It's not just Twitter clients anymore. Basically they're constantly turning off pieces of the protocol, or so effectively limiting them that they're turned off.
For example, it's no longer legal to display a tweet unless you use their rendering.
If you had asked anyone at Twitter in 2009 if they would ever do that, they would have (probably) said no.
PG was right, then -- it had all the trappings of a new internet protocol, and therefore was almost a miracle. Except for the catch, which he also noticed -- that a private company could change the deal at any time. And they did.
I remember that leak very well as I was part of the MSFT team that had the 'deep infrastructure deal' meeting that is mentioned in that post. It was an interesting time for Twitter - Ev had to leave early because he was flying out. I learnt later that it was to appear on Oprah.
Just wondering...why is it wrong for a private company to take control of their product and user experience? It's their service, why can't they set certain protocols on how to use it?
I agree. The problem lies with how people are defining success. If you choose to define success as a "wild level of user and developer adoption thanks to the service being free, fun and bankrolled by investors with hundreds of millions of dollars", then yes, twitter is successful and owes a lot of this to the open api and ecosystem that grew up around it.
The problem is, this isn't success. Everyone (users, app developers and twitter itself) has been enjoying a free ride for the last x years, courtesy of investment money. It's been one big successful party... except for the fact it's been anything but for the investors. And when it comes to judging a company's success, the latter is the only kind that matters.
Taken in that light, it's hard to feel sorry for those who were caught out when the music stopped and the party was over. The investors are free to take the company in whatever direction they like in pursuit of real success (i.e. turning a profit) and if that means strangling the third party eco system, then so be it. Whether or not that's a good idea (I'm not sure it is) is irrelevant. It's their call to make.
Twitter proliferated on openness - they owe a large part of their success to it.
Part of me feels they have broken a promise with early users - now that they have their critical mass and revenue pressure they can turn their back on 'open'.
The free marketer in me would hope that a new open alternative would win over in the long term, but it doesn't look likely.
Regardless of whether it's wrong, I think it's fair to say that Twitter would not have been as successful without the help of an open protocol and third party clients. So it was a bit unfair to all the people who wrote clients and helped Twitter gain popularity to start closing down the protocol.
Take control of their product and user experience can equally be read destroy other companies and force customers to use their sub-standard interface. There is harm in it. It doesn't seem hugely unreasonable for people to object. Whether Twitter's actions are worthy of legal interference I don't know. And heck, maybe it wasn't even practical for them to keep supporting the protocol. It is on first blush, however, a bit of a dick move.
Of course, in practice, they can do it - but they're screwing over a lot of people in the process and it's unclear why they should be allowed to other than that they own it. And ownership doesn't make sense to me as a strong-form legal principle when you're talking about things that affect many people. To quote Lord Denning:
'None of you nowadays will remember the trouble we had - when I was called to the Bar - with exemption clauses. They were printed in small print on the back of tickets and order forms and invoices. They were contained in catalogues or timetables. They were held to be binding on any person who took them without objection. No one ever did object. He never read them or knew what was in them. No matter how unreasonable they were, he was bound. All this was done in the name of "freedom of contract." But the freedom was all on the side of the big concern which had the use of the printing press. No freedom for the little man who took the ticket or order form or invoice. The big concern said, "Take it or leave it." The little man had no option but to take it. The big concern could and did exempt itself from liability in its own interest without regard to the little man. It got away with it time after time. When the courts said to the big concern, "You must put it in clear words," the big concern had no hesitation in doing so. It knew well that the little man would never read the exemption clauses or understand them.'
That's kind of why we have laws concerning bargaining power and things like undue influence, and why contracts are - at least in principle - meant to be more or less equally beneficial. I believe you can trace it back to The Wealth of Nations, and probably before - it's not as if Denning had the idea originally. But in terms of illustration of the conclusion of letting people who own a service dictate how it will be used absolutely. The peak of freedom of contract as a guiding legal principle has - at least in all hope - passed.
Silly contracts still get written, mind, that's just taken for granted. I have a contract with my phone provider for a two year plan, for instance, but if it ever came to court I very much doubt that would be upheld - it's far more likely that the court would say that while companies can be reasonably expected to make plans two years in the future I, as an individual (and consequently being far less powerful) can't.
"It's their service, why can't they set certain protocols on how to use it?
reply"
It is, but startup owners should realize that if you base your entire business on something like Twitter, one change could destroy your business overnight.
I would say not exactly: If you have a public protocol and it's intended to run in multiple places or in a distributed manner, then even if the rest of the world stops using it, you and your users can happily chug along in your own sandbox.
