Habla Livehelp

Tom Lawrence

staff:

Developers: OAuth support
For increased security and more resilient third-party apps, we’ve started early testing of OAuth support in the Tumblr API.
We’ve modeled our implementation after Twitter’s in many ways, supporting OAuth 1.0a with optional xAuth, upon request, where it makes sense such as mobile and native applications.
The existing authentication methods in the API are still supported for now, but we encourage developers to migrate to OAuth when possible.
To get started with Tumblr’s OAuth, integrate an OAuth consumer library into your project and get a consumer key by registering your app.
This is still in its early testing stages, so please report any issues. Thanks.

I’ve not written any software that implements OAuth yet so I’ll definitely give this one a whirl with some of my own Tumblr apps soon. This sure beats the current method for an authenticated read using POST requests with custom headers on an insecure connection passing along fellow user’s login details in plain text (I was surprised that you guys never implemented TLS/SSL).
It’s also great to see OAuth libraries available for so many platforms and APIs. Hell, there’s even an OAuth library for the .NET Compact Framework. Not that anyone actually cares about Windows Mobile but still, choice is indeed nice.
Next thing on your list of things to sort out in the API is replies! Seriously, they’re one of the most used features on Tumblr and devs can’t even implement them. Second on your list is Message/Ask; another highly popular feature that even the iPhone optimised page doesn’t feature. Finally, for the complete package, allow direct video uploads to Tumblr as opposed to using embed or data (that just uploads to Vimeo anyway).

staff:

Developers: OAuth support

For increased security and more resilient third-party apps, we’ve started early testing of OAuth support in the Tumblr API.

We’ve modeled our implementation after Twitter’s in many ways, supporting OAuth 1.0a with optional xAuth, upon request, where it makes sense such as mobile and native applications.

The existing authentication methods in the API are still supported for now, but we encourage developers to migrate to OAuth when possible.

To get started with Tumblr’s OAuth, integrate an OAuth consumer library into your project and get a consumer key by registering your app.

This is still in its early testing stages, so please report any issues. Thanks.

I’ve not written any software that implements OAuth yet so I’ll definitely give this one a whirl with some of my own Tumblr apps soon. This sure beats the current method for an authenticated read using POST requests with custom headers on an insecure connection passing along fellow user’s login details in plain text (I was surprised that you guys never implemented TLS/SSL).

It’s also great to see OAuth libraries available for so many platforms and APIs. Hell, there’s even an OAuth library for the .NET Compact Framework. Not that anyone actually cares about Windows Mobile but still, choice is indeed nice.

Next thing on your list of things to sort out in the API is replies! Seriously, they’re one of the most used features on Tumblr and devs can’t even implement them. Second on your list is Message/Ask; another highly popular feature that even the iPhone optimised page doesn’t feature. Finally, for the complete package, allow direct video uploads to Tumblr as opposed to using embed or data (that just uploads to Vimeo anyway).

Opera Mini 5 for iPhone

So can’t wait for this! Opera beats the crap out of Safari on both mobile and desktop platforms.

The app has already been submitted to Apple for review to be placed on the App Store and the folks at Opera are giving away a free iPhone 3GS for the person who guesses the closest time to when Apple approve the app. You can check that out here.

And as always, I heartily encourage people to try out the desktop version if they haven’t done so before (especially the Windows users out there, Mac users I’d probably wait for the 10.50 final release, although a beta version is available).

Remember when the Millenium Bug was the biggest threat coming?

jonmtm:

Since then:

  • We’ve been in a perpetual war in the middle east
  • Created a fear about a religion based on the actions of some of its extremists
  • Turned 18, turned 21, turned 24
  • Lived dreams
  • Said goodbye to the Concorde
  • Said hello to Flight of the Concords
  • Seen an increase in ukulele popularity
  • Seen bands split and reform
  • Seen the end of the boy band stronghold, and increase in singer-songwriters
  • Seen the rise of social networking websites
  • Been given MySpace, Facebook, Bebo, Twitter, Tumblr, Hi5, and more
  • Been given the gift of video hosting online, including Vimeo, YouTube, and all the porn star equivelents
  • I guess broadband internet speeds are faster
  • Were given iPods, iPhones, and other smart phones
  • Had the first black President
  • Seen an insanely high roster of celebrity deaths
  • Watched Disney Stars fall to pieces (see Britney Spears, and maybe that chick from High School Musical who’s nudes were all over the place)
  • Justin Timberlake
  • Watched Fame Acadamy, Pop Idol, American Idol and X-Factor run around making a mess
  • Had the tragedy that is Big Brother
  • Saw the follow ups to The Matrix and Star Wars Prequels and left feeling molested
  • Had Lord of the Rings
  • Learned what alcohol is BAD for you (read: none)
  • Created new mistakes, and didn’t regret
  • Listened to Millenium by Robbie Williams fourhundredthousandtimes
  • And used Last.fm to scrobbl that.

And really, what was the Millenium Bug going to do? Reset all our calendars and clocks to thinking it was the 1900s? Is that the biggest thing we had to fear back then? ‘Oh crap, I’ll have an appointment booked for Tuesday but I’ll be 100 years late for it oh no’.

I can’t believe its nearly 2010, you guys. Seriously, that is insane.

Interesting list, I have yet to type up some posts regarding the year and the decade myself but I’ve reblogged this as a response to your comment on the Millennium bug as it was theoretically a serious issue.

Warning: Geek Out Session Below

Many date systems focused on the last two digits of the year (those being ‘99’ and ‘00’ respectfully). Now, to a human, this is quite obvious that the century has rolled over and we’re now in the year 2000. However, to a computer this meant various things. For example, in the C programming language, the date/time function in the standard library would pass back the value -1900. The problem here is that a HELL of a lot of software out there is programmed in C (including software found in banks) and as a programmer, it would seem obvious to store a year in a variable as an integer - a whole number that can only be positive.

So if a programmer had not inserted code for correct validation and exception handling, this could cause a whole bunch of data catastrophes (ESPECIALLY if we’re talking about banking software) and we’re only talking about ONE programming language here. Any software based on a programming language that focused only on the last two digits could produce any type of result (not just ‘00’ as you would expect), depending on how the language was designed. Even logic systems found in electronics could completely fail with inverted bit patterns, causing many electronic devices to just stop working.

It sounds absolutely stupid in many respects, but it got many programmers worrying about how they had actually implemented date and time into their software and how those systems would cause other elements of their software to react.

Imagine if a nuclear facility’s monitoring software was to crash and perform improper exception handling due to a radiation logging system using two digit dates. It’s something most people didn’t want to know what the result would be.

Thankfully, with funding from various sources into patching and correcting software to either support the use of 4-digit dates or insert the correct validation or exception handling code, there were very few incidents that occurred due to what many claim as the so-called “Y2K bug”. It was somewhat pot luck for most as the majority of programmers just thought to themselves “Shit, exactly HOW did I implement date and time into that piece of software for that particular company? Is that my problem? Should I charge them to correct the program?”.

One good thing to come from the situation is we now know the importance of validation and exception handling for ALL systems in software, even the basic implementations, and storing our year variables as four-digit integers :p

Sorry for the total geek out post there, but it was something I felt the urge to discuss! Hope you enjoyed the read.