Tantek's Thoughts

Note: CSS restyling of Top 100 Technorati in progress. Current design based on WilWheaton.net. See Alternate Stylesheets (e.g. with the "Choose Style Sheet" favelet) for previous designs.

  1. 20040229

    1. 2359

      Snowy, wet and productive

      • web
      • travel

      This morning it snowed. Yes, you read that right. Here on the French Riviera. At the end of February. First thing in the morning. It snowed. David Baron took photos (he posted one). Of course the snow melted and then it was wet. Though it was wet yesterday as well from the rain.

      And it was/is cold. This means we stayed indoors a lot and got a lot done. It was a productive first day for the CSS working group. In the evening we went out to dinner, and the weather was cold enough to lose feeling in your ears after walking only 15 minutes. Dinner was excellent though.

      We ate at Le Brasier on Place Fontaine in the nearby small town of La Napoule where I had a delicious saumon pizza made with actual cooked fresh salmon (not smoked), cheese, tomato sauce, onions, crème fraiche, and oregano. Dessert was a healthy serving of mousse chocolat, washed down with an espresso.

      Of course then I stayed up until the early morning working on responses to CSS3 Basic UI last call working draft comments.


      1. David Baron
  2. 20040228

    1. 1617

      Mandelieu Arrival

      • web
      • travel

      I made it here to Mandelieu without incident. David Baron happened to be on the same flight, which had a connection in Heathrow airport. I have long despised traveling through Heathrow because they often leave insufficient time between connecting flights to first catch a shuttle from one terminal to another, and then trek the long and awkwardly laid out hallways past what seems like a lot of empty space to a gate with no visible indication of what flight is departing next.

      At least this time the cross-terminal transit was fairly efficient, and there was even enough time to grab a snack at the pizzeria Est Est Est in Terminal 1 of Heathrow Airport, London, which unfortunately I cannot recommend. I had a Margherita pizza which was pathetic — just a bit of cheese on some flavorless thin dough with a dabble of tomato sauce in a ring around the cheese.

      The flights were good though. Decent movie selection on British Airways , including Matrix Revolutions. The vegan meals were delicious. Landed in Nice just after sunset. Got overcharged (a bit) for a taxi to the hotel.

      Oh, what am I doing here? I am here for almost a week of W3C meetings at the appropriately named 2004 W3C Technical Plenary Week. It was here also two years ago. Last year it was in Cambridge. As usual, there's an IRC channel if you want to observe (or join in on) the fun:

      • /server irc.w3.org 6665 to connect
      • /join #tp to join the technical plenary channel
      • /list to see the list of channels
      • /who to see the list of who is on your current channel

      And while I'm in GMT+1, my posts are still in Pacific Time, which makes it just past 1am here so I should get to bed so I have a chance of being awake for the 9am CSS meeting (and breakfast beforehand) tomorrow morning.

  3. 20040225

    1. 1813

      New and updated CSS specifications!

      • web

      The past couple of days the W3C published a slew of new and updated CSS specifications.

      I have to admit, I've spent many many long hours over the past few years discussing, editing and writing some of these, so I can't help but be pleased. Congratulations to everyone who contributed to these documents. Thanks for all of your efforts and feedback.

      And speaking of feedback, you know where to go to give more.

    2. 1805

      Creative Commons supports rel="license"

      • web

      A few weeks ago I was coming up with examples of semantic XHTML for Kevin Marks and my presentation at ETech, real world semantics. Dan Connolly suggested on the wiki page that we discuss Creative Commons (CC) among other things.

      Thus one of the examples I came up with was to use a <link> tag with a relationship ('rel' attribute) of "license" to semantically link to a Creative Commons license (or any other type of license for that matter) for a web page. E.g. if you wanted to state that a web page was licensed under the Creative Commons "by" (attribution required) 1.0 license, you could simply put the following XHTML in the <head> of the web page:

      <link rel="license"
            href="http://creativecommons.org/licenses/by/1.0" />

      And just like a <link rel="stylesheet" ... > effectively says "this is a style sheet for the document", the above code says "this is a license for the document".

      As several documents can share a common linked style sheet and thereby avoid the tedious, error prone and inefficient hassle of duplicating style rules everywhere, documents can simply link to a common license, and thereby avoid the redundant, and "inelegant" use of RDF inside HTML comments (related: Escaped Markup Considered Harmful).

