My platform is called the Internet.<p>I'm still subject to some rules, but a heck of a lot less than the gardens you guys develop in like Facebook, Twitter and AOL.<p>I included AOL, because while it seems ridiculous someone would do that today, people used to until they got horrifically burned, and I think it'll be the same with non-decentralized platforms like FB and Twitter in the future.<p>Once again, my platform is the Internet.
I would see this as an opportunity. You have an app that has hit a hard limit in terms of number of users. It is clearly a useful app, otherwise it wouldnt be growing.<p>The natural thing to do is to make it a paid solution. You inform your users of the situation and say that in order to keep the service up and running you will start charging a signup fee of $1, the same as you would pay for a mobile app. Existing users however, get 50% off and only pays $.50. Give them a month to paypal it in and then another month of repeated notices to inactive users. After two months there will probably 50k+ users you will have to close out from the service. But give them an option to restore it easily.<p>This is not at all unreasonable, no one would ever be able to say so. Above all it would be a tremendous learning experience for you. Perhaps you also will make a few thousand dollars in the process, but that would not be the point.<p>There are two alternatives to this as I see it. Either you do the above but the only thing you require is for users to manually report in by clicking a button in a form. You will still be able to weed out thousands of accounts.<p>The other is to set up Smart Tweets 2, hosted on another ip, and refer new users there. Explain why and make it into something funny for the users.<p>Whatever you do, do something bold enough to make it to Techcrunch a second time.
They own the data. It is their ball and they can basically say "I'm taking my ball and going home".<p>Facebook, twitter, linkedin, google. All of them are in the data business and sell that data to make a profit. They don't really care about the devs as they just see the devs as a way to bring people to their ecosystem; the more small apps people write, the more ways they might be able to get data. But, if they get big enough, and all of them have, they can cut you out.<p>I understand building a business around someone else's data (cottage industry), and I would never say not to do it, but it isn't without peril. Though, I would do it in a heartbeat for a lifestyle business that I knew I could pivot on or build another if it failed.
Obviously Twitter does not want you to copy "their" data to Facebook. They don't get any money when you do that, and they want money. Hence, a problem.<p>There are several ways to get around this. Be a middleman that publishes to both Facebook and Twitter (a reverse FriendFeed). Or, sell a software product for the user to use to move his own tweets to Facebook. A user obviously owns his own content and can put it wherever he wants. A third party app, perhaps not.<p>(Also, why not just get tweets via whatever method the native UI gets tweets from? Do it from AWS if you are concerned about an IP ban.)
This story rings true for any popular app that depends on a third-party service, not just Twitter - often times you're at the mercy of the host once you become big enough to be noticed.
I've been following the thread and I commented on Michael's site, but thought it would be worth sharing on here as well: <a href="http://hoisie.com/post/a_warning_to_hackers_be_careful_building_on_twitters_api#comment-125739196" rel="nofollow">http://hoisie.com/post/a_warning_to_hackers_be_careful_build...</a><p>Since most of you probably don't know me, I'm director of the platform at Twitter.<p>Let me know any questions/comments you might have. I'm interested in an open discussion about it.<p>Ryan / @rsarver
As I realized after Twitter's "developer" conference last year, the era of the Twitter hacker is coming to a close (see <a href="http://twitter.com/#!/dacort/status/12005978721" rel="nofollow">http://twitter.com/#!/dacort/status/12005978721</a> and <a href="http://twitter.com/#!/dacort/status/12032959629" rel="nofollow">http://twitter.com/#!/dacort/status/12032959629</a>). If you are not building a business on their platform where money will pass into <i>their</i> business, good luck.<p>And who can blame them, providing 100,000 user tweets isn't free. Not sure where he got the idea he would never have to start paying.<p>Turning off basic auth also hastened this demise - whereas you could once pull whatever data you wanted from Twitter using a simple curl command, now you have to figure out OAuth. It's not that much of a challenge, but it is enough of a barrier to entry to dissuade somebody who's got a couple hours on the weekend and wants to have some fun.<p>Finally, their partnership with Gnip is yet another indicator that this is simply not the days of the wild west anymore. You want the data, you'll have to pay.
Of course you're at their mercy. This is the nature of using a third-party API. If you violate their rules, you have to accept the risk that they shut you down in the future, even if their rules aren't optimal for your app's user experience.<p>I recommend that you alter your product to conform to their rules, even if it makes your product a bit worse.
The way around API limitations on how much data can be read is to resort to scraping. However, Twitter probably blocks individual IPs that access too much data. The solution to that is to convince enough users to install software that lets you access their website - preferably via forwarded SSL so your users can't compromise data integrity. Users get some reward, presumably quite small, for relaying the requests for you.<p>The scheme could be opened to provide unofficial paid APIs for Twitter and other 'walled gardens' that make data available to unauthenticated users on the Internet.
With Microsoft monopoly you always had the opportunity to reverse engineering the OS but when you can't see the binaries because they are in the cloud you're in trouble, it's worse than closed source.<p>Building your business around web apis without an SLA is the most risky business, you don't have control,<p>Enjoy your 15 minutes of your application placebo fame!
I've had a whole litany of issues with Twitter doing stuff like this... but ultimately I've found work-arounds. It just depends how committed you are and how much time you have to make the fixes! More info here: <a href="http://pluggio.com/blog/" rel="nofollow">http://pluggio.com/blog/</a>
Crap, I have an app that was build long, long before display guidelines existed (heck even hash tags didn't exist yet!). No time or energy to fix it.<p>And honestly, that style guide is ridiculous. Way to alienate your earliest champions, Twitter.
Microblogging sites in China are thriving, much better than SNS sites (IMHO). I often come to wonder why FB is worth so much more than twitter. Could this actually be one of the cases when copycats win over the original?
I had an even worse issue (entirely my own fault) which highlights the dangers of building on another platform. My app (tweetbars.com) had a tiny flaw in that it didn't time out the Curl calls to the Twitter API, and my host didn't kill hanging php executions.<p>So when Twitter started hanging (and eventually timing out after a minute or so) the app basically ate one of their shared servers and the host took my whole account down for 24 hours.
John Kalucki from Twitter's platform team responded in the comments with what sounds like a reasonable alternative: <a href="http://hoisie.com/post/a_warning_to_hackers_be_careful_building_on_twitters_api#comment-125554983" rel="nofollow">http://hoisie.com/post/a_warning_to_hackers_be_careful_build...</a>
One should normally be careful of any freely available service - they owe you nothing. If you really want to partner with twitter for a cool app, approach them about signing an official agreement or something, no?
with regards to the lack of approval, could it be that Twitter want to avoid app redundancy to save resources? 100,000 of these requests can't be cheap for them, so when there are already multiple apps do what this one does does it not make sense for them to say they're no longer supporting most to save money?<p>I could be way off the mark, but that's how it appears to me. Multiple apps with the same purpose that require lots of resources... makes sense to stick with one or two high users and limit the rest.
the author has some points, but he does not mention something obvious to the reader: He was always at the mercy of Twitter. It was his fortune that the cap was increased at his whim, and only now have they stopped it, for whatever reason.<p>The reason sounds fishy, and his retort is well-founded, but it's an important thing to note.