980 people watch TV alone.
From 1990 people use non-Internet PCs.
From ca. 2000 people connect via the Internet.
From 2003 people use the Internet to gather in groups.
Thomas: Ben Hammersley is a hardcore journalist, for the Guardian among other, and he is also one of the leading proponents for the Semantic Web.
RSS is a format that enables us to share content from websites, and aggregate them in newsreaders. The semantic web is a big buzzword that has been around for a couple of years and is very big, but people are getting tired of it because nobody really has any idea what it is. Everyone has their own idea of what the semantic web.
Actually, we already have the Semantic Web. Today Ben will tell us what it really is. Pico, Ben's dog, is very cute and likes to curl up under blankets. Ben wants to be able to describe Pico to the Internet. He is 20 months old, 35 cm tall, etc. Ben is explaining the concept of triples, the basic building block of the semantic web. It has 3 parts: A subject, a verb and an object. So Pico is 20 months old = Pico hasAge 20months.
You can make Semantic Web statements in any language you want. English, French, XML. It doesn't really matter. Of course, natural languages don't really work in computers. Another reason we need a specific way to write down these triples is that we want to share our collection of statements. Consider "I love you" — the definition of the words change depending on who says them. We need to be able to define precisely what the terms mean.
The other key point of the semantic web is called the URI (Unique Resource Identifier). It looks very much like a URL, but is unique to the thing you are talking about. Ben is showing the RDF for his weblog. He points out the tag dc:creator, and dc refers to http://purl.org/dc/elements/1.1/, which is the Dublin Core, a collection of concepts. When Ben says creator, he refers to creator in the contect of the Dublin Core vocabulary. This gives us an unequivocal definition of "creator". Instead of saying "Ben made Ben's site", we can say "Ben-URI dc:creator Ben-Site-URI".
There are hundreds of vocabularies out there. You can start making up your own vocabularies. Examples of RDF vocabularies include RSS 1.0, FOAF, ThreadsML and MusicBrainz. Anything you can describe can be written using triples.
Compound queries are the killer app of the Semantic Web, and RDF enables us to do compound queries. Suppose you have tickets to Roskilde and want to give them to someone in your company. You might want to search for "everyone in this company who like these 3 bands, and live within 50 miles of Paris". RDQL is like SQL for the entire web.
There are some problems. First of all, the user interface. It's not Google. It's not one application. You cannot do semantic web searches with keywords. Keyword searches are easy. You cannot write down compound queries in Google without learning some kind of language. However, nobody really wants to learn new languages. How to get rich: Design a user interface that enable people to perform interesting queries with the data that people are putting out.
After we resolve the petty problems with how we put our angle brackets, the real issue is getting people to create this data, and to develop user interface.
The talk is called "How to go broke on the Internet". Almost no creative people (writers, musicians, etc.) make a living off what they love. Some artists have tried to find out how to use the Internet to make a living. Creative Commons is a license created by Lawrence Lessig. It's a system that allows artists to say that only some rights are reserved, and all other rights belong to the public. Cory put his latest book under the CC license.
Artists do not create because of rational economics. Creators desire audience rather than monetary compensation.
Today we have peer-to-peer networks and other technologies that undermine how authors make money off their works. Whether you believe the numbers of IPFI or RIAA, which you probably shouldn't, it is clear that there is something out there which threatens the way revenue is traditionally derived from the arts.
Today the industry is trying to get some of these technologies before, like radio and Betamax. Of course the media industry makes more money from pre-recorded media than from the box office. It seems that they have to be dragged to the money tree. Imagine how stupid the revenue model (with advertising) that was developed for radio would sound. Today we are presented with a variety of revenue models for art on the Internet, and we know how stupid they sound today.
In the day of Marconi, most people of course chose the wrong revenue models, and ended up not being artists anymore. Clay Shirky: There is no demand for music that is less flexible. No-one will say, "I will pay a premium for music that gives me less power." There is no market demand for proprietary formats, a little like MiniDisc and DVD.
This is why the DVD format hasn't really gotten anywhere. This is because no-one can make a DVD player without the permission of DVD Forum. Remember what happened with VHS in the early days. People starting taping births and police beating up people. Given the choice between a proprietary and a non-proprietary format, people will not choose a proprietary format.
Cory is asking someone to hold up a Nokia phone and take out the battery. It has 4 leads. One is positive, another is negative, and the others are crypto. There are $5 worth of crypto in the battery, and ¢50 of crypto in the phone to protect your conversations. Nokia spents an order of magnitude more to protect their business model than to protect your privacy.
