It’s a rare thing that I read a Doc Searls post and start shaking my head half way through. The recent $100 million investment in Twitter, bringing their valuation to $1 billion, has unleashed a torrent of criticism. Driving my daughter to school the other day, I heard an “analyst” on NPR chirp that Twitter couldn’t be worth $1 billion because it was just a fad, that people might stop using it tomorrow and the bottom would fall out. If using Twitter were a random activity that returned no value, I suppose that could be true. Just as people could decide to stop going to the movies, stop eating pizza or stop listening to “analysts” on NPR. If the value of something is disregarded at the start, it’s rather difficult to speak sincerely of valuation.
Searls’s criticism is a reprise of the open source silo meme. The drumbeat for the nationalization of Twitter has re-emerged. The capitalist pigs at Twitter have chosen to build a business rather than contribute their technology to the open source technology commons. Praise is sung for linux, rss, email and http. If only Twitter would see the light and release what they have to benefit the common good. Twitter’s business is just lumber from which other software developers should be allowed to create value. The complaint is that because Twitter is neither open nor decentralized, it has created an intractable engineering problem and does not contribute to the greater good of the web.
I would contend that Twitter is both open and distributed. Its characterization as a silo misses the point. Rather than using the silo as a criteria for openness, what if we look instead to the pipe. In the Unix command line, the standard output can be piped to the standard input of a new filter. Some very complex forms of processing can be created by chaining together a series of filters and piping data through it. The “chainability” of the javascript library Jquery is another good example of this model. The critique of the silo is its lack of interoperability, you can’t pipe to or from it.
Now, let’s look at Twitter. Can you pipe messages to Twitter? Can you pipe messages from Twitter? There was a time when I used Identi.ca as a primary micromessaging client. I typed messages into the Identi.ca web client and they entered the local pool, then I piped them to FriendFeed, where they also entered that ecosystem, FriendFeed sent them to Twitter, and Twitter sent them to Facebook. Examining this relay chain could you say that Twitter is a silo that owns my messages? Each of these venues represents a slightly different social graph and has a different tool set with which to display, prioritize and filter my messages within the context of the local graph. Twitter and Facebook are simply the most successful venues with which to read and write micro-messages (formerly called status messages). Google reader shares, SMS messages, Blog entries, et cetera can all be piped in and out of Twitter. Or if one prefers, Twitter can be left out of the chain entirely.
The mind share that Twitter and Facebook have built can’t be nationalized and distributed as lumber for a hypothetical socialist realist distributed micro-messaging ecosystem. If one is truly interested in open, look to the pipe, not the silo. Certainly there’s work that needs to be done on the pipe itself. Issues around real time, rate limiting, identity, social graphs, micro-communities, activity stream formats and track are all very important. But the real time stream environment is already here and operational. Many in the open source crowd are just rewinding the VCR and replaying the last battle. Steve Ballmer summed it up nicely in his interview with Mike Arrington, “we want to be first, best and interoperable.” Even Microsoft has embraced the pipe.
Your data is always already yours. It resides on your fingertips before they touch the keyboard.rnrn And you seem to make the point that the silo metaphor is leaking oil, and barely serves either its literal or metaphorical purpose in your rhetoric. Silos by definition are not interoperable. Turning “users” into “silos” is the ultimate contortion of a once stately metaphor.rnrnThere are no silos, think Network, nodes and the pipes that connect them.
Your data is always already yours. It resides on your fingertips before they touch the keyboard.rnrn And you seem to make the point that the silo metaphor is leaking oil, and barely serves either its literal or metaphorical purpose in your rhetoric. Silos by definition are not interoperable. Turning “users” into “silos” is the ultimate contortion of a once stately metaphor.rnrnThere are no silos, think Network, nodes and the pipes that connect them.
Your data is always already yours. It resides on your fingertips before they touch the keyboard.rnrn And you seem to make the point that the silo metaphor is leaking oil, and barely serves either its literal or metaphorical purpose in your rhetoric. Silos by definition are not interoperable. Turning “users” into “silos” is the ultimate contortion of a once stately metaphor.rnrnThere are no silos, think Network, nodes and the pipes that connect them.
Your data is always already yours. It resides on your fingertips before they touch the keyboard.rnrn And you seem to make the point that the silo metaphor is leaking oil, and barely serves either its literal or metaphorical purpose in your rhetoric. Silos by definition are not interoperable. Turning “users” into “silos” is the ultimate contortion of a once stately metaphor.rnrnThere are no silos, think Network, nodes and the pipes that connect them.
