Rolandt

The most common projects we turn down are those from people who want to create an online community but don’t yet have an existing audience.

It’s very hard to build a community if you haven’t yet earned the attention of an existing audience.

You don’t need a huge audience, but you do need a group of a few hundred people who will happily listen to what you have to say.

If you don’t have an existing audience, the alternatives are to buy an audience’s attention with social ads (expect to spend around $10 per each active member) or create a community concept so risky and so daring people will naturally spread the word.

Generally speaking, it’s best to earn the audience’s attention first. Follow the CHIP process. Create content, host events, interview people, and participate in existing groups.

You will always find it’s easier to start a new community if you’ve participated and supported related communities in the past.

sqlite-utils 1.11

Amjith Ramanujam contributed an excellent new feature to sqlite-utils, which I've now released as part of version 1.11. Previously you could enable SQLite full-text-search on a table using the .enable_fts() method (or the "sqlite-utils enable-fts" CLI command) but it wouldn't reflect future changes to the table - you had to use populate_fts() any time you inserted new records. Thanks to Amjith you can now pass create_triggers=True (or --create-triggers) to cause sqlite-utils to automatically add triggers that keeps the FTS index up-to-date any time a row is inserted, updated or deleted from the table.

Reading

The most likely explanation for negative interest rates is far simpler. The economy has become a giant kill zone. In venture capital circles, the term “kill zone” has become quite popular to describe the phenomenon of having no places to profitably invest.

How Monopolies Broke the Federal Reserve convincingly connects the dots between the rise of monopolies, negative interest rates, and the rate of innovation. I found it very convincing, and my experience working in tiny companies that compete with huge companies matched up with some of its discussion of the ‘economic kill zone’ that results from highly-resourced players.

But none of those people are the richest person here, which means they will keep succeeding despite—not because of—the man who is. He doesn’t know what they know; he doesn’t have to know. No one like him does.

The Adults in The Room is a remarkable final article by a politics editor at Deadspin that eviscerates the idea of experienced managers who swoop in and monetize media businesses.

It is so rare to find a role-playing game like this. There is no plot, no mystery, no dragons, no romance, no treasure. I still don’t know who I am or where I came from: my amnesia is never resolved. But I know why I am here, and that is enough.

I read Where it is Easy to Do Good on a park bench in Chicago after a lovely but overwhelming conference, and it hit me hard. I was more familiar with Everest’s art and generative fiction, but this short, impacting story reveals a really great writing talent too.

“My good friend Richard Serra is building out of military-grade steel,” he says. “That stuff will all get melted down. Why do I think that? Incans, Olmecs, Aztecs—their finest works of art were all pillaged, razed, broken apart, and their gold was melted down. When they come out here to fuck my ‘City’ sculpture up, they’ll realize it takes more energy to wreck it than it’s worth.”

This article about Michael Heizer’s city-sized new artwork.

I also watched a few things, like The Last Black Man in San Francisco (incredible) and Midsommar (just okay). It really seems like A24 is coming up with hits left and right - their releases are almost the majority of films I’ve seen recently.

Enigma

Right now my Enigma Machine notebook on Observable is making the rounds on the internet, achieving a sort of virality I’ve never enountered before - somewhere north of 8,000 folks have ‘liked’ it on Twitter.

Since it’s just such eye-candy, the feedback has been overwhelmingly positive. There was a little constructive feedback that I implemented to make sure that it’s as historically accurate as possible.

This was a lucky project to work on. Sure, there’s the implementation, which was tricky and required a few interlocking areas of expertise. But anyone else, given enough time, could have created something similar. It’s more that – by knowing Dana, by reading Jason’s book, by noticing that the Enigma was neat looking, by working at a company focused on explaining things, by having the ability to set some 9-5 time apart to work on this project for months on and off – and confirming that nobody had already claimed the task of creating a cool visualization of it - it all sort of lined up, in a way that’s much more luck than skill.

I’ve worked on lucky problems before – the saga of the DC Code was one of them, where a few key factors, where the political and intellectual winds blew in the right direction. And worked on unlucky problems, ones where I spent as much or more time and energy, but they were too early or too late for their time.

The other thing that I should write about this project, before it slips from my mind and I again complain about having accomplished nothing to my patient but irritated friends, is that it took four major revisions. The rotors looked sort of like chord diagrams, then they were skeuomorphic and 3D, and then they were flat and 1-dimensional, and then they were a mix of all those things. The final design was a combination of all of those, plus Tarek Rached’s pithy comment on an earlier prototype.

The final revision doesn’t fit cleanly into any of the classic visualization types that I reach for. Which is something that I usually try to avoid: familiarity is the most surefire path to understandability and usability, and inventing something new is risky. But the final design cleanly represented what this was: a bunch of stacked 1:1 maps that rotate and change over time. Many of the visualizations I admire are simply plain charts done really well, like those at the Economist, but the outside-of-the-box style is best exemplified by the work of Eleanor Lutz, of Tabletop Whale. They have a visual polish and deep creativity that begs for them to be printed and used as posters or wallpaper or full-body tattoos.

My colleague Matthew sent me a link to this compelling GIS application that allows exploration of the provincially-maintained roads on Prince Edward Island by annual average daily traffic:

The busiest road on the Island in 2018 (the latest year for which data is presented) was the Hillsborough Bridge between Charlottetown and Stratford, with 35,053 vehicles per day:

Four roads were tied for least busy, each with an average of 42 vehicles per day.

The first two comprise this stretch of road that includes the Aberdeen Rd., the Mill Rd., and the Mickle Macum Road, near Naufrage:

The third is the Gowan Brae Rd. near Souris:

A road that sees 42 vehicles a day is seeing under 2 cars an hour, on average. That’s a pretty sleepy road.

As you might expect, the collection of busiest provincial roads on the Island are those that run into and around Charlottetown: each of the roads highlighted in yellow on this map had a daily average of more than 12,000 vehicles:

Because the application supports exporting all of the data about the roads and the traffic counts to GeoJSON, it’s possible to cook up your own visualizations too. Here are the roads in QGIS, colour-coded by traffic count (the redder the line, the more traffic):

This visualization looks like a circulation system of the body. And, of course, that’s exactly what it is.

Kudos to the Department of Transportation, Infrastructure and Energy for releasing this data, and developing such a useful tool for exploring it.

The best workshops are transformative.

They leave people feeling smarter and more confident than before.

They equip participants with new information.

They challenge participants to use this information to solve engaging problems.

They leave participants with a peer group they can call upon for support afterwards.

You can have workshops for your top members, newcomers, or attendees of your conference. When done well they are powerful, and woefully underused, tools.

p.s. 5 tickets remaining for my 80-person workshop at CMX Summit this week.

Total functions are functions that give you a valid return value for every combination of valid arguments. They never throw errors and they don’t require checks of the return value. Total functions help you write more robust code and simpler code, free from defensive checks and errors.

What is a total function?

Total functions are functions that give you a valid return value for every combination of valid arguments. They never throw errors and they don't require checks of the return value. Total functions help you write more robust code and simpler code, free from defensive checks and errors. https://share

Transcript

Eric Normand: What is a total function? By the end of this episode, you will know how this idea helps you write more robust systems.

Hi, my name is Eric Normand and I help people thrive with functional programming.

This is another important idea for math that we can borrow in our programs. It’s used often in functional programming. We’re not talking about totals functions. We’re talking about mathematical, pure functions. These are calculations, they don’t have side effects.

However, it does apply to actions, to side effects, if you want. You can take this concept, it’s a flexible idea, and apply it to those things. We’ll just be talking about it for pure functions, mathematical functions, in this episode.

All right, so let’s get started. What is a total function? Quite simply, a total function has an answer for every combination of valid arguments.

That kind of asks a new question — what is a valid argument? We’re going to say that in type languages it is very clear. If you’ve got static-type…

Eric: …arguments to a function are all values that belong to the argument types.

If it says int — whatever that means, in whatever language you use — then every possible int, if it’s an int, then it’s a valid argument to that function. If it say string, that means every possible string. Any way you can make a string, then that’s a valid argument.

For untyped languages, it’s a little bit more complicated. Usually when we’re programming in untyped languages, the types are in our heads, and we have some unwritten, implicit, informal idea of what we mean as valid arguments.

That’s what I mean, these informal rules about what are the valid arguments. That’s what you’ve got in an untyped language.

If you can make them more explicit, if you can have some kind of checking on your arguments, some kind of preconditions, contracts, some kind of assertions, whatever you want to put in there to claim this is what’s valid, all those things will be helpful for making this more useful.

That said, I myself, for my own code, when it’s just me, I kind of know what the valid arguments are. I tend not to use too many assertions and things. Sometimes, I do. They’re helpful sometimes. I just want to put that out there. We’ll discuss this a little bit more, later in the episode. It’s a hairy discussion.

Total function has an opposite. Just to clarify again, just to repeat, for every valid argument. If it says it takes ints, I can pass any int and get an answer. That’s a total function. The opposite is a partial function. Maybe one of those ints doesn’t work.

A classic example is division. They type of division, let’s say, it takes two numbers and returns a number. Divides the two numbers and then returns a number, but you can’t divide by zero, so it throws an error or it does something. Returns some infinity, or not a number, or something like that. Returns nil, null, whatever.

All those things are kind of like the function line. The function said I take two integers and I return an integer, but it didn’t. It returned something else or threw an exception. It didn’t even return. It broke when you passed a zero for the second argument, for the…

Eric: We can see how this creates problems. First of all, robustness. If you have a divide by zero error in your program, it’s going to fail in production and it’s not going to be as robust. It’s is not correct.

If you do want to make it correct, what do you do? You put an if statement before you divide, just to check. That makes more conditionals.

This is what the type system is supposed to be handling for us. We don’t want them check that the arguments are valid. We want to just run division, and we want the compiler to find these problems. This is in a typed language.

Another example of people doing this — I see this occasionally and it really bugs me, because I really try to do total functions as much as possible — let’s say you have a person class, and there are two subclasses. There’s employee, and there’s volunteer. Employees get a salary. Volunteers don’t.

On person, there is a method called salary. That means you can call salary on employees and volunteers. When you call salary on employee, it gives you a number. It’s like $10,000 a year or whatever. Then on volunteer, it throws an exception. The message says something like, “Volunteers don’t get paid.” That’s the friendly error message.

To make this total, you would need the salary method to return a number in every single possible case, instead of throwing an exception. That means that the volunteers’ salary should return zero. That’s the mathematical definition. I know there might be some reason why you didn’t return zero there.

You threw an exception, because you’re probably doing something wrong somewhere else in your code, and you want to signal, “Hey, we’re trying to get the salary of some volunteer. That’s probably an error.” That means like, “We don’t know if they’re a volunteer or an employee. We’ve lost track.”

You know what? I think that you’re modeling it wrong. If only employees have a salary, the salary method shouldn’t be on person, it should be on employee and have your compiler check that.

What a subclass is supposed to mean is that you can forget what the subclass is. You just care about the main class that it’s all subclassing from, the superclass.

If you don’t know that it’s a volunteer, that’s supposed to be OK. Just saying that this is an issue that can come up, even in something like an object-oriented model, where you have different methods that you think should be called. Maybe it’s an indication that your method is in the wrong place.

That’s one way to handle it, is to change the definition of what valid means for the arguments like we did with this person. It’s not valid to call salary on just any person. It has to be an employee.

Another way that you can handle this is just make sure that you handle all the cases. Make volunteers’ salary method return zero. That’s it, just make it happen. Don’t use nils and exceptions. Find a value that makes sense.

Zero happens to work in this case because their salary actually is zero. Sometimes, people will return a null instead of an empty array when there’s no answers.

It normally returns an array of something, an array of answers. If there’s no answer, it returns null. Why don’t you return the empty array? In that case, that’s still correct. Now, that function is total, and it works in a wider range of cases.

You don’t have to check the return value. You can call length on it and know that there’s zero things in there. Likewise, you can replace all those exceptions, trying to find some value that works within the already assigned return type.

That’s not always possible. Sometimes, you’re going to have to augment the return type. I guess that’s number three — augment the return type. What does this mean? You could make the empty case explicit. Change your type so that it has an empty case.

This is in an untyped language where you’re dynamically typed. You’re a lot more flexible. Make nil explicitly accepted. I do that in Clojure. I say it’s either a number or a nil. This is what numbers mean, and this is what nil means.

I don’t want to overload nil with too much meaning. In this specific case, this specific type that I’m returning, number means this. Nil means this. I don’t do that all the time, but sometimes I do.

Now, if I’m returning something like a variant, a tagged union, I might add a case to that variant. That’s like no answer. You could use a maybe or option type, instead of using null. If you’re going to throw an exception, change the return value to a Maybe Int.

If you had division, you could make division total by saying, “Division doesn’t return a number. Division returns a maybe number.” If you could pass zero in for the denominator, I’m going to return nothing. You’re going to have to check that.

Optionally, if there’s errors and things, you can make those part of the type. In Haskell, you would use an either, but somehow fold the errors into it. The way I imagine this is if you have a HTTP client, usually you can have it configured where it never throws an exception if the HTTP request doesn’t go through. You can have it actually return the error as a response.

If the client got in touch with the server and the server had an error like a 404 or a 500, that is a valid response. It’s an error, but it’s a valid response. The error modes have been baked into the response type. There’s codes, status messages, and things like that.

Even some clients — the good ones — will turn a time-out, which means they don’t even talk to the server sometimes. It’ll turn those into the same kind of error type.

The status 700, which isn’t a real HTTP type of status type, but we’ll call it status 700 time-out. Something like that. Then you can handle that in the same way with the same kinds of logic as you would handle even a valid response, a correct response.

The last way to handle these partial functions and make them total is to augment the argument type. In our case of division, instead of saying it takes two integers or two numbers and returns a number, we say it takes a number and a number that’s not zero. That’s a type — number that’s not zero — and it returns a number.

Now, we have to make this new type called number that’s not zero. Somehow, somewhere else, we will make one of those. When we make it, we check to make sure it’s not zero.

Now, this function knows that it trusts this type. It’s not going to be zero. It can unbox it and divide. It would call whatever internal division that’s unsafe partial function inside.

Here’s a warning. These last two things where you’re augmenting the return type and augmenting argument type, they still need to check somewhere. They’re just deferring the if statements. Before, you were doing unsafe division. You weren’t checking if it’s zero. You’re going to have to catch that error.

That’s like a conditional. Somewhere it’s like, “Oh, if it throws an error, do this. Otherwise, you get an answer, do this.” It’s a branch. You have a branch. Now, if you turn it into a maybe — the return value becomes a maybe number — you still have to branch. You still have to say, “Is it a nothing or is it a just number?” You still have to do that branch.

