Arena

~ the goodbye ~

Arena's audience was the people writing the specification, and I was never going to be in that room. It existed so that an argument on a mailing list could be pointed at a running window and settled. By 1998, when my own line finally worked, the arguments were finished and I was living inside the answers without knowing where any of them had been tested.

Dave Raggett started it in 1993, for the plain reason that he had written a proposal called HTML+ and there was no browser on earth that could render it. Version 0.91 went out publicly on 24 October 1994. From July of that year Håkon Wium Lie worked with him on it at CERN, porting it to Unix and wiring it into the W3C's reference library, and from July 1995 the INRIA team at Sophia Antipolis took over the maintenance with Yves Lafon on forms and graphics formats.

It ran on X11, which is to say on workstations. The hardware it needed, from NeXT machines to the Sun and SGI boxes, sat in research departments rather than in houses. That was the correct audience for what it was, and it is also why almost nobody has ever seen it running.

~ the chronicle ~

~ what it was built to prove ~

Cascading style sheets shipped in Arena as an experiment in version 0.96, years before a browser with users would take them seriously. Håkon Lie had co-authored the CSS proposal and Arena was where he could show it working instead of describing it. Every time you set a font in a stylesheet today you are using something that had to be demonstrated somewhere first, and this is where.

The same is true of the PNG format and of the tables and math in the HTML 3 draft. Arena was the machine those things had to survive before anyone would put them in a real browser. It was slow and it looked academic and none of that mattered, because a testbed does not have to be pleasant. It has to compile.

~ the handover ~

By 1996 the W3C had decided that Amaya would be its testbed instead, and said openly that it did not have the resources to keep two browsers alive. What it did next was try to find somebody outside to take Arena. That worked, briefly. On 17 February 1997 Yggdrasil Computing picked up development, and the browser continued as a Linux community project for another year and a half.

The last release, version 0.3.62, came out on 25 November 1998. Nothing announced it and nothing replaced it, because by that point the thing Arena had been proving was already sitting in Internet Explorer and Netscape with somebody else's name on it.

~ what replaced it ~

Demonstration software dies quietly by design. Nobody has an account on it and nobody has files trapped in it, so when it stops there is no announcement to write. What survives is the argument it won, which is why the standards work of the mid-90s reaches us as a set of things that simply are the case rather than as a set of decisions somebody had to prove.

Arena is the reason CSS is not a paper standard that never shipped. That is a real thing to have done, and the price of doing it was being forgotten completely by everyone who benefited.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »