Dick Costolo recently spoke about Twitter as a platform on Charlie Rose show - "The future of Twitter is that we'll have a true platform, not just an API that allows developers to create an alternate Twitter experience, but an API that allows third parties to build on top of Twitter in a way that creates accretive value for the user, much how Amazon allowed third-party merchants to build into Amazon."
He made it abundantly clear that Twitter doesn't want its API to be used for alternative twitter clients. IFTTT does not necessarily create alternative twitter consumption client but it can be used to accomplish that.
On a side note (and it may not be popular with HN community); so far we have seen API used (majority of times) for alternative twitter consumption clients. May be with these API changes, we might see more innovation using Twitter as a "platform"? Rather than people trying to create alternative clients.
Well, IFTT's twitter hooks were exactly that kind of innovation, and it was just shot dead. How is anyone going to use it as a platform if you can't even read data from it?
Screen scraping is an interesting topic. From what I understand, a ToS is technically a legally binding contract, but if you send a screen scraper from a server that you are not physically at, then it negates it. Also, as I understand it, there is some discrepancy over whether a ToS is truly legally binding.
Well, actually breaking a site's ToS and benefiting materially from it is now considered a Federal crime. Aaron Swartz is being charged by the Feds for just that. It didn't involve screen scraping but use of a bot to download data, even though he was entitled to download that data normally, just not using a bot.
This is not true. Aaron Swartz is being charged for distributing non-free content that he happened to obtain that way. You make it sound like he's being charged with a felony for scraping free content.
Like last year’s original grand jury indictment on four felony counts, (.pdf) the superseding indictment (.pdf) unveiled Thursday accuses Swartz of evading MIT’s attempts to kick his laptop off the network while downloading millions of documents from JSTOR, a not-for-profit company that provides searchable, digitized copies of academic journals that are normally inaccessible to the public.
In essence, many of the charges stem from Swartz allegedly breaching the terms of service agreement for those using the research service.
The case tests the reach of the Computer Fraud and Abuse Act, which was passed in 1984 to enhance the government’s ability to prosecute hackers who accessed computers to steal information or to disrupt or destroy computer functionality.
The government, however, has interpreted the anti-hacking provisions to include activities such as violating a website’s terms of service or a company’s computer usage policy
Hmm, since the case hasn't been concluded yet, does that mean that that is a valid thing to charge a person with?
Also, the part you quoted ends with:
> "The government, however, has interpreted the anti-hacking provisions to include activities such as violating a website’s terms of service or a company’s computer usage policy, a position a federal appeals court in April said means “millions of unsuspecting individuals would find that they are engaging in criminal conduct.” The 9th U.S. Circuit Court of Appeals, in limiting reach of the CFAA, said that violations of employee contract agreements and websites’ terms of service were better left to civil lawsuits."
Also of interest:
> The rulings by the 9th Circuit cover the West, and not Massachusetts, meaning they are not binding in Swartz’ prosecution. The Obama administration has declined to appeal the ruling to the Supreme Court.
He's exaggerating. Theoretically they could send the feds after you, but in practice it will never be done. Screen scraping and botting websites is widespread in the industry and has even been used by several startups featured on Hacker News.
Actually, I think LinkedIn founder said in the interview that they (LinkedIn) was better off after the feature was turned off. Twitter chose to break the feature, but LinkedIn benefitted. Something about signal vs. noise...
It may have hurt in the case of IFTTT but if we assume the OP is right then this would simply be an edge case. Just because it hurt one 3rd party using the API doesn't mean all third parties will be prevented from innovating. Assuming Twitter is creating API usage rules to steer developers more toward building a platform rather than just replicating Twitter's functionality (which is what most 3rd parties are doing, the OP is right) its just not feasible or a productive use of time to verify that any given app isn't reproducing functinality (hence why they even have rules to begin with).
I totally get their concern about people making their own Twitter clients. But it's difficult to really push hard and tell the world you're trying to become a platform when the single most useful site for letting people connect your platform with other really cool stuff is disconnected because of your new policies.
Twitter seems to just have a leadership team with a complete lack of vision.
Tweetie actually did a great job of delivering relevant ads in a way that wasn't very obnoxious at all. Then Twitter bought the client, stripped out the ads, and started delivering really lame and irrelevant "promoted tweets"
The amount of cognitive dissonance is amazing. They need to ban 3rd party clients to protect monetization, yet they don't monetize their clients in a meaningful way. They want to be a platform, yet they do everything possible to make that effort fail and piss off potential partners.
All I see from them are expensive campaigns linking traditional media to whatever the twitter platform is supposed to be. I can't see how a business model that boils down to giving teenagers a way to communicate with Justin Bieber's assistant creates any value.
What does "a true platform" even mean? Currently my mind blanks when I try to think of any useful applications of Twitter, given that the only way to see tweets is on their web site now (effectively, I know there are some half assed clients still available for the time being).
Also betting on their API seems crazy now. Big companies or startups that spend other people's money might still be able to do it, but for a bootstrapper it seems like a recipe for disaster.
That's entirely the point. A "true platform" is, to Twitter, one where developers come up with new ideas that aren't obvious (hence, your mind being blank trying to come up with one on the spot), then Twitter copies / acquires them and extracts all of the value. It's the exact same thing that Facebook does on its platform (see Dalton Caldwell's experience at Facebook recently), and what OS vendors have done for decades.
I used to have a lot of ideas for Twitter apps, but now I am just confused. Not sure what is allowed and what isn't, and in any case, as you mention, whatever you would build would only have a very limited lifetime anyway. So it is not even fun to think about new ideas for Twitter anymore.
I wonder if that means Twitter want to roll out Facebook style "apps", so it has more control over things. At the moment it's not really a platform if you can't build anything that stores or displays tweets.
Twitter is realtime communication at it finest. They are the breaking news channel. Last year there was earthquake in east bay. First thing I did (even though it was middle of night) was to check my twitter when I sensed the vibrations. And I wasn't disappointed. Within seconds, I knew what was going on and how strong the earthquake was.
The amount of data they generate is humongous and can be very valuable. Data scientists can go through that data can do lot of pattern matching generate interesting insights. From how your product is doing in market or even complex predictions. That sort of stuff twitter wants developer community to build not yet another twitter client.
I think you are confused. What happened (and what this post is about) is they are preventing this kind of application to be built. You can't do pattern matching or whatever without access to the data...
> You can't do pattern matching or whatever without access to the data...
If you want "access to the data", you can get a firehose subscription from gnip.com or datasift.com
I once built an elaborate options pricer, only to find that my "access to the data" - which was basically a scraper from finance.google.com, was cut off. So I switched to finance.yahoo.com and soon enough they cut me off too! So I asked my professor and did what he told me - pay the subscription to IV ( http://www.ivolatility.com/data/historical_data.html ) and voila, the socket was instantly open and I could suck in as much data as I wanted.
It was 10 dollars for a month's access to a few MB of cleaned up csv historical option prices of a specific ticker.
And that's how resellers work - you pay exactly for what you want. Nobody's saying you have to pay 360000, like the other post implies.
Except that for Twitter you do. A few MB = 30 seconds of tweets. I don't know why you think financial feeds are related to twitter, other than both being data feeds.
The $360k price tag is not "implied", it's the actual price for access to 50% of twitter's messages (see the link I posted). 100% access can only be negotiated with Twitter itself. To be fair, from the information I could find, the cheapest plan from gnip is $500/month[1], datasift is $3000/month[2], both quickly go up. That means you can't just whip up a prototype in the weekend for some new idea anymore. This pricing probably filters out 99% of the things that could be built on it, limiting them to projects with investor backing and ready business models (which twitter itself still doesn't have).
I completely agree, there's a huge amount of data there that can be processed in a useful way. The problem is, at what point does something become a Twitter client? IFTTT let people use tweet data in a bunch of interesting ways, not just as another way of displaying things. If anything was using Twitter as a platform, IFTTT was.
I think you're wrong on that. Clients are the most visible use of the API for obvious reasons but I'd have to see data to convince me that it's disproportionate.
I think that is where Twitter's value degenerates. It's basically personalized RSS feeds with a smackload of users.
If you take away third party API capabilities for those RSS feeds in an effort to push people towards making "apps", you're going to lose developers who were in it for the functionality, not the community.
Comments
Dick Costolo recently spoke about Twitter as a platform on Charlie Rose show - "The future of Twitter is that we'll have a true platform, not just an API that allows developers to create an alternate Twitter experience, but an API that allows third parties to build on top of Twitter in a way that creates accretive value for the user, much how Amazon allowed third-party merchants to build into Amazon."
He made it abundantly clear that Twitter doesn't want its API to be used for alternative twitter clients. IFTTT does not necessarily create alternative twitter consumption client but it can be used to accomplish that.
On a side note (and it may not be popular with HN community); so far we have seen API used (majority of times) for alternative twitter consumption clients. May be with these API changes, we might see more innovation using Twitter as a "platform"? Rather than people trying to create alternative clients.
Well, IFTT's twitter hooks were exactly that kind of innovation, and it was just shot dead. How is anyone going to use it as a platform if you can't even read data from it?