It’s not reducing the complexity, really. It’s deferring it somewhere else. You still have all these other ways of dealing with maybe. For example, it’s a functor. It’s a monad. You can use it in other ways. It fits in with what you’re doing elsewhere in a nice way maybe. It’s a little bit better.

Likewise, if you augment the argument type, you still have to make one of these things somewhere else. You probably don’t have one when you need it. You have a number, and you need to make a number that’s not zero, however how you make one of those things.

When you make one of those, something’s got to check if it’s zero. What if it’s zero? It’s going to return a nothing. There’s a maybe on the other side, too. You still have the branch. It’s still in there. It’s just now you have the type system on your side.

You can push all these if statements out to the edges. That’s the metaphor that people like to use — that there’s edges. When input comes in, the user types something in, check it for all these things — the zeros, etc.

Now, in the happy path where it’s not zero and you pass it through, there’s no checks. It’s all good. When it comes out, you got an answer. You’re pushing it to the edges. All the way, as close as possible to where the user input comes in or to where the user output goes out. Those are called the edges.

You’re deferring it, pushing it out, so you’re able to create a space in the middle where your functions are all total, where everybody’s happy. Where you don’t have to check things before you call methods on them. You don’t have to worry about error cases, etc.

Of course, this is a design decision. It is up to you to decide how far you want to push stuff to the edges. Whether it’s worth it in particular cases, etc. A lot of people push it really far to the extreme.

Some people don’t mind checking if a number is zero right before they divide. Some people don’t even check and they just let it throw there, and say, “Well, we’re just going to have errors.” That’s fine. It depends on your software and what you need to do with it.

I do want to come back to this idea that types do help with totality. I’m going to use Haskell as an example. Haskell has several built-in functions that are partial. I talked about division. It is partial in the base Haskell language. Also, head, which gives you the first element of a list. If the list is empty, that throws an error. That’s partial, this is the built-in function called head.

In general, Haskellers, they either look at that as a mistake or a wart on the language because they want total functions. They might rewrite head to turn it into like it returns a maybe. Something like that. They might call it maybe head or something like that.

Type systems do help in a number of ways.

One is they give you this way of defining what is valid so that the compiler checks it. We talked about how difficult it was in a dynamically typed language to even talk about what valid arguments means. Type system gives a definition of what a valid argument is.

It also will, in many cases, be able to check whether you’ve handled every case. If you have what’s called a tagged union, or a variant, or you have multiple cases, different constructors for the same type, there’s a flag you can set on the compiler that says, “Make it an error if I don’t handle one of those cases.”

That’s the compiler being on your side, helping you make total functions.

Types can help with these total functions. The language can help as well, for instance in Haskell. Another design decision of the Haskell has nothing to do with the type system, but there’s no null pointers in Haskell. That’s really nice.

Null pointers were a mistake. [laughs] The inventor of null pointers admitted this. They didn’t make the same mistake in Haskell. They learned from that, so there’s no null pointers.

Nulls, especially in languages like JavaScript or Java where you have a type system…Like in Java you have a type system, but you can always put a null basically anywhere. The type system won’t help you with that at all. That sucks.

As far as dynamically-typed languages go, I do want to say that this is one of those sore spots in dynamic languages where there’s one way to look at it, which is that every function accepts any kind of argument.

It’s true. [laughs] In a certain way of looking at it, I can go into my Clojure REPL and type plus, string and null. It will try to add them and it’ll say, “Ah, these aren’t numbers. Bleh.”

It’s true. I could run it. I did run it. I got an exception, so in some way every function is partial because you can always find some value of some type that wasn’t expected that will throw an exception. It’s true.

Except, that most…This is the same of object-oriented languages that are dynamic, you can send any message. There’s no checks on what messages you’re allowed to send. Very often, you’ll get something…in JavaScript, you get undefined is not a function because it couldn’t find that method in the object. Or you’ll get method not found or whatever it’s called in your language.

That’s another ding where it says, “You can pass any message to anything.” Everything’s partial. Every method is partial because you can find some object that doesn’t answer that message. Yes.

But there’s another way to look at it which is that there is an intended type for every argument. Sometimes that type is really complicated, and sometimes it’s not specified. Sometimes, it is not explicit, but there is an intended type for every argument.

Does that mean that they’re all total? No, not saying that either. It’s a lot more total [laughs] than with the other perspective. These are things we’re used to dealing with in dynamic programming, dynamic-type languages. We’re used to thinking the compiler’s not going to help me. I got to remember the types.

Well, that’s all I want to say. That is the extent that we have a good definition of valid in dynamically typed languages. There is an intended type. It doesn’t mean it works everything.

For instance, I think in JavaScript, in Clojure, divide still takes two arguments, two numbers that still the intended type, and it sometimes is an error. It’s still undefined for divide by zero. It’s just the way it is. It’s the same in Haskell.