      The beauty of using the 'rel' attribute is that this technique not only works for <link> tags, but also for <a href>.

      A typical page that is licensed under a Creative Commons license has a nice (CC) Creative Commons button <img> button on it, typically coded something like this:

      <a href="http://creativecommons.org/licenses/by/1.0">
       <img src="http://creativecommons.org/images/public/somerights.gif"
            alt="Creative Commons License" >

      Just by adding rel="license" to the code:

      <a rel="license"
       <img src="http://creativecommons.org/images/public/somerights.gif"
            alt="Creative Commons License" >

      the page now simultaneously, all in one place, indicates the visual (using the image) association with the license, a user accessible association with the license (with the user clickable hyperlink to a human readable page), and an auto-discoverable association with the license (with rel="license").

      After our ETech session, Kevin Marks and I reprised our presentation at Concon. Almost simultaneously (synchronicity), Mike Linksvayer posted to the cc-metadata list, referenced our presentation at ETech as inspiration, and pointed out that <a rel="license" href="..."> is just as good as RDF, and can be used to generate RDF. He also noted a very important feature of the lowercase semantic web:

      A neat thing on the presentation side is that CSS selectors can actually change the document rendering based on rel attributes -- making the metadata not just close to the presentation, but part of it.

      I wonder who will be the first to post a user style sheet that demonstrates this.

      Subsequently Mike updated both the Creative Commons Metadata documentation page with the <a rel="license" href="..."> technique, and the Creative Commons Choose a License page to generate the rel="license" markup. So it's official.

      A big thank you to Mike Linksvayer for making it happen.

      Another nice aspect of this technique is that by indicating where to get the license for a document in this simple machine parsable way, you eliminate the need to redundantly embed RDF license markup inside markup comments in all your pages. Just link to the license and let the license maintainers at the other side of the href worry about the RDF.

      As soon as I saw that the Creative Commons folks had adopted and recommended rel="license", I updated all of my Creative Commons licensed CSS Examples accordingly.

      Why not do the same and make your CC licensed content discoverable as well?

      If your CC licensed content already was discoverable with RDF markup nested in comments, use rel="license" instead to make your pages smaller and to avoid comment nested markup hacks.

      Thanks to the Creative Commons folks for providing another great example of real world semantics.


      1. Eric Meyer License To rel

        Eric, it is unlikely that folks will include the "c-commons" value in the rel attribute as it is redundant information (the URL in the href already has that info, and you need the URL, but you don't need the extra rel value). Also, *[rel~="license"] is too general and may pick up false positives from non hyperlink elements. Finally, I think a user style sheet needs to key off of an href to a particular CC license so that it can display the appropriate icons for that license. Here is one such selector to start with.



        P.S. Awesome photo of Carolyn.

      2. Eric Meyer
      3. Mike Linksvayer
      4. Serge K. Keller
  4. 20040224

    1. 0125

      Concon roll and late night coffee

      • friends
      • city
      • web
      • tech

      Let's see if I can remember the folks I met or bumped into a week ago at Concon:

      I'm sure a week after the fact my memory is randomly failing to recall some folks. Consider yourself encouraged to speak up and let me know.

      Just after Concon, Jonas, Kevin, Peter and I took off in Jonas' jeep to go find Boris (Bopuc) and his friend who were both staying at the Holiday Inn on Van Ness. We found them and grabbed coffee at Royal Ground Coffee at 1605 Polk st. where there were a few open wifi )( networks.

  5. 20040223

    1. 2359

      Concon notes

      • web
      • tech
      • city

      The past seven days have been insanely busy, but fortunately, also insanely productive. I took a bit of an involuntary blog break, so there's some catching up to do, starting with the posting of notes from last Monday night's Concon.

      Paul Schreiber and I took notes over Rendezvous using Subethaedit (which was quite convenient because I left my computer on during my talk and Paul took notes on my presentation while I was speaking).

      For now here's the original plain text notes. Maybe I'll mark them up later.

      Jonas also had a few things to say.

  6. 20040219

    1. 2224


      • web
      • tech
      • city

      Danny O'Brien kindly organized a nice post ETech event this past Monday night called ConCon which was hosted by the RX Gallery, the same place that is currently hosting a classic video games exhibit.

