Amaya

» see w3.org/Amaya as it lived, on the wayback machine

~ the goodbye ~

Amaya's fifteen years line up almost exactly with my own time online, and I never once opened it. That is not an accident of taste. Amaya was built for people who edit the web, and in 1999, on a 56k line in Turkey, I was firmly a person who read it.

The W3C started it in July 1996 with INRIA, the French research institute, as the successor to Arena in the role of house testbed. Where Arena had existed to prove that a draft specification could be rendered at all, Amaya was given a larger job: it had to browse and it had to edit, in the same window, with the same document in memory. You opened a page, put the cursor in it, and typed.

It was ahead on the things standards people care about. Amaya rendered MathML and SVG when the browsers with actual users rendered neither, and it carried working CSS at a point when that was still an argument rather than a fact. RDF and XPointer went in later, along with the ordinary comforts of caret navigation and a spell checker. It was also, by every account including the W3C's own, slow and strange to look at, because a reference implementation is graded on being right rather than on being liked.

The odd part is that the thing Amaya wanted did happen to me, just not through Amaya. By the mid 2000s I was writing to the web constantly, in signature boxes on Turkish hardware forums where you listed your components and argued about them. That is an ordinary person editing a live page. It arrived through a text field on somebody else's server rather than through a program on my machine, and the difference between those two routes is most of what this grave is about.

~ the chronicle ~

~ what it was for ~

The argument underneath Amaya is older than Amaya. The first web browser, the one Tim Berners-Lee wrote on a NeXT machine in 1990, was an editor as well. You could change a page you were looking at. When Mosaic made a read-only client the normal shape of the thing in 1993, that capability fell out of the web almost by accident, and the W3C spent years treating this as a mistake worth correcting rather than as a decision that had been made.

Amaya was the correction. It is a strange sort of product, one built to demonstrate that the split between reading and writing was never necessary. Fifteen years of releases is a long time to keep making that point.

~ the decline ~

Two things went wrong and neither of them was technical. The first is that a reference implementation cannot win a browser war it is not entering. Amaya had no marketing budget and no bundling deal, and it competed for attention against products with hundreds of millions of users apiece.

The second is worse for the premise. The editing problem got solved on the other side of the wire. Wikis and blog forms put the writing surface inside the browser you already had, so nobody needed a special client to edit the web. By 2005 the read-write web was a live argument again, and every implementation of it had shipped as a web page rather than as an application.

~ what came after ~

Version 11.4.4 went out on 18 January 2012. The last change to the code was on 22 February 2013. The W3C's page for the project now says, without ceremony, that it is no more developed, which is the plainest death notice in this entire inventory.

Amaya lost the implementation and won the premise. Every text box you type into on a page, every field that saves back to the document you are looking at, is the read-write web that Amaya spent fifteen years insisting on. It got there by a route Amaya did not build and would not have chosen. Being right about the shape of something is not the same as being the thing.

@}-->--

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

« previous  ·  random  ·  next »