Yeah, all right. Let me recap. It’s been total functions. It’s very useful for robustness if your functions are total. You have to do fewer checks. You’re not going to get the errors.

If you’re using only total functions, no errors, none. It’s not possible. Maybe you get something else. Something like out of memory or stack overflow, something like that, but you’re not going to get error because you chose the wrong arguments.

Total means you have an answer for every combination of valid arguments. The opposite is a partial function. Usually it means it throws an exception, or it returns null or nil, or some other error thing it does.

Why? It gives you robustness. It also improves your simplicity. You just have fewer conditionals within your center, inside the edges. A lot of times you’re just pushing conditionals to the edge, but inside you’ve got this nice total, no errors, no conditionals. Everything works right inside.

I went over four ways that you can increase the totality of your function. The first one was that you could just move stuff around because things are on the wrong types, or classes, or you just designed it poorly, not using the constraints of the language properly.

The second one was sometimes you just mistakenly forget to handle a case. You thought it should be an exception, but really, within the stated type or the intended type, there is a valid response for that. Salary on a volunteer should be zero or it shouldn’t even be on there.

The third one is you can augment the return type. If it turns out that there isn’t a good response, you’ve got to change the type. You’ve got to make it a maybe. Maybe you can add cases to your type so that you’re explicitly calling out the empty cases.

You could fold the error conditions, the failure modes, into the type, just like HTTP does. It uses the same response format for 200 responses as it does for 400 responses. Still got headers, still got a status code, still got a body. It’s all the same.

The fourth one is to augment the argument type. You could say this is a number that’s not zero. It’s somehow somewhere else you’re making sure that those are getting created correctly, but this function doesn’t have to worry about that and that’s nice.

When you do this, you can push these conditionals, checks, and things out to the edges. If you’re checking stuff from the arguments it’s going closer to user input and you’re pushing the return type checks further out into where it’s getting output.

If you enjoyed this episode, if you learned something, there are a lot of other episodes. You can find all the past ones at lispcast.com/podcast. All the past episodes have audio, video, and text transcripts. Whatever medium you’re into, we’ve got it.

You’ll find links to subscribe, including the video, the RSS for the text, also the RSS or iTunes link for the podcast audio. You can also find social media links at that same address lispcast.com/podcast.

Please get in touch. I love discussing this stuff with people that’s why I’m doing it, broadcasting, trying to find people who I can relate to out there. It’s a lonely world. Got to find people that like the same stuff you like, have a lot in common.

My name is Eric Normand. This has been my thought on functional programming. Thank you for listening and rock on.

The post What is a total function? appeared first on LispCast.

Over the last few years, Logitech-owned Ultimate Ears has firmly cemented itself as one of the top portable Bluetooth speaker companies. As expected, the UE WonderBoom 2 carries that legacy forward.

Throughout the summer, the WonderBoom 2 has been my go-to cottage speaker. Its compact size, passable sound, battery life and waterproofing has made it the perfect speaker for me — and likely you as well depending on what you’re looking for from a portable speaker.

While it may not sound as good or last quite as long as the ultra-popular UE Boom 3, it makes up for those shortcomings with its ultra-portable size and reduced price.

Even though this speaker isn’t for audiophiles, the value it offers is impressive.

Same look, but that’s not bad

The WonderBoom 2 looks almost exactly like its predecessor, but with two new buttons: one on the top and one on the bottom. The speaker’s volume buttons are slightly larger as well.

The button on the top of the speaker borrowed from last year’s UE Boom 3 is called the ‘Magic Button.’ It’s a fairly simple button that acts similarly to what is sometimes featured on headphone cords. This means that you press it once to play/pause and twice quickly to skip to the next song.

The button on the bottom — the one in the shape of a tiny pine tree — triggers the ‘Outdoor Boost Mode.’ This feature makes songs sound louder and crisper but also drops the bass level a bit. Either way, this button is crucial for listening to music outside in wide-open spaces.

The other two buttons on the top are for power and syncing the speaker. Beyond the new buttons and the features they bring, the WonderBoom 2 also looks awesome.

It’s about the size of a softball and weighs 420g.

The speaker is waterproof and has an IP67 water and dust resistance rating. That means that it’s able to be submerged in a metre of water for about half an hour. While I haven’t left the speaker in the water for that long, I’ve gone swimming with it and it’s also sat outside overnight. The Wonderboom 2 took both tests like a champ. It’s also worth noting that the speaker floats.

Overall, I think the speaker features a great design that’s both stylish, portable and includes interesting new features.

Hey, Listen!

Regarding audio quality, the speaker’s sound isn’t as clear as the larger UE Boom 3, but it’s able to turn up the volume loud enough when you need to. The Outdoor Boost mode also significantly helps give it an extra kick in some scenarios.

While UE markets the speaker as perfect for the great outdoors, I also found that it excels inside.

Playing music while I work at my desk or when I’m in the shower is great since the little speaker can easily fill a room with sound. While the sound stage isn’t perfectly balanced like a Sonos or other larger, more expensive smart speakers, it offers excellent audio for its price tag.