The same way we read everything without an API, Pinky. Screen scraping.
screen scraping = break ToS = get arrested by Feds.
It is the new world kids and we all better break out our reinforced tin foil hats.
Screen scraping is an interesting topic. From what I understand, a ToS is technically a legally binding contract, but if you send a screen scraper from a server that you are not physically at, then it negates it. Also, as I understand it, there is some discrepancy over whether a ToS is truly legally binding.
Really? Would you get arrested by feds? Are you really breaking any federal laws? Isn't denying you access all they can do to you?
Well, actually breaking a site's ToS and benefiting materially from it is now considered a Federal crime. Aaron Swartz is being charged by the Feds for just that. It didn't involve screen scraping but use of a bot to download data, even though he was entitled to download that data normally, just not using a bot.
Scary stuff.
This is not true. Aaron Swartz is being charged for distributing non-free content that he happened to obtain that way. You make it sound like he's being charged with a felony for scraping free content.
Actually I am correct.
http://www.wired.com/threatlevel/2012/09/aaron-swartz-felony...
Some extracts:
Like last year’s original grand jury indictment on four felony counts, (.pdf) the superseding indictment (.pdf) unveiled Thursday accuses Swartz of evading MIT’s attempts to kick his laptop off the network while downloading millions of documents from JSTOR, a not-for-profit company that provides searchable, digitized copies of academic journals that are normally inaccessible to the public.
In essence, many of the charges stem from Swartz allegedly breaching the terms of service agreement for those using the research service.
The case tests the reach of the Computer Fraud and Abuse Act, which was passed in 1984 to enhance the government’s ability to prosecute hackers who accessed computers to steal information or to disrupt or destroy computer functionality.
The government, however, has interpreted the anti-hacking provisions to include activities such as violating a website’s terms of service or a company’s computer usage policy
Hmm, since the case hasn't been concluded yet, does that mean that that is a valid thing to charge a person with?
Also, the part you quoted ends with:
> "The government, however, has interpreted the anti-hacking provisions to include activities such as violating a website’s terms of service or a company’s computer usage policy, a position a federal appeals court in April said means “millions of unsuspecting individuals would find that they are engaging in criminal conduct.” The 9th U.S. Circuit Court of Appeals, in limiting reach of the CFAA, said that violations of employee contract agreements and websites’ terms of service were better left to civil lawsuits."
Also of interest:
> The rulings by the 9th Circuit cover the West, and not Massachusetts, meaning they are not binding in Swartz’ prosecution. The Obama administration has declined to appeal the ruling to the Supreme Court.
When you repeatedly change your MAC address to avoid efforts to block your screen scraper, maybe then you will have a problem.
They can't really know your MAC address anyway since they're more than a hop away from you, most likely.
That was more a reference to the Swartz case. But there too, jstor was more than one hop away.
Good point, forgot about that.
He's exaggerating. Theoretically they could send the feds after you, but in practice it will never be done. Screen scraping and botting websites is widespread in the industry and has even been used by several startups featured on Hacker News.
Unauthorized use of a computer, computer trespass... and it's probably over state lines.
Screen scraping != get arrested by feds. TOS' are not ratified laws, lol.
But IFTTT also allowed you to post tweets to your LinkedIn or Tumblr account (as far as I know). So twitter didn't like it.
Twitter already has LinkedIn integration ...
Twitter pulled this a few months ago http://thenextweb.com/socialmedia/2012/09/19/linkedin-referr...
Was not aware of this. There goes that useful feature.
Actually, I think LinkedIn founder said in the interview that they (LinkedIn) was better off after the feature was turned off. Twitter chose to break the feature, but LinkedIn benefitted. Something about signal vs. noise...
Infact linkedin is lot better these days
It may have hurt in the case of IFTTT but if we assume the OP is right then this would simply be an edge case. Just because it hurt one 3rd party using the API doesn't mean all third parties will be prevented from innovating. Assuming Twitter is creating API usage rules to steer developers more toward building a platform rather than just replicating Twitter's functionality (which is what most 3rd parties are doing, the OP is right) its just not feasible or a productive use of time to verify that any given app isn't reproducing functinality (hence why they even have rules to begin with).
I totally get their concern about people making their own Twitter clients. But it's difficult to really push hard and tell the world you're trying to become a platform when the single most useful site for letting people connect your platform with other really cool stuff is disconnected because of your new policies.
Twitter seems to just have a leadership team with a complete lack of vision.
