Home > Archive >  2009 >  December >  22

Open is in the eye of the beholder

Tuesday, December 22, 2009 by Dave Winer.

When I say the Twitter API may be an open standard, I mean something different than when Jonathan Rosenberg says Google likes open standards.  

I mean it's open in that anyone can implement it now.  

A smart developer can implement the Twitter API in a matter of weeks.  

Rosenberg means that the process of defining the standard is open. He would start a process to define a standard that in two or three years a team of 20 programmers could implement in another two or three years. Those are the kind of results that his version of "open" delivers. 

In an ideal world I prefer defacto standards that are ratified by unanimous consent by standards bodies. The deliberative process yields bad results for several reasons: 

1. The defacto standard doesn't just shrivel up and die, despite the best efforts of the latecomers to force it to. Twitter isn't going to shut off its API and kill its developer base because a working group formed. 

2. You end up with two ways of doing something, and two ways are worse than one, no matter how much better the new way is. As Anil Dash says, the Twitter API is now, for all practical purposes, finished.  

3. The new way is never better because the deliberative process yields a design by committee. SOAP failed to be a source of interop for just this reason. The working group had hundreds of members and pleasing even some of them meant creating a spec that was so vague and rambling as to be pointless as a standard. It provided almost no interop. (BTW, I am an author of the first version of SOAP, so I got to watch the deconstruction of a standard by a working group from the inside.) 

Dare Obasanjo of Microsoft asks us to look at the process by which RSS begat Atom as a cautionary tale. He says the designers of Atom had "no choice" but to be totally incompatible with RSS, but that's not true. They could have started with RSS, and only deviated where absolutely necessary. I urged them to do this. 

All this talk about open standards processes is surfacing now because WordPress and Tumblr implemented the Twitter API, thereby putting the "defacto" ball in play. The BigCo's don't like it because, like WordPress and Tumblr, they're late to the game and don't want to play on the same field as the little companies. They want to force Twitter into a standards process, I assume. It's something Twitter might consider, but they should be cautious.  

Microsoft welcomed the W3C to the standards process when Netscape threatened the dominance of the Windows desktop in the mid-90s. The W3C gave them a greater-than-equal voice to Netscape, even though they had nowhere near Netscape's market presence in the dawning market of the web.  

In 2009, Google is just beginning to be a presence in realtime. It's fairly thrilling to see realtime search results show up in Google searches. They should keep going but hold back some of their power to let the little companies have a chance to do what they do better -- drive innovation. If Google does what Microsoft did, and unleashes all their destructive power, they will destroy the market. And we will be left with a scorched battlefield instead of the growth that's promised by the blossoming prototype of the news system of the future that I believe Twitter is. 

I've said all along, as I said with Netscape in 1994 -- Twitter desperately needs competition to toughen them up and make them more responsive to market opportunities. To give them a sense of urgency they lack. But what they and we don't need is Google and other big companies to stall the market in the name of being open. Their process is open only if you're a BigCo, and shuts out exactly the people we want in there. The gutsy bright-eyed young entrepreneurial minds at rising stars like WordPress and Tumblr. They are ready for a standard now, not someday in the future, after a huge working group is finished with it. Not two years after that when everyone has forgotten the Open Microblogging Intitiative or whatever it ends up being called. They're ready now, and so is the market.  

So Google, Microsoft, Facebook, et al should step back and consider how they can help now, and not throw obstacles in the path of a surging market. 

My argument to Microsoft 15 years ago was that they would get the lion's share of any growth that is achieved by the upstarts, so it worked against their interests to slow them down. That still is true today, but there are new players who own a large share of the future, no matter what. Don't screw it up.