Inside, the UE WonderBoom 2 is packing two 40mm active drivers and two 46.1mm x 65.2mm passive radiators.

These are slightly smaller than the Boom 3’s speakers, and you can tell when you’re listening to both speakers side-by-side, but when isolated, they both sound great.

What else is packed into this thing?

Beyond the two new buttons and the sound quality, the speaker also has a few hidden tricks.

For example, you can pair two WonderBooms together. If you have two speakers near each other, then you need to press and hold the Magic button until you hear the pairing sound to play music in stereo.

UE has also included a button combo for checking the speaker’s battery level. Holding both volume buttons will play one of three sounds that indicate the speaker’s battery level.

You can also pair two phones or other devices to the speaker at the same time so you and a friend can take turns playing songs.

What is a bit odd is that the WonderBoom 2 doesn’t connect to the same app as the UE Boom 3. While I admittedly rarely use the app with the Boom 3, the extra features it offers make it worthwhile in some cases.

UE says the WonderBoom 2 features 13-hour battery life and in my experience, that number holds up. I’ve never had it playing for 13 hours straight, but the WonderBoom 2 has easily made it through a weekend at a cottage without needing to be charged. Just like the UE Boom 3, the company once again isn’t including a charging brick in the box, which is annoying and a strange decision on UE’s part.

The post UE WonderBoom 2 Review: The perfect portable speaker appeared first on MobileSyrup.

For the NASA Earth Observatory, Adam Voiland describes about two decades of fires:

The animation above shows the locations of actively burning fires on a monthly basis for nearly two decades. The maps are based on observations from the Moderate Resolution Imaging Spectroradiometer (MODIS) on NASA’s Terra satellite. The colors are based on a count of the number (not size) of fires observed within a 1,000-square-kilometer area. White pixels show the high end of the count—as many as 30 fires in a 1,000-square-kilometer area per day. Orange pixels show as many as 10 fires, while red areas show as few as 1 fire per day.

There are a lot of fires, but a bit surprising given the news lately, the total area burned each year is decreasing.

Tags: fire, NASA

In the wake of the terrifying violence that shook El Paso and Dayton, there have been a lot of questions around the role of the Internet in facilitating communities of hate and the radicalization of angry white men. Digital affordances like anonymity and pseudonymity are especially suspect for their alleged ability to provide cover for far-right extremist communities. These connections seem to be crystal clear. For one, 8chan, an anonymous image board, has been the host of several far-right manifestos posted on its feeds preceding mass shootings. And Kiwi Farms, a forum board populated with trolls and stalkers who spend their days monitoring and harassing women, has been keeping a record of mass killings and became infamous after its administrator “Null”, Joshua Conner Moon, refused to take down the Christchurch manifesto.

The KF community claim to merely be archiving mass shootings, however, it’s clear that the racist and misogynistic politics on the forum board are closely aligned with that of the shooters. The Christchurch extremist had alleged membership to the KF community and had posted white supremacist content on the forum. New Zealand authorities requested access to their data to assist in their investigation and were promptly refused. Afterwards, Null encouraged Kiwi users to use anonymizing tools and purged the website’s data. It is becoming increasingly clear that these far-right communities are radicalizing white men to commit atrocities, even if such radicalization is only a tacit consequence of constant streams of racist and sexist vitriol.

With the existence of sites like 8chan and Kiwi Farms, it becomes exceedingly easy to blame digital technology as a root cause of mass violence. Following the recent shootings, the Trump administration attempted to pin the root of the US violence crisis on, among other things, video games. And though this might seem like a convincing explanation of mass violence on the surface, as angry white men are known to spend time playing violent video games like Fortnite, there has yet to be much conclusive or convincing empirical accounts that causally link videogames to acts of violence.

One pattern has been crystal clear, and that’s that mass and targeted violence seem to coalesce around white supremacists and nationalists. In fact, as FBI director Christopher Wray told the US Congress, most instances of domestic terrorism come from white supremacists. From this perspective, it’s easy to see how technological explanations are a bait and switch that try to hide white supremacy behind a smoke screen. This is a convenient strategy for Trump, as his constant streams of racism have legitimized a renewed rise in white supremacy and far-right politics across the US.

For those of us who do research on social media and trolling, one thing is for certain, easy technological solutions risk arbitrary punitive responses that don’t address the root of the issue. Blaming the growing violence crisis on technology will only lead to an increase in censorship and surveillance and intensify the growing chill of fear in the age of social media.

To better understand this issue, the fraught story of the anonymous social media platform Yik Yak is quite instructive. As a mainstream platform, Yik Yak was used widely across North American university and college campuses. Yak users were able to communicate anonymously on a series of GPS determined local news feeds where they could upvote and downvote content and engage in nameless conversations under random images to delineate users from each other.