If it's only used or useful in the context of a privately-owned/controlled environment... then you're in trouble, or likely to get there..
Google Reader ignored an important part of the RSS 2.0 spec, but I kept producing content that they couldn't parse, and if people used the tools that existed before GR, they didn't notice any difference, even though GR came to dominate RSS aggregators.
BTW, that's probably a reason why Twitter never could really support RSS, not because the spec didn't cover what they were doing, rather Google didn't.
I would say yes, but unfortunately most of those that took that gamble have been shut out. If Twitter were really smart about the whole thing, given the amount of money they have, they should let others develop off of their platforms and acquire them as they start to gain traction.
I was almost positive I had read the idea of Twitter as a protocol before (or something very similar), even though on Twitter[1] pg notes he never published this specific piece despite having written it in 2009. I really appreciated the simplicity of thinking of Twitter that way, and the definition has stuck with me.
In case anyone else wanted to know where they read it, it's under Request for Startups, #3[2].
Paul, curious if your enthusiasm for twitter has waned in light of the direction they have taken since? Obviously it is still an influential platform, but the grand vision we all had doesn't seem to be materializing...
My enthusiasm as a user hasn't waned. I use it increasingly often. But I'm not so enthusiastic about funding companies based on it. There are probably still opportunities for companies to grow by using Twitter initially, but it would be risky to start a company that depended on Twitter.
I wouldn't really call it a protocol. If Twitter is a protocol, then every exposed JSON-emitting API is a protocol - and perhaps that's true. But Twitter isn't unique in that respect. People have been publishing messages without specified recipients for a long time - it's called broadcasting, or publishing, or web publishing.
I wouldn't say the recipients are totally unspecified, either.
What makes Twitter unique is the artificial limit imposed on messages that can be broadcasted indiscriminately. And that makes it more of a new form of communication than any kind of protocol. It leads to a new form of expression, and a new way of getting a fairly accurate global view of the attitudes and cultural tides that are flowing through humanity in real time. That does make it truly earth shaking.
Because publishing a web page on any blog site lets you specify the recipients ?
To be a "protocol" imply a sort of "standard" shared , improved and used by different "actors" ( that's the reason for a protocol : two different people aggreing on a way to communicate). Twitter has always been centralized and closed in that sense.
Tweets do have recipients, which makes them different from a web page.
At the next level up from the one you describe it is very much a protocol.
Users get to decide whose tweets they follow and don't. And senders can block receivers if they want to. Twitter may have a very small something to say about that (spam, community standards) but for the most that's a free choice on the part of both the sender and the receiver.
Agreed. Twitter is in no sense a protocol. I think it became a great deal once people realized it is the only place for real time information flow, or at least that's how it's perceived among the Twitter users I know.
Yes, I thought this also. It's a protocol like email or IRC, except it's privately owned. And it's actually suited to being monopolised by a single company.
I also wondered if you could characterise something like Facebook the same way. If you strip away all the extra functions, you have a protocol for setting up a private network of friends, identified by their real names, who can then broadcast to each other. Other social networks didn't have quite the same setup - not private by default, not real names.
But Twitter is a protocol owned by a private company. That's even rarer.
They had an opportunity to be a free and open platform that spurns innovation instead they stifle it and impose ridiculous restrictions upon those who can only make the service better. In the beginning Twitter was a great platform to build on, but no doubt due to growing pressure to make money, Twitter have driven away a large number of developers, maybe Twitter will be fine without them but for how long?
Since Tumblr has supported multimedia content for a while, I'd say no. Twitter was (and is still mostly) text only. Protocols don't have to be text-only, but that's what makes Twitter feel like one, I guess? Just my two cents.
To my mind the fact that including pictures or video is less ad-hoc with Tumblr makes it more of a protocol, not less. Honestly Twitter doesn't feel particularly special to me - Facebook, Google+, even Livejournal feel like they're the same "kind of thing". Heck, at times Youtube behaves the same way. But I never got hugely into twitter, maybe it's different if you're a real fan.
>To my mind the fact that including pictures or video is less ad-hoc with Tumblr makes it more of a protocol,
I see what you mean, it's like Twitter but customizable to the core - true to any good protocol. What made twitter really a protocol was that so many clients were made for its liberal API.
Just think how much more enjoyable it would be if futurists kept their analyses and predictions to themselves until there was some actual evidence to support them though.
I predict that as a group they'd be much more respected.
Of course, we'd have to come up with a new name for them. "Histori...sts"
When is a repost justified if not when it comes true? It would still be justified if the prediction was spectacularly off, in which case we can all have a good laugh.
Makers of Twitter clients were eventually reminded of this fact in a rather painful way.