Again and again, technology is used to harm us. Businesses will fail because of this and blame the Internet. This happens because they commit a fatal sin of capitalism, to create something that there is no market demand for.
Somehow, there is a market demand for compensation for artists. Most people agree that artists should be compensated for their work. In Europe there has not been a strong sense of fair use. You pay a levy on your blank media and you can do anything you want with that media. There is no provider of artist compensation on the Internet. Apple iTunes compensates the labels.
We're losing end-to-end, the right to build a service on the Internet without asking someone's permission. We're losing academic freedom. The University of Wyoming is building a packet sniffer that opens up every packet and look for enfringement. Imagine if Joe McCarthy had gone to Wyoming, where all the deans wear cowboy hats, and asked them to listen to every conversation and find communists.
Under the EU's IPR Directive, if you sign an affidavit that your rights are being infringed upon by another company's customer (like an ISP), to seize that company's computers for 31 days.
Can we fix this? Can we fight these threats? Cory believes that we can find a solution that we can reconcile technology with copyright. It is called a compulsory license. The idea is that you pay a compensation, a fixed fee, and in exchange you get the right to use copyrighted works. The US already has the Internet radio license. You can play any song you like, and a fixed proportion of the license fees go directly into the artist's pocket. We can fix the copyright wars with a compulsory license and some techonology. We need statistical sampling, but billions of dollars are already distributed using statistical sampling, such as the Nielsen families who record the TV that they watch.
What we should watch out for is a levy on media without compulsory license. The Swedish implementation of the EU Copyright Directive imposes a levy on blank media, but in exchange you get — nothing. You are still a criminal.
If we are careful, we can solve the copyright wars if we talk to the government, to technologists and to artists. The one way we cannot solve it is to continue to walk down the road that the recording industry wants us to walk down and to attempt to serve market demands that do not exist.
We're only 16 minutes behind schedule. Mygdal is introducing this year's reboot and the 4 teams that are facilitating the conference. Alexander Kjerulf is the Open Space host. The Bigger Picture folks will visualize and draw the knowledge that arises from the event. And 3 anthropologists will investigate the people of reboot and to document the human angle. Pollas is introducing the technology that will facilitate reboot.
ComputerWorld is reporting (in Danish) that Basse Bergqvist, an entrepeneur who is something of a legend in Denmark for his failed Internet ventures, is starting up a new company that offers a bundle consisting of a broadband line and unlimited domestic telephony, kind of like Vonage. This will cost DKK 400 per month.
I've wondered fro a while whether it will be possible to launch a Vonage-like service in Denmark with unlimited phone calls. The problem is that most of the people you would want to call in Denmark are connected to the traditional PSTN system, which is billed by the minute. Unlike the United States, there are no toll-free local calls. Per-minute costs might be lower because one end of the loop (legs 1, 2 and 3 in Ciscospeak) are VoIP, but still, it seems very risky to offer unlimited phone calls. Perhaps it's possible to insure against this kind of risk?
Aaron Swartz has created a rendition of Tufte's PowerPoint essay in PowerPoint format.
| You are 34% geek | You are a geek liaison, which means you go both ways. You can hang out with normal people or you can hang out with geeks which means you often have geeks as friends and/or have a job where you have to mediate between geeks and normal people. This is an important role and one of which you should be proud. In fact, you can make a good deal of money as a translator.
Normal: Tell our geek we need him to work this weekend. |
Just saw George Clooney's directorial debut. I like it. So does Ebert. This movie tells the story of a tv producer who works as an assassin for the CIA, wastes his life, but ends up marrying the girl (Drew Barrymore, very cute). Ebert claims that “all autobiographies are fictional,” and he is probably right, but that does not make this movie or the story any less appealing. A 9 on IMDb.
At one point Barris tells Julia Roberts that he always wanted to be a writer who would be quoted by ‘lesser minds,’ but he has realized that he is himself one of those lesser minds. He creates low-quality game shows that humiliate contestants. He kills people because he likes it. Is being a lesser mind the prototype autobiographical nightmare? I think that is the morale of the closing scene.
Barris and Patricia (the Julia Roberts character) quote literature a few times. Interestingly, Kathleen Donohue believes that one of the Nabokov quotes (“All the information I have about myself is from forged documents”) is apocryphal.
I just noticed that the website of Berlingske Tidende, Denmark's biggest newspaper, uses HTML and CSS positioning instead of table hell. It's a shame that it doesn't validate, but it's still something. Works in all major browsers too.