      In short, a fast-forward summary of ETech sessions:

      The format is a casual, five-minute lightning talk with a friendly audience, with the emphasis less on five minutes and more on questions and answers. We'll just go through names until we run out of time. Then we'll have fun.

      Kevin Marks and I gave a reprise of of our real world semantics session.

      Ambiguous.org posted one line summaries of all the five minute talks . I liked they're summary of our talk:

      Kevin Marks & Tantek Çelik: hey everyone, we found some real live people entering metadata!

      That sounds about right. Real live people entering metadata, which is what happens when you keep the formats simple and easy to integrate with what they already understand (XHTML). They also posted several moblog photos, including a photo of me presenting.

      Niall Kennedy also wrote a brief summary ConCon. As did Danny.

  7. 20040217

    1. 2118

      ETech roll

      • web
      • tech
      • friends
      • travel

      etcon butterfly Before my memory fades too much, I figured I would jot down all the folks I bumped into at ETech. Some of these folks I met for the first time. Of course I'm using XFN to markup my relationships, and thus they will all have the value "met" somewhere inside their 'rel' attribute. If this is the first occurance of the value "met" in reference to a particular person/hyperlink from me, then you (or an XFN crawler) could conclude that I met them at ETech. Did I miss anyone? Email me your URL.

      In alphabetical order by last name (but the order itself has no meaning so yes it's semantically an unordered list), and some with additional ETech related links (I'll add more as I find them).

      And there were many more folks there that I either I saw but didn't get the chance to meet, or heard they were there but didn't see. Yes, it was a bit overwhelming.

      Trivia: Rael Dornfest and I discovered a superb Indian restaurant near the conference hotel: Monsoon.

  8. 20040212

    1. 1231

      Fixing the OSX Finder

      • mac

      Eric Meyer pointed out to me that he had already dealt with the Finder shortcut problem I previously mentioned.

      The way the Finder has worked for years and years (like10+):

      • Cmd-N creates a new folder

      The way the OSX Finder works:

      • Cmd-N creates a new Finder window

      What were they thinking?

      What genius thought it would be a good idea to violate millions of man years of muscle memory for quickly creating a new folder on the Macintosh?

      Were they possibly thinking that by copying Windows' behavior (in Windows File Explorer, Ctrl-N creates a new File Explorer window), that they would make it easier for Windows users to switch?

      Or was this just yet another instance of NeXT user interface folks overriding Mac user interface folks?


      Just fix what Apple broke.

      Eric has documented how to correct the problem, and set Cmd-N to create a new folder in the Finder, as it was always meant to.

      Now, anyone know how to fix Windows File Explorer to do that as well?

    2. 0012

      ETech session posted

      • web
      • tech

      Kevin Marks and I just wrapped up our ETech participant session. I've posted the presentation:


      Now, onto the l0pht party.

  9. 20040211

    1. 1108

      Emerging technologies

      • web
      • tech


      Yesterday I added GeoURL to my blog in like two seconds (maybe minutes). It was trivial. An excellent example of "RealWorldSemantics", also known as the "lower case semantic web". Go check it out.


      1. Jonathan Stanley

      XFN news: diveintoxfn and 1 in 12

      Well known blogger and accessibility (among other things) advocate Mark Pilgrim has reintroduced a blogroll to his blog and it's XFN Friendly.

      Phil McCluskey just reported that 'The response so far to rubhub has been great.'. Apparently: "rubhub has crawled close to 1000 sites, with a 1 in 12 ratio of XFN users to non-XFN users."

      Wow. That's an 8% uptake. Amazing.

      Trying OSX

      I've decided that ETech provided an excellent opportunity to try using OSX. Fortunately OSX enables OS9 applications to run through the use of its "Classic" compatibility process. I'm talking about OSX v10.3 Panther in particular, as I'd tried out previous versions (alphas, betas, 10.0, 10.1, 10.2) and they were all terrible, pathetic, and not worth using. In short it became apparent very quickly that they were a complete waste of time. v10.3 is very different.

      I've managed to setup my primary applications which store user data in preferences-like locations (Entourage, Internet Explorer) to share their data across my systems using aliases. E.g. my "Microsoft User Folder", "Favorites.html" and "Internet Scrapbook" are all aliases that point to the real location of that data. I'm using IE5.1.7/MacOS9 on Classic because I don't know if there is a way to get IE5.2/OSX to use a shared Favorites and Internet Scrapbook.

