Adobe Flash

» see adobe.com/flashplayer as it lived, on the wayback machine

~ the goodbye ~

flash started in 1996 as a vector animation tool called futuresplash. macromedia bought it within months and renamed it flash. the technology let you author small animations and interactive content that would play in a browser plugin. by 2000 the plugin was installed on essentially every desktop computer in the world. flash had won the web animation market by default.

adobe bought macromedia in december 2005 for 3.4 billion dollars, mostly for flash. the next five years were flash's peak. youtube initially used flash for video playback. browser games were almost all flash. animated banner ads were almost all flash. interactive product demos were almost all flash.

the iphone shipped in 2007 with no flash support. apple decided not to support flash on ios. steve jobs wrote a famous open letter in april 2010 explaining why: flash was buggy, insecure, battery-draining, and built for desktop computers with mice. the letter was a strategic move but it was also mostly correct. the iphone's growth slowly turned the lack of flash from "apple's problem" into "flash's problem."

adobe announced flash's end of life on 25 july 2017, with a final shutdown date of 31 december 2020. browsers (chrome, firefox, edge, safari) all dropped flash support on or near that date. on 12 january 2021 adobe's runtime actively blocked all remaining flash content from playing. the plugin was dead.

~ the chronicle ~

~ what flash actually was ~

flash was several things at once. it was an authoring tool you used to create animations. it was a file format (.swf) that the animations were saved in. it was a browser plugin that played those.swf files. it was a programming language (actionscript) that let you make the animations interactive.

the combination was unusual. most web technologies were one of these things. HTML was a format. javascript was a language. browser engines played them. flash bundled the authoring tool, the language, the format, and the runtime into one closed system. that integration was both flash's strength and its eventual weakness.

flash files were small (vector graphics compressed efficiently), played smoothly (the runtime was optimized), and could be authored visually (the macromedia flash IDE was the design tool of choice for an entire profession of flash animators). for the kind of interactive web content that was popular from 2000 to 2008, flash was clearly the best tool.

~ the steve jobs letter ~

in april 2010, steve jobs published an open letter on apple's website titled "thoughts on flash." the letter explained why apple's mobile devices would not support flash. the arguments were technical:

• flash was a closed proprietary system, while the modern web was moving to open standards
• flash had bad battery life on mobile devices
• flash had security vulnerabilities, with many CVEs each year
• flash was designed for mouse interaction, not touch
• HTML5 video and canvas could replace flash for most use cases

each of these arguments was largely correct. the letter was strategically motivated (apple did not want adobe controlling a major content runtime on iOS) but the technical reasoning was sound. the broader industry slowly accepted apple's framing over the next five years.

adobe did not respond well. they argued for flash's importance, then announced flash for android (which was buggy and slow), then quietly removed it from android in 2012. the strategic battle had been lost by then.

~ the security problem ~

flash had a real security problem. the runtime was complex. the language was dynamic. flash files could be crafted to exploit bugs in the runtime, leading to remote code execution on the user's computer. these exploits were used for years by malware authors as a primary distribution vector.

adobe shipped flash security patches almost monthly. browsers eventually implemented "click to play" features that required users to explicitly enable flash before each piece of content played. the constant patching, the blocking by default, and the obvious user friction made it harder for content to depend on flash.

by 2017 most browsers had moved flash to a permission-required model. by 2018 the only places that still used flash were specific enterprise applications, old educational tools, and the rapidly-shrinking flash games archive at sites like newgrounds.

~ how it died ~

adobe announced the end on 25 july 2017. the announcement gave 3.5 years of notice. the actual death date was 31 december 2020. on that date, flash player stopped accepting new content from any source. browsers blocked it. adobe distributed a "kill switch" runtime that actively prevented existing flash content from playing.

the kill switch was important. adobe did not just stop developing flash. they actively pushed an update that would make existing flash files unplayable. this was meant to reduce the security risk from old flash content sitting on disks and webpages everywhere. it also meant that anyone who had flash files they wanted to keep playing had to keep an old flash runtime around or use an emulator.

~ what we lost ~

the interactive web of 2002 to 2010. an enormous archive of flash content (games, animations, interactive presentations, art projects) became unplayable in early 2021. ruffle (the open-source flash emulator written in rust) has restored playability for some of this content, but coverage is partial. genuine cultural artifacts from the flash era are now harder to access than they were in 2010.

the homebrew animator and game developer ecosystem. flash was the tool of choice for individual creators making animation and games as hobby projects. when flash died, this ecosystem fragmented. animators moved to after effects (commercial, expensive). game makers moved to unity (capable but heavy) or html5 (powerful but harder). nobody since has replaced the specific friction-low approachable flash authoring environment.

the case study in proprietary platform decline. flash shows what happens to a proprietary technology that briefly won a market and then lost when an open alternative arrived. the lesson is that proprietary platforms have a half-life. the more successful they get, the more incentive there is for open replacements to emerge. the open replacements eventually win because they fit the architecture of the wider system better.

@}-->--

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

« previous  ·  random  ·  next »