Seriously? Quit complaining! Developers have known about this for months and the date has always been in June 2010 (as far as I'm aware). Any problems occurring because of the switch are 100% because of the developer's poor planning.
Or lack of interest; from the posting: "When Twitter breaks all the apps in the OAuthcalypse, they will break all of mine, and I have no intention of fixing them."
I have to say that I got a bad impression of Dave Winer late last decade when he posted a private email of mine on his blog without asking. And in our back and forth he mentioned how he was quite confused about people being upset by his habit of doing this....
ADDED: I wasn't too upset, except that he wasn't cooperative about my desire to edit it for public consumption.
What was the email about? Do you have a pointer? Why didn't you want it published? Did you say in the email it was private? Was it in response to something public? Was it abusive?
It was about the broader issues illuminated by an AGIS/Conxion peering conflict that prevented a number of people from accessing his web site in October 1997.
I didn't want it published "as is" since it wasn't written for public consumption. There were personal details that detracted from the message, adjustments I wanted to tone down some of the rhetoric and some people (sources) who I referred to who needed better credit and/or real links.
I didn't say it was private, but in some of it I was "talking" directly to him. And it was certainly netiquette at the time that email was by default private; he said that was "crap", that he'd gone over this issue "countless times", that "off the record" was a privilege that had to be negotiated ahead of time and that he was "being very generous" in offering to delete it.
That message from him left a rather bad taste in my mouth.
It wasn't abusive, although it was harsh on some bad players of the era, like UUNET, e.g. their actions prompted the editor of Boardwatch to commission a cover depicting the head of UUNET planting blue barrels of ANFO in MAE-East and I think MAE-West, referring to the OK City bombing a couple of years earlier.
The general issue of the day as I put it in my email:
"[T]he ISPs that are refusing to peer with others on equal terms are basically no longer offering Internet connectivity, but are instead offering a private network that happens to be connected to parts of the Internet. As I like to describe it, their unique selling proposition is "sign up with us, and we'll connect you to a steadily smaller portion of the Internet"."
UUNET's dominance at that time allowed them to play this sort of game and we still see it occasionally when a low cost provider irritates another. The end result of these power plays as I said at the time was "paying customers of the disconnecting ISP demand full connectivity or take their business elsewhere", although obviously that was a lot easier in the dialup era.
Actually they telegraphed the change long before that, but the problem is that OAuth changed in the interim and is now a moving target. So any developer that got on board and implemented OAuth is getting burned twice.
Um, Twitter has been up for years. I have two old apps, and Twitter is needlessly making me put more work into them. It's especially painful because I don't even remember the code very well anymore. The Twitter library I use in one of them is not actively supported anymore, so it might be a huge hassle to switch to another library.
Don't know, it is the first time I experience this, to be honest. Do APIs really change that much? I have experience with new features being added, which you'd preferably use in your future projects.
I am not super-complaining, just saying I see a reason for complaints. Especially since the update seems to make thinks more complicated without any real benefit.
This change provides an added benefit of cleaning up the Twitter app ecosystem. Twitter is changing, and if an app developer won't make this change they probably wouldn't ever be inspired to update their app to use new features^, and users would go elsewhere (in this case sooner rather than later).
^ Single-use apps--visualizations come to mind--deserve a pass. They do one neat thing and don't necessarily need to grow and evolve outside their niche. I'm sure one app or another will be missed.
Ah, the iPhone argument. What if an app simply doesn't need constant updates and a recurring flood of new features? What are we, a society of update junkies? Some apps are simply there to do something useful, not for showing off new features.
That's why RSS is frozen. No developer should (or can) code against a moving
target. RSS has been in the same place for a long time and look at all that has
developed around it. If we kept changing our minds about how it worked,
eventually it would have amounted to nothing. But no one had the power to make
those changes, despite how much people complained -- so it stayed put.
This is what works.
Lies, bullshit, projection, and blatant hypocrisy is par for the course with Dave Winer. He repeatedly edited the 'frozen' spec documents incompatibly without changing the version number. A good introduction: http://diveintomark.org/archives/2004/02/04/incompatible-rss
He's one of the worst spec writers of all time, responsible for other epic fuckups like OPML, XML-RPC, and SOAP.
Since when have web platforms ever not been moving targets? It's a pretty simple equation really, if people see enough value with being part of the twitter platform they will update their code, if not, well twitter of decided that it doesn't bother them to lose those people by making this change.
The twitpocalypse came and went, and developers updated there apps to use 64-bit ids for tweets. It was a necessary growing pain that the API had to go through. This is no different.
All Twitter apps? Certainly not mine because I, like most, actually read the big notice that was on the api wiki for most of '09 saying basic auth would disappear soon.
It is mindsets like his that destroy and stagnate langauges, and fill them with bloat to support outdated legacy code.
I don't see why anyone should be truly content with a platform or environment that refuses to evolve, progress will obviously cause breaking changes, but in the far majority of cases, the resulting fix will improve the existing application.
