‘In pursuit of a more secure walled garden’

Upcoming API Change Will Break Facebook and Instagram oEmbed Links Across the Web Beginning October 24

Sarah Gooding, writing for WordPress Tavern:

In an extraordinarily inconvenient API change, Facebook and Instagram will be dropping unauthenticated oEmbed support on October 24, breaking content across millions of websites. The change will force users to generate an app ID with a developer account in order to continue embedding links via oEmbed

In 2008, Leah Culver, one of the collaborators on the oEmbed spec, said it was created to be “an open web API standard for fetching embed code based on a URL.” Requiring authentication in order to use oEmbed links seems like a violation of its intended purpose. For more than a decade, oEmbed has made it possible for users to easily share media across websites and social networks, without having to touch any code. It underpins a flourishing, connected landscape of web sharing that opens up new audiences for posts that might otherwise be buried in a social network’s fast-moving timeline.

In pursuit of a more secure walled garden, Facebook will now require all publishers to obtain developer app credentials in order to embed content that was previously available through simple URLs. Many users will be understandably frustrated when they find they can no longer embed Facebook and Instagram links the way they could in the past. Some will not be motivated to surmount the hurdle of setting up a Facebook app and may resort to posting screenshots or omitting the content altogether. A feature so widely used by non-technical users should not be suddenly locked away behind developer credentials.

A big middle finger from Facebook to the Open Web, in other words.

In response to Facebook’s API change, WordPress will be removing Facebook as an oEmbed provider in an upcoming core release. This will break a lot of content — many years’ worth of posts in some instances, and will require users to install a fallback plugin. WordPress plugin developer Ayesh Karunaratne has created a new plugin called oEmbed Plus that brings back support for Facebook and Instagram content embedding. It guides users through the process of setting up Facebook developer app credentials.

For those who are using the Gutenberg plugin, the Facebook and Instagram blocks have been removed as part of tomorrow’s version 9.0 release. oEmbed links will continue to work until Facebook’s API change goes into effect.

In all seriousness, anyone who uses embedded Facebook content or connects to Facebook services from their websites should consider whether doing so is morally and ethically justifiable.

Whatever Happened to CNET?

Om Malik:

As someone who has witnessed the meteoric rise of CNET and its role in technology media from the very beginning, I was shocked to learn that they are being sold for $500 million to Red Ventures, a North Carolina-based media investor. 

After all, CBS had acquired CNET for $1.8 billion in 2008 — but then I remembered that CBS already sold off the China operations. And they milked the property nicely after a turnaround by former CEO Jim Lanzone and his team. What’s left is finally getting a new home, and at a price much higher price than traditional private equity buyers were willing to pay. 

Wow, that’s a name I’ve not thought about for a long while. CNET — and ZDNet, which it acquired in the 2000s — was a big part of my web browsing, and later podcast listening, habits for a long time.

Sadly, as Om points out in his piece, Red Ventures are more interested in the reviews and referrals side of CNET, and that makes me wonder what’ll happen to what’s left of the journalistic side of their operations. Here’s hoping those folks find new gigs.

I wonder what the future will hold for the other parts of CBS Interactive, particularly Last.fm ?

What’s so hard about PDF text extraction? ​

As it turns out, it can be very hard indeed, depending on how the PDF was created.

I have a love-hate relationship with PDF files. While I love them as an archival format, not all PDFs are made equal. When I was doing print pre-press work, for instance, text could get mangled depending on whether we had the matching fonts available.

Son of Clippy?

Mike Stone has some thoughts on the late, not-much-lamented Clippy:

For those that don’t know, Clippy was the default avatar for the Microsoft Office Assistant in Office 97 to 2003. The actual name was Clippit, but if you point that out, most people won’t know or care.

Now, to be fair, the Office Assistant was terrible. It was so terrible that Microsoft even spoofed firing it in an ad campaign.

I draw a line between Clippy itself and the Microsoft Assistant, and try to look at what was intended with Clippy instead of how the botched implementation really set back computer-based assistants for a long time.

