Microsoft Liquid Motion

~ the eulogy ~

liquid motion sat in the part of the late nineties web i could read about but not personally remember using. in 1998 i was 14, my home internet was a 33.6k modem on ttnet, and the web pages i opened were the donanimhaber forum and a handful of mp3 search results. the kind of person who would have used liquid motion as a 1998 web designer was building a corporate homepage on a t1 connection in seattle. i was watching a yellow notepad icon spin on netscape 4 while a homepage hit-counter gif finished downloading.

microsoft liquid motion was an authoring tool for web animations that compiled its output into java applets. it shipped in 1998 on top of technology microsoft had acquired in may 1997 from a san francisco company called dimension x, which had built one of the first java multimedia toolkits. the pitch on the box was simple. flash needed a plugin in every browser. liquid motion would run on the standard java runtime that already sat inside ie and netscape. you would author once, the applet would render anywhere.

by 1998 macromedia flash had already won the web animation argument. flash 3 was out and the player install base was past 50 percent of web users. the production tooling around flash had a head start of years. liquid motion shipped a 1.0 with web productions and pro editions. it never picked up a designer base, and microsoft discontinued the product in 2000. the replacement was a tool called vizact that animated html content inside the page directly instead of stuffing the animation inside an applet box.

~ the dossier ~

~ how it landed ~

the dimension x acquisition closed in may 1997 for an undisclosed amount. dimension x had been the first java-only startup. founded in march 1995, it had shipped two products. liquid motion was a java animation authoring tool. liquid reality was a java-based vrml browser. microsoft kept the liquid motion piece and killed liquid reality. the team folded into the windows web team. the relaunch under the microsoft brand happened in 1998 with the new tooling integrated with frontpage 98 as the front-end html editor.

the press treatment was lukewarm. trade reviews pointed out that liquid motion files were heavier than flash files for the same animation. they took longer to start up because the jvm had to spin first, and the java rendering pipeline did not match flash's bezier curves so the output looked slightly worse. the cross-browser portability argument was undercut by the fact that the microsoft jvm and the sun jvm did not always agree on rendering details. you could in principle ship an applet that rendered cleanly in ie 4 and broke in netscape 4.5.

~ the java animation problem ~

java in the browser had a structural disadvantage that liquid motion could not work around. the applet model required a full jvm cold start on first load of a page that had a java tag in it. the start-up cost was several seconds on a 1998 home pc. flash startup cost was a fraction of that because the flash player was a tight c codebase tuned for one job. liquid motion authors could ship beautiful clips and watch them die on first impression because the user closed the browser tab before the applet loaded.

this part i can read about now but did not see at the time. by 1999 the conversation in web design magazines had already concluded that flash had won. microsoft's own properties were quietly building in flash too. the people who would have been liquid motion's natural audience had moved on. the microsoft team that supported the tool spent more time replying to support tickets about jvm version mismatches than shipping new features.

~ what replaced it ~

microsoft killed liquid motion in 2000. the official successor was vizact 2000, a tool that animated html and css inside the page instead of compiling to an applet. vizact lasted about a year before microsoft cancelled that line too in 2001. the deeper strategy continued in a different shape with silverlight in 2007, microsoft's third attempt at a non-flash rich client runtime for the browser. silverlight made it further than liquid motion did and died for the same reason a decade later, undercut by html5 and the iphone's plugin ban.

none of the liquid motion files survive as a live format. the java applet model was deprecated by oracle in 2017 and removed from browsers shortly after. any.jam or.lmo file from 1998 that did survive needs a 1998-era browser and a working java 1.1 jvm to render. the dimension x team mostly scattered. the company itself is a wikipedia footnote now. the first java startup, absorbed and digested in three years.

@}-->--

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

« previous  ·  random  ·  next »