Tweetie actually did a great job of delivering relevant ads in a way that wasn't very obnoxious at all. Then Twitter bought the client, stripped out the ads, and started delivering really lame and irrelevant "promoted tweets"
The amount of cognitive dissonance is amazing. They need to ban 3rd party clients to protect monetization, yet they don't monetize their clients in a meaningful way. They want to be a platform, yet they do everything possible to make that effort fail and piss off potential partners.
All I see from them are expensive campaigns linking traditional media to whatever the twitter platform is supposed to be. I can't see how a business model that boils down to giving teenagers a way to communicate with Justin Bieber's assistant creates any value.
What does "a true platform" even mean? Currently my mind blanks when I try to think of any useful applications of Twitter, given that the only way to see tweets is on their web site now (effectively, I know there are some half assed clients still available for the time being).
Also betting on their API seems crazy now. Big companies or startups that spend other people's money might still be able to do it, but for a bootstrapper it seems like a recipe for disaster.
That's entirely the point. A "true platform" is, to Twitter, one where developers come up with new ideas that aren't obvious (hence, your mind being blank trying to come up with one on the spot), then Twitter copies / acquires them and extracts all of the value. It's the exact same thing that Facebook does on its platform (see Dalton Caldwell's experience at Facebook recently), and what OS vendors have done for decades.
I used to have a lot of ideas for Twitter apps, but now I am just confused. Not sure what is allowed and what isn't, and in any case, as you mention, whatever you would build would only have a very limited lifetime anyway. So it is not even fun to think about new ideas for Twitter anymore.
I wonder if that means Twitter want to roll out Facebook style "apps", so it has more control over things. At the moment it's not really a platform if you can't build anything that stores or displays tweets.
Twitter is realtime communication at it finest. They are the breaking news channel. Last year there was earthquake in east bay. First thing I did (even though it was middle of night) was to check my twitter when I sensed the vibrations. And I wasn't disappointed. Within seconds, I knew what was going on and how strong the earthquake was.
The amount of data they generate is humongous and can be very valuable. Data scientists can go through that data can do lot of pattern matching generate interesting insights. From how your product is doing in market or even complex predictions. That sort of stuff twitter wants developer community to build not yet another twitter client.
I think you are confused. What happened (and what this post is about) is they are preventing this kind of application to be built. You can't do pattern matching or whatever without access to the data...
> You can't do pattern matching or whatever without access to the data...
If you want "access to the data", you can get a firehose subscription from gnip.com or datasift.com I once built an elaborate options pricer, only to find that my "access to the data" - which was basically a scraper from finance.google.com, was cut off. So I switched to finance.yahoo.com and soon enough they cut me off too! So I asked my professor and did what he told me - pay the subscription to IV ( http://www.ivolatility.com/data/historical_data.html ) and voila, the socket was instantly open and I could suck in as much data as I wanted.
So, basically, pay or shut up? Access to 50% of Twitter messages is gonna cost you only $360,000 per year. That will certainly foster innovation.
http://www.readwriteweb.com/archives/twitter_to_sell_50_of_a...
Do you mind saying roughly how much the IV subscription costs? All I could find was "email for quote".
It was 10 dollars for a month's access to a few MB of cleaned up csv historical option prices of a specific ticker. And that's how resellers work - you pay exactly for what you want. Nobody's saying you have to pay 360000, like the other post implies.
Except that for Twitter you do. A few MB = 30 seconds of tweets. I don't know why you think financial feeds are related to twitter, other than both being data feeds.
The $360k price tag is not "implied", it's the actual price for access to 50% of twitter's messages (see the link I posted). 100% access can only be negotiated with Twitter itself. To be fair, from the information I could find, the cheapest plan from gnip is $500/month[1], datasift is $3000/month[2], both quickly go up. That means you can't just whip up a prototype in the weekend for some new idea anymore. This pricing probably filters out 99% of the things that could be built on it, limiting them to projects with investor backing and ready business models (which twitter itself still doesn't have).
[1] http://web.appstorm.net/how-to/social-media-how-to/gnip-the-...
[2] http://datasift.com/pricing#costs
I completely agree, there's a huge amount of data there that can be processed in a useful way. The problem is, at what point does something become a Twitter client? IFTTT let people use tweet data in a bunch of interesting ways, not just as another way of displaying things. If anything was using Twitter as a platform, IFTTT was.
I think you're wrong on that. Clients are the most visible use of the API for obvious reasons but I'd have to see data to convince me that it's disproportionate.
I think that is where Twitter's value degenerates. It's basically personalized RSS feeds with a smackload of users.
If you take away third party API capabilities for those RSS feeds in an effort to push people towards making "apps", you're going to lose developers who were in it for the functionality, not the community.
A "true platform" doesn't change with management's whims. No dev should trust Twitter's APIs.