Tastes like chicken

http://wso.williams.edu/~dgambrel/catgirl/cgold.gif

http://www.catandgirl.com/

(from sbisson)

Instead of PKI

Remember I said that PKI deserves to die?

This Clay Shirky essay states the problem with names in a really nice way (it's an independent reinvention of the ideas in this one but in clearer language)

Mark S Miller has taken the shape of the problem and shown that it's the shape of the solution! Here's how names should really be handled.

At first I thought "who would go to the effort of writing <pn>Alison</pn> all the time?" But the answer is, of course "anyone who can be bothered to write <lj user="purplerabbits">" - ie, it seems, nearly everyone who uses LJ.

This is really neat. And it's just become an essential part of my ideas on how LJ can become decentralised.

I can imagine the user interface already. You see a story that says

went to the cinema with ?Connor, who insisted we watch Spiderman.

Click on the "?Connor" and you get a little interface that tells you about the name in terms of names you already know.

* This user prefers the name Connor, but you already have a user of that name in the database (edit)
* Grant knows ?Connor as Daniel (accept)
* David knows ?Connor as Gingernuts Johnson (accept)
* Enter your own choice here _______ (accept)

[ ] Publish choice

Click on the appropriate "accept", and that name will be subsituted in that and all future appearances:

went to the cinema with Gingernuts Johnson, who insisted we watch Spiderman.

(U: replaced "?Ciaran" with "?Connor")

Update: A burst of nostalgia for the day we lost our naming innocence, just over eight years ago (and prescient words a few months earlier)

Neat XML things

I seem to be using LJ as my bookmarks, so here's a URL I keep losing: a nicer way to write XML in Python-like syntax

http://www.langdale.com.au/SOX/

html>
    head>
        title> My Home Page
    body>
        h1> Contact Details
        p>  I can be contacted at
            a>  href=mailto:me@myplace.net 
                this address
            except when on vacation.

and a similar, XSLT-specific thing:

choose
  when .test "$count > 1"
    call "by-columns" ("column-count":"$count")
  otherwise
    <para>
      "No columns"