Anne van Kesteren


Quite recently the HTTP State WG at the IETF was started up again to work on cookies. As far as standards go cookies have a somewhat unfortunate history. Initially there was a specification from Netscape that was fairly low on detail and did not quite match reality. Then there was RFC 2109 which had some more detail but did not match reality. Then there was RFC 2965 which introduced new syntax hoping the whole web would move to something new. It obviously did not match reality. Now a new cookie draft is almost done and ready to get an RFC number. Adam Barth is the main driving force behind this fixing-the-cookie-problem project. This draft defines in careful detail how cookies actually work, how dates in cookies are to be parsed, and other annoying details implementors had to reverse engineer from each other to date.

Anyone else see the HTML1, HTML4, XHTML2, HTML5 pattern?