I remember the Microsoft Office Assistant from Office 97 through to 2003. In retrospect, having Clippy be the default avatar probably didn’t help the image of the Assistant feature, and most people never seemed to work out that other options were available. Personally, my favourites were Scribble the cat (from Office 97), and Lynx the cat from Office 2000.

So, what was the intent of Clippy?

The idea here was to have an assistant that anticipated your needs and assisted you in your work. The theory behind it really isn’t all that different from voice assistants we have now, like Alexa, Siri, and the Google Assistant. It just pre-dated them by more than a decade and wasn’t voice driven.

I honestly think there’s room for a new “version” of a desktop assistant like Clippy on the modern desktop.

Now, before I get bombed from orbit, I realize that individual workflows vary significantly. How someone using a Mac works is significantly different from someone running i3 on Arch, so when I say that there’s room for a version of a desktop assistant like Clippy on the modern desktop, I don’t mean that to be for everybody.

I’m more talking about something like a less crappy version of the Bonzi Buddy. Perhaps a graphical front end for Mycroft that was more interactive and had the capability of anticipating your actual needs.

I could definitely see a use for an on-screen assistant as an option for those new to Windows / macOS / Linux / whatever, or people with disabilities. One thing I’ve noticed lately is that, for all their advances, modern operating systems pretty much expect you to know roughly what you’re doing and don’t think to ask “do you need some guidance to get started?” Actual paper manuals are long gone, and how-to or for-dummies guides are an additional purchase.

TuneOut Radio — The Dumb-ening

Back in May of this year, I mentioned that several Internet radio stations were in the process of being ejected from TuneIn Radio’s directory. At the end of that post, I mused as to what other stations might get culled.

Last night I got the answer. Opening the TuneIn Radio app for the first time in a while on my iPad, I was greeted by a request to allow ‘personalization’, which is code for tracking plus additional adverts shoved in my ears. Needless to say, I clicked ‘no, thanks’ to that. Proceeding to my list of presets, I was annoyed to discover that all the internet-only stations were marked as ‘not available in your region’.

Continue reading “TuneOut Radio — The Dumb-ening”

Cutting The Cruft

This afternoon, I deleted a load of apps from my iPad Mini 4.

One of them — Good Sudoku — was becoming a time-sink as well as a source of annoyance at its occasional freezes.

Another was the WordPress app, which has never really proven that much use to me other than a source of additional notifications.

Most of the rest were apps that I’ve not opened in ages, if at all, and am unlikely to do so anytime soon.

I might add a few of them back at some point. Then again, maybe not.

I have come to the realization that, while my iPad can and does perform many useful functions for me, it doesn’t need to do everything.

“We need new paradigms, not more new tech.”

Big Tech uses our roads, our airports, our post offices — and rather than help prop up the system to work in the interest of all, the industry eschews responsibility and creates housing and employment crises by relying on gig-workers and contractors, effectively reducing the number of workers in the tech workforce eligible for employer-based healthcare. Big Tech fails to hire African Americans in high-paying jobs, and it actively suppresses wages for women and other people of color, especially Black women, who continue to make significantly less than men. It benefits from talent nurtured in public school systems and cherry-picks researchers favorable to its interests, without investing in the public infrastructure that enables those skills to develop. What we get back is an ecosystem of workers who can’t afford to live or pay rent, or experience homelessness, in wealthy tech corridors.

In short, instead of being partners in building the public good, Big Tech continues to profit from its erosion. Rather than contribute to the public coffers so that we can fund the public institutions we so desperately need to support, the titans of Big Tech are trying to come to the rescue through philanthropy and expressions of personal goodwill, making private donations at a tiny fraction of their personal wealth into the charitable and non-governmental organizations of their choosing. But these kinds of private drops into the desperately under-funded public pool, in exchange for eschewing the kind of tax responsibilities that working-people face, further erodes coordination and fair distributions of power and resources.

Source: The Loss Of Public Goods To Big Tech – NOEMA by Safiya Noble