Microsoft XNA

» see msdn.microsoft.com/xna as it lived, on the wayback machine

~ the obit ~

the death notice was an email to unpaid volunteers, and it leaked inside a day. at the end of january 2013 microsoft wrote to the people holding mvp awards in directx and xna to tell them the category was being retired from the award program on 1 april 2014. the reasons given in the mail were that directx was no longer evolving as a technology and that xna game studio was not in active development. the story broke on 1 february 2013 and microsoft spent the following week explaining that it was in fact actively investing in directx and that xna remained a supported toolset. sitting in the same statement was a line about there being no plans for future versions, which is the only part of it that turned out to matter.

xna stood for xna's not acronymed, which tells you what the culture around it was like. underneath the joke it was a managed layer over directx written for c#, and it took care of the parts of game programming that stop people from finishing games. the content pipeline turned your art files into something the runtime could load. sprite batching and controller input arrived already written. you got a game loop for free and could spend your evenings on the game instead of on the plumbing.

the split that defined it was money. writing xna games for windows cost nothing at all. putting one on an xbox 360 meant joining the creator's club, later the app hub, at ninety-nine dollars a year or forty-nine for four months, and that membership was the only door onto the console. microsoft made the toolchain free and charged for the shelf, which is the reverse of how development tools had been sold for the previous twenty years.

game studio express landed free in august 2006, the year i was a computer engineering student and c# was what the department put in front of you. a free compiler with a real debugger behind it mattered more then than it sounds now, because everything else in that category had a price in dollars. the part that did not travel was the annual fee. ninety-nine dollars is a small number in the country that sets it and a different number everywhere else, and it came round again every year whether or not one person had bought the thing you made.

~ the rap sheet ~

~ how it landed ~

xna was announced at the game developers conference in san jose on 24 march 2004, two years before anybody could touch it. the first community preview of xna build came in march 2006 and game studio express followed on 30 august. microsoft built a whole recruitment apparatus around it: dream build play, a global contest with seventy-five thousand dollars attached, opened in january 2007, and dreamspark handed students a free year of premium membership so they could put their coursework on a console.

the versions came steadily for five years. game studio 2.0 on 13 december 2007, then 3.0 on 30 october 2008 adding the zune, then 4.0 on 16 september 2010 adding windows phone 7. the 4.0 refresh of 6 october 2011 was the last one ever shipped. by the time the mvp mail leaked, the product had already been silent for fifteen months and everyone using it knew.

~ the day it died ~

xbox live indie games was built on xna and could not exist without it, so the framework's death sentence was the storefront's as well, even though the two events are four years apart on paper. what actually ended in 2013 was the promise. microsoft had spent seven years telling hobbyists that c# was a legitimate way to make a console game, and then withdrew the claim in a mail to volunteers rather than in a blog post with a name signed under it.

the walk-back made the position worse rather than better. a supported toolset with no planned versions is a product being kept alive on paper, and no studio starts a two-year project on one. the clarification told everyone what the leak had told them, in longer words.

~ what replaced it ~

mono.xna started as a port of the api onto the open source mono runtime, and monogame grew out of that codebase together with silversprite. fna forked from monogame with a narrower purpose written into its own documentation: preserve xna by reimplementing xna. the api microsoft stopped maintaining now has two independent implementations that outlive it, which is a fate normally reserved for languages rather than for game frameworks.

the practical result is that xna games kept shipping after xna died. terraria was built on it. bastion and fez came off it. stardew valley was written against xna 4.0 and released in february 2016, three years after the framework was declared finished, and only moved to monogame later for the ports. the thing that killed xna was that microsoft stopped caring about it, and that turned out to be survivable.

@}-->--

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

« previous  ·  random  ·  next »