Your data is always already yours. It resides on your fingertips before they touch the keyboard.rnrn And you seem to make the point that the silo metaphor is leaking oil, and barely serves either its literal or metaphorical purpose in your rhetoric. Silos by definition are not interoperable. Turning “users” into “silos” is the ultimate contortion of a once stately metaphor.rnrnThere are no silos, think Network, nodes and the pipes that connect them.
Your data is always already yours. It resides on your fingertips before they touch the keyboard.rnrn And you seem to make the point that the silo metaphor is leaking oil, and barely serves either its literal or metaphorical purpose in your rhetoric. Silos by definition are not interoperable. Turning “users” into “silos” is the ultimate contortion of a once stately metaphor.rnrnThere are no silos, think Network, nodes and the pipes that connect them.
Your data is always already yours. It resides on your fingertips before they touch the keyboard.
And you seem to make the point that the silo metaphor is leaking oil, and barely serves either its literal or metaphorical purpose in your rhetoric. Silos by definition are not interoperable. Turning “users” into “silos” is the ultimate contortion of a once stately metaphor.
There are no silos, think Network, nodes and the pipes that connect them.
Hm, but piping my data and stuff around different silos doesn't make them less of a silo, it just makes them interoperable silos. Doesn't make me that much better off as a user, if user autonomy is what I am after
And as Doc points out elsewhere – a choice of silos is not a free choice.
But even practically, piping just isn't good enough – my data is still one step removed from me and I cannot control or analyse it the way I might want to. Open means that my data is mine and can be 'ported' by me (not the platforms de jour) to an application of my choice. If we are to have silos at all, let's make the users 'silos', i.e. autonomous platforms connected and networked on the web.
As for Twitter, I wish them well. The calls for their nationalisations are repulsive to me, their business plan is irrelevant to me and I wouldn't expect anything less than a very good exit strategy for them. They earned it, good luck to them. What I want is the identi.cas of this world to flourish…
[…] of silos vs. pipes, beginning with my post Values and Valuation, then continuing in Cliff’s The Silo & The Pipe: Doc Searls gets Venezuelan, and in the comments below that post. While I don’t wish to abandon the silo metaphor (or any […]
We're saying some of the same things in different ways. Microsoft failed in the attempt to replace the Network, and so had to join it. In a network, the potential for connection is critical. A silo is unconnected by definition.
Apple's music play isn't a silo, it's a successful business. There's no barrier to creating music players, stores or applications. Microsoft has done just that. But there's no requirement that once you are successful– you must give the store away. It's entirely possible that a direct distribution model, or a music micro-community model will disrupt the central store model in the next few years. But that's an issue for users to decide which solution is more valuable. Lala.com is a very compelling model. Amazon's MP3 downloads work perfectly with my iPod and iTunes.
Yes, interoperability — pipes are enough. Now, “pipe” politics look different than “silo” politics. Real time and latency is a big issue — does the pipe update in real time? Rate limiting is an issue — does a node limit traffic and why. Censorship, does a node censor any kind of traffic? How can identity be piped from node to node? How can micro-communities form across nodes? These are all political questions in the realm of the pipe — the connection.
The metaphor is important because it implies boundaries to the conceptual analysis. Looking to the pipe rather than the silo opens a new field of issues. And it recasts 'silo' issues in the context of the Network. In some cases, Microsoft is more interoperable than open source counterparts.
Twitter is a business that's trying to build sufficient scale to create a return on investment for their investors. Rather than ask 'at what cost?' — we might ask, 'at what benefit to the Network?'
Interesting that you mentioned Netscape and LDAP. One thing I've been pondering lately is the closed nature of the twitter directory. There is user info just floating out there (it comes with the tweet when you pull from the API) but there is no access to the user directory itself… Twitter really does appear to be trying very hard to be open with developers, but I wonder if this is their “ace in the hole”
What makes the Network “the silo that killed all other silos”? The Internet has its protocols and standards, but it is not by nature something that seeks to contain, for its own self-serving purposes, everything that depends on it.
Motivations matter. Microsoft failed to control the network, as Craig Burton pointed out long ago, in A Bulldozer Through the Intersection. But Microsoft wanted to control the Network. According to Craig (and bear in mind how long ago this was — April, 1996), Microsoft began losing that game when Netscape and friends adopted LDAP — an open directory standard. The story here, now almost forgotten, was vendor vs. vendor hockey at its finest, and one of the ways that Netscape's early work fighting Microsoft paid off later for Google, which in many ways is Netscape 2.0.