      The conference has provided an excellent opportunity to try out Rendezvous, a new technology available only in OSX on the Mac. So far I've used only two applications that use it: iChat and SubEthaEdit. There are dozens of people in my Rendezvous window in iChat, and I can easily drag and drop them to my Buddy List since Rendezvous is broadcasting their AIM/mac.com screen names. This is the easiest way to build a Buddy List that I have seen. Much easier than having to type in lots of email addresses over and over.

      SubEthaEdit (formerly known as Hydra) is basically a text editor that allows for folks to co-edit the document with you over Rendezvous. You click a button to share the document, and then anyone running SubEthaEdit can see that you've made it potentially available to share. They click on it to try to edit it, and then you approve them (or not) as co-editors. Everyone's edits show up in different colors. When you save it saves locally on your machine.

      In fact, Kevin Marks and I are using SubEthaEdit to cooperatively develop our presentation for our Participant Session later tonight (10pm in Ballroom C) (RealWorldSemantics), even while attending different presentations here at ETech.

      A few good things I've noticed about OSX:

      • It boots (on my iBookG3/900) anywhere from just under a minute to about three minutes. I have no idea why there is such a disparity.
      • It boots Classic faster than OS9 boots natively
      • It launches Entourage (the OS9 version in Classic) faster than Entourage running on OS9. Perhaps the file system in OSX is so much faster that it speeds up even Classic applications that are file I/O bound?
      • BBEdit 7.1 (a super text editor that I highly recommend) automatically launches native in OSX or native in OS9 depending on which you have booted.
      • When one of my Classic applications appears to hang, I can switch to some other OSX app and do some work while the Classic applications finishes what it was doing or I decide that the Classic app has actually crashed and decide to force quit it. In other words, in OSX, no one application can hang my machine. That's nice for those of us that have short attention spans and can't stand response latency of any amount.
      • Wake up from sleep (and going to sleep) is nearly instantaneous. My backlighting on my iBook now "feels" as responsive as the proverbial refrigerator light. No weird "waiting for the network to do something to close down" etc.

      I'm still annoyed that cmd-N opens a new window rather creating a new folder (which now requires cmd-shift-N) in the OSX Finder. I vaguely recall hearing about some 3rd party extension that fixes this. Hopefully someone around here will know where to find it.

    2. 0953

      ETech Summary

      • web
      • tech

      It's been a very busy time here at ETech, as expected.

      Yesterday the conference kicked off with an inspiring keynote by Tim O'Reilly and Rael Dornfest, followed by a very scary keynote by Helen Greiner of iRobot. Bottom line: it looks like iRobot is pioneering the precursors for Skynet. Sure, those vacuum cleaners look cute. But wait til you check out the indestructible military bots which "can carry many kinds of payloads". Yes she actually said that. Suddenly I'm not really interested in a vacuum cleaner bot.

      All in all, the sessions have been an interesting mix of technology and social analyses. More interesting than the sessions however have been the chance meetings with people whom I've only either met online or have only heard of but not spoken with.

      Shortly after arriving Monday afternoon and getting settled in the conference hotel, I headed down to the hotel lobby bar where the wifi was plenty and people were sociable. There I met Rael Dornfest (conference chair) who then introduced me to Tim O'Reilly.

      There's been running Orkut-jokes throughout the conference. It's like an Orkut-a-holics anonymous convention. Proof in point: Tim O'Reilly just approved my Orkut friend request, during this morning's keynote by Marc Smith. Yes, the auditorium is filled with open laptops and everyone is on email, IRC, rendezvous, iChat etc. The network performance is a bit spotty.

      Marc Smith's keynote featured tons of cool charts and visualizations of Usenet usage patterns, visual profiles of different types of Usenet participants, who apparently behave similarly over a number of years.

      However, the following keynote, the Nokia presentation, was (is) terrible. The presenter started out with a single slide of random images that made no sense, and left it up for half an hour while he droned on with buzzwords loosely strung together in a dullingly monotone voice, e.g. "maximum innovation with peer-to-peer openness". Oh look, he just changed the slide to a photo of a cell phone. Oh look, he just changed it back to the random image collage.

