A happy “accident”

Andy Plesser and the folks at Beet.tv do a lot of great interviews with people in technology… if they’re not on your reading list, I’d recommend adding them. But today they had an interview with Adobe’s Jen Taylor, which ended up on Techmeme with the title “The Vidoe Revolution Happened by ‘Accident'”. News to me! 😉

If you watch the interview, at about 1:20, Jen says “It’s funny, I don’t think we ever anticipated the success we see today with online video. I often tell the story of how video got into the Flash Player, and it was a complete accident… I don’t think we ever understood what we had seeded [?] in that.”

True enough — Jen was saying that we hadn’t foreseen the explosion of digital video on the web, where everyone expects parallel over-the-air and over-the-web viewing of all types of realtime experiences, where people routinely upload videos of their own for people the world over to watch. That part was indeed a happy surprise.

But the capability itself wasn’t an accident. Jon Gay and Robert Tatsumi brought FutureSplash to Macromedia, and stayed on the combined Player/Authoring team for a few years. Jon took a sabbatical, and came back with a vision of two-way video communications, all implemented by a single tiny codec inside the Player.

To get an idea on the emphasis on two-way video communication back at that time, check out Tim Anderson’s interview with Jeremy Allaire, back in April 2002, during the introduction of the term “Rich Internet Applications”:

“We’re introducing a new technology for communications applications… The Flash player that was recently released includes within it all the client capabilities needed for these communications applications. It allows you to deliver real-time, peer-to-peer or one-to-many or many-to-many real-time communications applications, with shared data, audio and video. All that is built into the client today. We’ll be introducing a new communications server later this year. It’s codenamed Tin Can, and it allows you to build these communications applications… .. You can do real-time shared audio and video.”

I spoke with Paul Betlem in the kitchen here on Townsend St today, and he said that his big memories from the time were the emphasis on two-way communication, the integration with a webpage instead of being a branded video “player” (like those from Real, Microsoft, Apple), and the concern that the addition of a codec, however small, might slow consumer adoption rates. It was a controversial decision internally — a premeditated gamble — and we all sort of held our breath on how it would turn out.

To get an idea of how advanced Flash developers saw it at the time, check out these notes from Mike Chambers, of a July 2002 session by Danny Mavromatis and Mike Davidson of ESPN.com. While Flash video may not have had all the features and options of existing video architectures, it had some unique, no-hassle, audience-inclusive benefits which helped bring about the situation we see today.

It took a few years for people to understand that video was now much easier to deliver… back in 2006 I noted that Flash video was a “voice in the wilderness” for its first few years. I think it was a few far-sighted Flash-savvy content developers who really proved to the world what could be done.

But I don’t think anyone at Macromedia had a clear vision of how user-generated content would take off, and how there would soon be giant video sites and live-streaming of public events and such — the technology was a very deliberate decision, and it was the global popularity which was the “happy accident”.

2 Responses to A happy “accident”

  1. Eric Wittman says:

    Thought I’d add some “old timer” comments here. As the Flash Product Manager for Flash 3 and 4 and the Director who ran Flash Product Management and Marketing for Flash 5 and MX (6), I can assure everyone here that implementation of Flash video was not an accident, rather a calculated product strategy. Without revealing too much secret sauce, here’s what we did:
    Even during the Flash 3 days, we wanted to find ways of getting a lightweight video codec into Flash Player to do basic video playback. Shockwave had high quality video playback at the time and while not wanting to compete with big brother, we did want to enable some video playback initially to support user interface design use cases customers were requesting. The team didn’t have a lot of video experience so the initial path was to partner with pure play video players, Apple (QuickTime) and Real Networks (Real Player), to embed Flash Player into their respective players. While this didn’t exactly fit the desired use case, customers like Disney and MSN did use Flash for video UI. This additionally allowed the team to become more familiar with video playback on the Web (which at the time was nascent and horrible) and equally important from an strategy execution standpoint, get video authoring functionality implemented into the development tool. For those Flash historians out there, folks will remember we could import video files into Flash authoring and export those back out as QuickTime or Real files.
    During the Flash 4/5 timeframe, we still felt that the solution we were providing wasn’t meeting the full range of customer demands so began the quest to find a video codec that was good enough in quality but small enough in footprint to avoid bloating Flash Player. Eventually through a partnership and co-development work with Sorenson, we found our codec which made its way into Flash Player 6 (officially called Sorenson Spark).
    It was during the Flash MX development period that Jon Gay took a sabbatical to clear his mind after a long time of hard work on Flash. One of the results of his hiatus was the precursor to the Tin Can project, initially released as Flash Communications Server, now Flash Media Server. Jon came back from sabbatical, plunked down a PC tower with this custom rolled software, and showed a live web cam playing video back into Flash Player. Jon had a vision of taking the video capabilities in Flash to the next level for not only static video playback but also dynamic video and communications. Quickly the implementation in Flash Player 6 of Sorenson Spark grew to not only include the decoder, but the encoder to enable real-time video exchange and support Jon’s communications vision and the Tin Can project.
    While the initial quality of Flash video wasn’t as good as others, it gained instant popularity because “it just worked” for consumers. Over time, the codecs and quality got better to the point today were full-screen, HD quality content is being delivered to hundreds of millions of people over the Internet. While indeed not envisioning the massive adoption seen today, Flash video was a calculated part of the product strategy and one that has paid off handsomely for Adobe, companies building solutions around it, and for the web audiences as a whole.

  2. Sarah Allen says:

    I’ll have to agree with Eric that there was very little accidental about Flash video. I was part of the “TinCan” team in 2000 which added two-way audio-video to Flash Player 6 and created what is now called the Flash Media Server.
    This post quotes a bunch of marketing at the time that was designed to drive server sales, highlighting the new features which required the new server product.
    Remember that in 2000, when the Flash Media Server development began, all evidence pointed to there being no market for streaming video servers. Microsoft was giving away Windows Media Servers and Real Media announced that they would open source their server that year. The business model that drove the initiative was around server sales, so it needed to do more than just replace Windows and Real Media servers. Of course, Macromedia was in the business of selling software… with the Flash Player free, the company was seeking ways to get into the server business.
    It seemed obvious at the time that adding good video playback to the ubiquitous and free Flash Player would change the landscape, but it wasn’t clear that alone would significantly drive an increase in revenue. I wouldn’t expect Macromedia (now Adobe) has seen much profit from YouTube.
    Also, personally I had the feeling that if we played up the importance of (one-way) streaming video, we would not get the chance to do the really exciting innovative two-way features, so I didn’t work too hard to make the case for it 🙂