Ingy 2.ö

Monday, June 30, 2008

Coffee & Diffie


I've been hanging out in Palo Alto for the last week, meeting new coworkers at Socialtext, and visiting old friends. This last Saturday morning I made a new friend. I was waiting for my coffee and I spied a man sitting at a table using a laptop. Not a rarity around these parts by any means. But this was a man I was pretty sure I'd seen before at a Foo Camp some years ago... "Excuse me, are you Whit Diffie?" "Yes, I am." "Hi, my name's Ingy döt Net"... Whit and I talked a bit and then I ran off to have breakfast with Lyssa.

This morning I ran into Whit again and we ended up talking for about an hour or so. I think I am cursed with this condition that when I meet someone with interesting ideas, I feel compelled to turn those ideas into projects. Whit left me with two interesting ideas...

1) While discussing the concept of "ingy.net", he impressed on my that a query url is a long lasting resource of good information on a subject. For instance http://ingy.net (this blog) may become dead some day, but http://www.google.com/search?q=ingy+dot+net is much more robust. This led me to the idea of creating a blogware wrapper, that basically contained the latest google info about me, without me needing to lift a finger.

So far I have create http://i.ingy.net. For now, it is a simple redirect to google, but may be more interesting at some point. It might be an interesting meme to make domains like 'i.' or 'iii.' to mean, "general information". At this point "www" has very little meaning, but lots of people type it in regardless.

2) Whit is interested in NSA factoids and during some talk about this and Wikis, we came to the conclusion that a wiki would be a decent tool for a group of people with that common interest to share what they know. The main difference is that this data is more structured in nature and wikis are pretty much about free form content.

This discussion rekindled my interest in two older ideas of mine: Kwiki and CogBase. I think that there is a lot of potential in what I would called a "typed wiki", a wiki who's pages have a type and that type implies a certain schema (aka required data). That way you could set up a wiki that contained both free form prose, and also pages that made you fill in a form for certain data. Imagine a baseball wiki. You could have pages on players that required all the typical "baseball card" info and also free formed pages that talked about things like a certain famous game.

Whit was quite an interesting man, but given my current project schedule, perhaps dangerous to meet with too often. ;-)

2 Comments:

  • "typed wiki's" are called structured wiki's. A good example is e.g. wikilens, a phpwiki enhancement.
    http://wikilens.org/wiki.php/WikiLens/StructuredDataTutorial

    By Blogger rurban, At July 2, 2008 10:37 AM  

  • It seems more useful to do "type inference" e.g. on Wikipedia: find structural patterns that Wikipedia pages tend to meet and then make refactoring recommendations (or actually carry them out!) according to the findings. E.g. consistent application of templates, consistent application of headings, consistent application of categories, etc.

    By Blogger rp, At March 12, 2009 11:11 AM  

Post a Comment



<< Home