ThunderHawk

~ the goodbye ~

Bitstream was a type foundry. It had been drawing and selling typefaces out of Cambridge, Massachusetts since 1981, and if you have ever run a Linux desktop you have read something set in Bitstream Vera without knowing whose it was. Around the turn of the century the company decided that the problem with the mobile web was a rendering problem, which from a type company is not a strange conclusion at all.

They announced the technology on 6 June 2001, put out a beta that October, and shipped ThunderHawk in December. On 20 May 2002 the Pocket PC edition arrived and was marketed as the first full-featured web browser for that whole class of device. Full-featured meant the real thing: HTML 4.01 and CSS, rendered properly, on a screen measured in a couple of hundred pixels.

The browser went to version 2.1 with Java support in 2005 and reached 3.00 in 2008. After that there is nothing. No shutdown post, no farewell page, just a product that stops appearing in release notes and a website that eventually stops answering. Bitstream was already building its replacement by then.

~ the chronicle ~

~ the trick ~

The client on your phone did almost no work. It opened a connection to Bitstream's servers, which fetched the page you asked for and rendered it at full desktop width with real fonts. What came down to the phone was a compressed picture of that rendering, in a transport format Bitstream had designed for the job. The visible part of the page came first and the rest arrived in the background while you were already reading.

What you got was a split screen: an overview of the whole layout up top and a magnified strip underneath for the bit you were actually reading. Everybody else in 2002 was reflowing pages into a single narrow column or serving you a WAP deck with nine links on it. ThunderHawk kept the page's shape and shrank it, on the theory that you already knew what the site looked like and were trying to find your way around a map you recognized.

~ who it was really for ~

Mobile operators and handset makers, mostly, with a side door for individuals who had a Pocket PC and $49.95 a year to spend on browsing. Sonera in Finland put the beta into a pilot program. That is the customer shape of a company selling rendering technology, and it is also why almost nobody reading this ever saw it.

Turkey in 2002 is the other half of that answer. I was in internet cafes playing Counter-Strike and checking MSN, at rates per hour that made a yearly browser subscription look like a joke, and the phones around me were sending text messages. Mobile data here was WAP priced by the kilobyte, on devices that had no business loading a desktop layout. A product that assumed you owned a Pocket PC and paid a US subscription in dollars was never going to turn up in that room.

~ the second attempt ~

Bitstream kept the engine and rebuilt the product around it. Bolt went into private beta on 15 January 2009 and public beta at Mobile World Congress a month later, same server-side rendering, same compression lineage, except this time it was free and aimed at ordinary people with ordinary phones. It got a real audience and lasted until the end of 2011.

Both deaths have the same cause written on them. Server-side rendering was a brilliant answer to handsets that could not render, and handsets stopped being that. The iPhone shipped a browser in 2007 that drew the actual page on the actual device, and after that the only thing Bitstream's farm of servers was saving you was bandwidth.

@}-->--

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

« previous  ·  random  ·  next »