If someone is too lazy to take the chance to improve an application of their own, then they truly don't care about it in the first place, but then again if they are statisfied with the stagnation of their entire platofrm and environment, it makes sense that they would be statisfied with the stagnation of their own programs as well.
Software evolves, poor programmers complain and resist, good programmers go with the flow, but great programmers embrace this.
I'm surprised no one has mentioned his freakout over JSON, yet (http://www.scripting.com/stories/2010/04/26/theToxicCoralRee...). Apparently, Twitter turning off XML output and standardizing on JSON is "even more dramatic" than the Basic Auth-to-OAuth transition.
Frankly, I don't get it. Dave's argument appears to be that because the tooling he uses (Radio Userland?) doesn't support JSON, and therefore he will stop writing tools for Twitter (he wrote Twitter-related tools?).
He's the only person I've ever read who seems to prefer XML to JSON. Is there anyone else out there who feels the same way? What's the perceived advantage of XML over JSON? Can anyone shed some light on this?
I'm not going to defend Winer's point of view, but I also prefer XML for some tasks. Namespaces make XML extensible in a way that JSON is not.
In XML, I can add an attribute or a child node anywhere without affecting anything else. There is no easy way to extend JSON without jumping through ugly hoops to recreate a more XML-like data model.
I think Dave is confused about the difference between a protocol like RSS with many producers and consumers (very hard to change, just look at SMTP!), and a service API like Twitter's with a single producer (it changes, some apps die that were unsupported already, big deal).
"None will make it through this transition without being reconceived."
Say what? I can't imagine many apps will require reconcieving because an authorization API changed.
21 comments
[ 69.6 ms ] story [ 486 ms ] threadI have to say that I got a bad impression of Dave Winer late last decade when he posted a private email of mine on his blog without asking. And in our back and forth he mentioned how he was quite confused about people being upset by his habit of doing this....
ADDED: I wasn't too upset, except that he wasn't cooperative about my desire to edit it for public consumption.
I didn't want it published "as is" since it wasn't written for public consumption. There were personal details that detracted from the message, adjustments I wanted to tone down some of the rhetoric and some people (sources) who I referred to who needed better credit and/or real links.
I didn't say it was private, but in some of it I was "talking" directly to him. And it was certainly netiquette at the time that email was by default private; he said that was "crap", that he'd gone over this issue "countless times", that "off the record" was a privilege that had to be negotiated ahead of time and that he was "being very generous" in offering to delete it.
That message from him left a rather bad taste in my mouth.
My email was in response this public posting on his blog, start with the 4th paragraph of this page: http://www.scripting.com/1997/10/16.html
It wasn't abusive, although it was harsh on some bad players of the era, like UUNET, e.g. their actions prompted the editor of Boardwatch to commission a cover depicting the head of UUNET planting blue barrels of ANFO in MAE-East and I think MAE-West, referring to the OK City bombing a couple of years earlier.
The general issue of the day as I put it in my email:
"[T]he ISPs that are refusing to peer with others on equal terms are basically no longer offering Internet connectivity, but are instead offering a private network that happens to be connected to parts of the Internet. As I like to describe it, their unique selling proposition is "sign up with us, and we'll connect you to a steadily smaller portion of the Internet"."
UUNET's dominance at that time allowed them to play this sort of game and we still see it occasionally when a low cost provider irritates another. The end result of these power plays as I said at the time was "paying customers of the disconnecting ISP demand full connectivity or take their business elsewhere", although obviously that was a lot easier in the dialup era.
Doable, but annoying.
Platforms evolve. APIs get deprecated. Stuff either breaks or gets updated.
I am not super-complaining, just saying I see a reason for complaints. Especially since the update seems to make thinks more complicated without any real benefit.
^ Single-use apps--visualizations come to mind--deserve a pass. They do one neat thing and don't necessarily need to grow and evolve outside their niche. I'm sure one app or another will be missed.
He's one of the worst spec writers of all time, responsible for other epic fuckups like OPML, XML-RPC, and SOAP.
I don't see why anyone should be truly content with a platform or environment that refuses to evolve, progress will obviously cause breaking changes, but in the far majority of cases, the resulting fix will improve the existing application.
If someone is too lazy to take the chance to improve an application of their own, then they truly don't care about it in the first place, but then again if they are statisfied with the stagnation of their entire platofrm and environment, it makes sense that they would be statisfied with the stagnation of their own programs as well.
Software evolves, poor programmers complain and resist, good programmers go with the flow, but great programmers embrace this.
Frankly, I don't get it. Dave's argument appears to be that because the tooling he uses (Radio Userland?) doesn't support JSON, and therefore he will stop writing tools for Twitter (he wrote Twitter-related tools?).
He's the only person I've ever read who seems to prefer XML to JSON. Is there anyone else out there who feels the same way? What's the perceived advantage of XML over JSON? Can anyone shed some light on this?
In XML, I can add an attribute or a child node anywhere without affecting anything else. There is no easy way to extend JSON without jumping through ugly hoops to recreate a more XML-like data model.
"None will make it through this transition without being reconceived."
Say what? I can't imagine many apps will require reconcieving because an authorization API changed.