Tragically, Yik Yak was plagued by the presence of vitriolic and toxic users who engaged in forms of bullying, harassment, and racist or sexist violence. This included more extreme threats, such as bomb threats, threats of gun violence, and threats of racist lynching. The seemingly endless stream of vitriol prompted an enormous amount of negative public attention that had alarming consequences for Yik Yak. After being removed from the top charts of the Google Play Store for allegedly fostering a hostile climate on the platform, Yik Yak administrators acted to remove the anonymity feature and impose user handles on its users in order to instil a sense of user accountability. Though this move was effective in dampening the degree of toxic and violent behavior on Yik Yak’s feeds, it also led to users abandoning the platform and the company eventually collapsing.

Though anonymity is often associated with facilitating violence, the ability to be anonymous on the Internet does not directly give rise to violent digital communities or acts of IRL (“In-real-life”) violence. In my ethnographic research on Yik Yak in Kingston, Ontario, I found that despite intense presence of vitriolic content, there was also a diverse range of users who engaged in forms of entertainment, leisure, and caretaking. And though it may be clear that anonymity affords users the ability to engage in undisciplined or vitriolic behavior, the Yik Yak platform, much like other digital and corporeal publics, allowed users to engage in creative and empowering forms of communication that otherwise wouldn’t exist.

For instance, there was a contingent of users who were able to communicate their mental health issues and secret everyday ruminations. Users in crisis would post calls for help that were often met with other users interested in providing some form of caretaking, deep and helpful conversations, and the sharing of crucial resources. Other users expressed that they were able to be themselves without the worrisome consequences of discrimination that entails being LGBTQ or a person of color.

What was clear to me was that there was an abundance of forms of human interaction that would never flourish on social media platforms where you are forced to identify under your legal name. Anonymity has a crucial place in a culture that has become accustomed to constant surveillance from corporations, government institutions, and family and peers. Merely removing the ability to interact anonymously on a social media platform doesn’t actually address the underlying explanation for violent behavior. But it does discard a form of communication that has increasingly important social utility.

In her multiyear ethnography on trolling practices in the US, researcher Whitney Phillips concluded that violent online communities largely exist because mainstream media and culture enable them. Pointing to the increasingly sensationalist news media and the vitriolic battlefield of electoral politics, Phillips asserts that acts of vitriolic trolling borrow the same cultural material used in the mainstream, explaining, “the difference is that trolling is condemned, while ostensibly ‘normal’ behaviors are accepted as given, if not actively celebrated.” In other words, removing the affordances of anonymity on the Internet will not stave off the intensification of mass violence in our society. We need to address the cultural foundations of white supremacy itself.

As Trump belches out a consistent stream of racist hatred and the alt-right continue to find footing in electoral politics and the imaginations of the citizenry, communities of hatred on the Internet will continue to expand and inspire future instances of IRL violence. We need to look beyond technological solutions, censorship, and surveillance and begin addressing how we might face-off against white supremacy and the rise of the far-right.

Abigail Curlew is a doctoral researcher and Trudeau Scholar at Carleton University. She works with digital ethnography to study how anti-transgender far-right vigilantes doxx and harass politically involved trans women. Her bylines can be found in Vice Canada, the Conversation and Briarpatch Magazine.

https://medium.com/@abigail.curlew

Twitter: @Curlew_A

Headline image via: Source

Just a month ago I wrote here about my reservations concerning the use of mobile phones as hotel room key. A hotel I will be staying at in the near future yesterday started sending me multiple (unasked) SMS’s to download their hotel app to ‘make my stay smarter’. Sure, I will trust download links in unrequested SMS! Today as I’ve ignored their messages I received an e-mail imploring me to do the same.

The app they ask me to use is called Aeroguest, and their pitch to me is about easier check-in/out, using chat to contact staff, and using my phone as door key. The first two I’d rather do in person, and the last one is not a good idea as explained in the above link.

Why such an app might be seen as attractive to the hotel, becomes clear if you look at the specifications of the app. A clear benefit is direct repeat bookings, saving the expensive middle men that booking sites are. In my case I almost always book through the hotel’s website directly. And if I enjoyed my stay I usually book the same hotel in a city for my next visit. I do use booking sites to find hotels. In this case I’ve stayed in this hotel several times before.

The stated benefits for the guest (key, chat, check-in/out, choosing your room) are a small part of the listed benefits for hotels in using the app, such as up-selling you packages before and during your stay. An ominous one, when seen from the guest’s perspective, is ‘third party services’ access presumably meaning potential access to your booking / stay history and maybe even payment / settlement information, requested preferences etc. Another, more alarming one, is “advanced indoor mapping” which I take means tracking of guests through the hotel which can yield information on time spent in hotel facilities, time spent in the room, how often / when the key was used, and by matching it with other guests, whom you might be meeting with that is also staying in the hotel. In Newspeak on the apps website in the data and analytics section this is described as “With transparency, you can proactively accommodate your guests’ needs.” Note that the guest is the one who is being made transparant. That is quite a price in exchange for being able to choose your specific room when checking in with the app.

