Microsoft Chromeffects

~ the eulogy ~

There were two names for one project. Chrome was the set of programming interfaces that let ordinary software reach DirectX, assembled from the Dimension X team Microsoft bought in May 1997 and engineers off the original Direct3D group. Chromeffects was the part the public was supposed to see: an XML wrapper that let a web page invoke all of it, the way Microsoft's own VML was supposed to handle flat artwork. Microsoft announced the software development kit on 21 July 1998, showed the technology at SIGGRAPH the same month, and put a developer release out in August.

The pitch was a web with depth in it. You wrote HTML and XML, with C++ or one of the script languages behind it, and Chromeffects composited 3D scenes and video on the client machine instead of shipping finished frames down the wire. That was the clever argument, because a scene description is tiny compared to the animation it produces, so the demos were supposed to look expensive and download cheap. The requirement underneath was less generous: Windows 98 with DirectX, plus a graphics card able to accelerate any of it.

Testers came back almost unanimously negative. The complaints were about speed and bugs rather than concept, which in a normal year would mean another version. 1998 was not a normal year for Microsoft multimedia, since the company was in court watching Apple and RealNetworks describe what happened to media companies in its orbit, and a proprietary replacement for a web standards process is a poor exhibit for the defense.

The decision came in a reorganization in November. Eric Engstrom, who ran the project and the multimedia group, was moved over to MSN as a general manager for web products. Deborah Black inherited what was left of Chromeffects alongside NetShow. The Register wrote it up on Friday 13 November 1998 with the company line attached: "We are stepping back and redesigning Chromeffects to better meet both our partner and customer needs." Nothing came back.

In 1998 I was fourteen, in Turkey, on a 33.6k modem that TTNet metered by the minute. A product whose minimum requirement was Windows 98 and a 3D accelerator was not aimed at the room I was sitting in, and the bandwidth argument that made Chromeffects sound smart to American developers would have been wasted anyway: the bottleneck on my line was not the size of the animation, it was that the connection dropped when somebody picked up the phone.

~ the dossier ~

~ the name ~

Ten years before Google shipped a browser called Chrome, Microsoft had a graphics layer called Chrome, and it died so quietly that the name was free when somebody else wanted it. Searching for the first one now is close to impossible, because every result belongs to the second one. The Wikipedia article on Microsoft Chrome exists mostly to tell you it is not the thing you were looking for.

~ where the pieces went ~

Nothing was thrown away, which is the usual fate of a cancelled Microsoft technology. The animation work turned up as DirectAnimation inside Internet Explorer. The Java-facing layer had already shipped separately as Liquid Motion, which got its own few years and its own death. The staff went back into the DirectX team, where the same engineers kept working on the thing that had been underneath all of it.

The one part nobody licensed was the business model. Chromeffects needed PC makers to put it on machines before anyone would author for it, and through the middle of 1998 not one of them took it. A web technology that requires a specific operating system release on the viewer's desk has to win that argument first, and this one never got past the first conversation.

@}-->--

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

« previous  ·  random  ·  next »