The larger motive, or both Netscape and Google, was growing the whole marketplace. They could still dominate that marketplace by doing better work, or by offering more goods and services. But they would not own that marketplace in the sense that they controlled it as a private domain.
The distinction here is between an open market space where many can participate and a closed market space that one company controls.
I don't think Twitter is being like the old Microsoft here. I think it is trying to be as interoperable as it can. But I think some of its moves, such as sticking with bit.ly and not working (far as I know, and I would welcome correction on this) toward an open standard for URL shortening, serve to control and contain the marketplace rather than to open and enlarge that marketplace — one which they already dominate, and which opening would not threaten.
I also don't buy the Beatles analogy. The Beatles were one band. They never controlled the music industry. Today, to a much greater degree, Apple does that. And they achieved that position by verticalizing a stack of dependencies. Today the vast majority of little music players are iPods (and now iPhones). These operate only with iTunes software. They interoperate with approximately nothing else. Want to buy music for your iPod? Go to Apple's store. Want applications for the iPhone? Apple's iTunes store, again. A large third party marketplace has grown around all of this, and it's all inside Apple's well-crafted vertical space. I don't think it's wrong, or antique, to call that a silo. In fact, I can't think of a better metaphor, and it has nothing to do with politics; nor is it obsolete open source lingo. It has to do with linguistics and the concept of the container. What is Apple's iTunes/iPod/iPhone contained market space most like? A stovepipe? A smokestack? A bottle? Call it what you will, a metaphor is unavoidable. Pick one.
Again, I don't think the Twitter folks are much like Steve Jobs & Co. But I'm sure many of their investors would like them to be. If so, at what cost?
Which brings me back to the question I asked earlier. Is interoperability enough? I say it's not.
I exhumed the moldy metaphor of the 'capitalist pig' intentionally. If you'll recall, in it's original context, one was either a 'capitalist pig' or a ' commie bastard.' Like the 'silo' it's a binary opposition that has outlived its usefulness.
If the Network is the computer, a silo is not part of the computer. Once MSFT lost its bid to become the Network (The silo that killed all other silos), it had to join the Network. Thus their new focus on 'first, best and interoperable.'
Interoperability is a question of the 'pipe' — the connection of one node to another. Oauth is the plumbing Twitter uses to pipe identity to this comment stream.
Don't confuse plumbing with a value proposition. As Steve Gillmor might say, Twitter is like the Beatles. Two guitars, bass and drums is plumbing, but they have every right to the music they make with those instruments. Twitter has taken the micro-message and the directed social graph and created something special. Technically, there's no barrier to entry, anyone can create a micro-messaging service. It's Twitter's mindshare that would be very difficult to overcome.
The Beatles cause a fail whale on Amazon because their Box Sets are sold out. Twitter's fail whales are similar — you can buy a different box set if you like, you're just not getting The Beatles.
Good points. I agree with many of them, especially around piping. There are rich metaphorical and technical lodes to mine there. But, while I'll agree to leveraging the “open source silo meme”, I object to being lumped in with a “drumbeat for the nationalization of Twitter” and accused (by inference, but clearly) of considering Twitter's folks “capitalist pigs”.
What I've been working toward, for many years, is recognition of the business-supporting qualities of open infrastructure. You're right that the pipe should be considered part of that, as should other structural conventions of digital plumbing. I'd like to explore that. But you're wrong to suggest that I advocate a “hypothetical socialist realist distributed micro-messaging ecosystem”. I don't.
By the way, it's interesting that Twitter is offered as a way to log myself in to DISQUS. Is that an example of piping? I suppose so. Also interesting that, while I've been part of the effort to establish OpenID as a (not the) standard for this kind of thing, using Twitter (for me at least, in this case) is much easier. My fault, I'm sure, for not remembering any of my many OpenID ID/login/password combinations.
What matters is that there are alternatives. Means for logging in are substitutable. Both Twitter and OpenID supply plumbing, much as a lumberyard supplies lumber.
The problem comes when the plumbing from Company A only works within its own walled garden.
So I grant that Twitter and Facebook interoperate. Are you saying that's enough?
Last thing… Below is a checkbox that allows me to “Tweet this comment as @dsearls”. Since I'm way over 140 characters I don't know what that means. But I'll check it to find out.