      The nice thing about boring presentations is that they are the perfect opportunity to catch up with blogging.

  10. 20040208

    1. 2154

      Social software snippets

      • friends
      • web

      There's been a lot of activity around XFN recently. The XFN: What's Out There? page is quickly outgrowing its original design. Here's a selection of what's new:

      All this in less than two months after XFN was introduced. I continue to be amazed by how fast people are picking up and supporting XFN. The GMPG folks designed XFN to be very simple to use and understand from the beginning, with the hope that this user-centric focus on simplicity would spur quick adoption. It seems to be working.

      Somewhat relatedly, Jonas wrote an excellent essay analyzing various social software technologies, like FOAF, vCard, XFN, Friendster, Orkut and Tribe. A few interesting quotes:

      FOAF, which seems to be little more than vCard data in a new format (RDF). Why the designers of FOAF chose to use RDF rather than an existing and widely adopted format remains a mystery to me, maybe there are advantages, maybe it's just the old NIH [Not Invented Here syndrome].

      Indeed. There does seem to be quite a lot of NIH on the Web, as folks working on Web technologies mistakenly conclude that a problem that is new to them is a problem that hasn't been solved before, and thus deserves a novel solution. I heard about vCard many years ago, and today, as Jonas says:

      vCard has a long list of supporting tools, from most mail applications to PIMs and address books.

      And vCard is an open standard. So Jonas poses a good question, why was FOAF invented? Perhaps an attempt to justify RDF? Is this an example of a solution looking for a problem? A clever acronym looking for a reason to exist?

      Speaking of which, the biggest irony I see about FOAF is its name, which stands for "Friend of a Friend", and yet the technology has nothing to do with "friends". Like Jonas said, it's simply vCard recast in RDF, except for the "knows" relationship, which itself is quite meaningless (quite ironic for a Semantic Web effort), as it provides no more meaning than a plain hyperlink. Why work so hard for so little?

      FOAF could be saved however, simply by adding an XFN module, thus enabling FOAF to finally fulfill its namesake, and represent friendship rather than just claiming to.

      Jonas goes on to compare XFN and various centralized services:

      XFN has a few advantages over communal systems such as Friendster, Orkut, or Tribe, in that it correctly recognizes and displays asynchronous relationships, such as Anna calling Bob a friend, but Bob considering Anna merely an acquaintance. Relationships in 'the field' are rarely as clean cut and simple as Orkut's mutual friends paradigm would suggest.

      I think this is an important point. Not only are relationships more than just "friend" or not, but they are not necessarily symmetric. XFN handles both of these aspects (relationship spectrum and relationship asymmetry).

      Jonas brings us right back to what we started discussing at the beginning of this post:

      XFN, assisted by tools such as RubHub could be the answer.

      By the way, Jonas isn't your average geek, he is a sociologist by training and schooling. When Jonas talks about social software, he's not just talking a bunch of smack like any other technophile. He's an expert telling you how it is.

      And for everyone who made it this far, here are my photos from Friday night's Orkut launch party.


      1. Henning
  11. 20040205

    1. 0554

      Finished my CSS 2.1 CR edits!

      • web

      Phew. This past monday we (the W3C CSS working group) caught up with resolving all outstanding CSS 2.1 issues to date, and yesterday Ian finished updating the CSS 2.1 issues list with resolutions, each annotated with respective action items for specific working group members.

      After several hours of editing, I just finished making the necessary edits per those resolutions to the CSS 2.1 chapters that I'm responsible for, so unless new issues are raised that affect those chapters, I'm finished with my CSS2.1 Candidate Recommendation (CR) edits!

  12. 20040204

    1. 1311

      ETech Conference and Real World Semantics

      • web
      • tech
      • travel

      etcon butterfly I've decided to go to the upcoming O'Reilly Emerging Technologies conference (ETech). I'm still looking over the schedule to figure out which conference sessions I want to attend.

      I've also created a page on the ETech Wiki for RealWorldSemantics, my proposed ParticipantSession for signups and suggestions.

      If you are attending ETech, and want to discuss "Real World Semantics", please visit the Participant Sessions page, append a '+' to the end of the description for "Real World Semantics", and add your name and suggestions to the RealWorldSemantics page.

      See you at ETech!