I’ve replied to the hotel my reasons for not wishing to use the app (linking to my previous blogpost), and told them I look forward to checking in at reception in person when I arrive. When I arrive I am curious to hear more about their usage of the app. For now “making my stay smart” reads like the “smart cities” visions of old, it may be smart, but not for the individuals involved, merely for the service provider.

Only 17 years separate these two oblique angled shots of the Downtown peninsula. Since our 2002 image was taken, over 26,000 residential units have been added Downtown and in the West End. That’s around 140 additional buildings of 10 or more storeys. Thousands more units are under construction and in the development stream, and even then the peninsula is by no means ‘built out’ – although sites are fewer, and harder to find.

There’s still a gap on the far right, on the waterfront, where the Plaza of Nations, and further Concord Pacific sites have yet to be built. There are a number of sites reserved for non-market housing inland behind and between the condo towers built by Concord on the former Expo lands, and a recent deal should see over half developed as non-market, with others returned to Concord for more market development.

On the left of the image Vancouver House is nearly complete, (so Trish Jewison, who photographed the 2019 shot from the Global BC News helicopter took the picture recently). From this angle the twisting taper of the building is almost invisible. In the middle of Downtown, the Wall Centre’s upper floors were reclad almost as dark as the bottom, so the distinctive two-tone effect in 2002 has been lost. From this distance the Empire Landmark wasn’t so obvious in 2002, but in 2019 it’s gone, and the replacement condo towers will be shorter. The Shangri La and Trump Hotel and condo towers almost line up from this angle, so only one tall tower appears in the distance.

Over on the right, the BC Place stadium has its new(ish) retractable roof, surrounded by new towers, with the distinctive rust red of the Woodwards Tower behind. The original ‘W’ was still in place in 2002 – now it’s down on the ground, and a replacement revolves in its place. Not too many new office towers have been added to the Central Business District, but that’s changing. Ten office buildings are currently being built, the most office space ever added to the city at one time, and much of it already leased. The biggest building is the Post Office, getting a pair of office towers added on top, with the huge building (that fills an entire city block) changing to office and retail space.

0900

I’ve long been a fan of Alan Kay. We met twenty five years ago as we were building a consulting firm that blended strategic and technology insight. One of Alan’s favorite observations is “point of view is worth 80 IQ points.” Choosing a better vantage point on tough problems is time well spent, especially when there is pressure to get on with it.

I’m not sure I can count the number of times I’ve heard or said that we live in a knowledge economy. That we are all knowledge workers who live and work in learning organizations. Yet, we continue to celebrate the industrial revolution in those organizations. We celebrate scale and growth and control. We worry about the problems of accelerating change but assume that working harder and longer will suffice to keep pace.

There is a better vantage point. It is to treat knowledge work as craft work in a technological matrix. Craft work integrates materials, tools, and practices to create artifacts that simultaneously embody the skill and expertise of crafters and meet the practical and esthetic needs of patrons.

Examining each of those elements from a craft perspective illuminates what it takes to become effective as a knowledge worker and remain so as change continues to accumulate. It’s our 80 IQ point move.

Materials – make them visible to make them manageable

Industrial work is built on repeatability; my iPhone 6 Plus is fundamentally identical to yours; any differences are cosmetic. Give me the same consulting report you prepared for your last client and we have a problem. The output of knowledge work derives value by being unique.

Knowledge work produces highly refined abstractions; a financial analysis, a project plan, a consulting report, a manuscript, or an article. A piece of knowledge work evolves from germ of an idea through multiple, intermediate representations and false starts to finished product. Today, that evolution occurs as a series of morphing digital representations which are difficult to observe and, therefore, difficult to manage and control.

A pre-digital counterexample reveals the unexpected challenges of digital work. I started consulting before the advent of the PC. When you had a presentation to prepare for a client, you began with a pad of paper and a pencil and sketched a set of slides. Erasures and cross outs and arrows made it evident you were working with a draft.

This might be two weeks before the deadline. You took that draft to Evelyn in the graphics department on the eighth floor. After she yelled at you for how little lead time you had given her, she handed your messy and marginally legible draft to one of the commercial artists in her group. They spent several days hand-lettering your draft and building the graphs and charts. They sent you a copy of their work, not being foolish enough to share their originals.

Then the process of correcting and amending the presentation followed. Copies circulated and were marked up by the manager and partner on the project. The graphics department prepared a final version. Finally, the client got to see it and you hoped you’d gotten it right.

Throughout this process, the work was visible. Junior members of the team could learn as the process unfolded and the final product evolved. You, as a consultant, could see how different editors and commentators reacted to different parts of the product.

Today’s digital tools make the journey from idea to finished product easier in many respects. When knowledge artifacts are digital, however, they are hard to see as they develop.

So what? Only the final product matters, right? What possible value is there to the intermediate versions or the component elements? Let’s return to the bygone world of paper again. Malcolm Gladwell offers an interesting observation in “The Social Life of Information:”

”But why do we pile documents instead of filing them? Because piles represent the process of active, ongoing thinking. The psychologist Alison Kidd, whose research Sellen and Harper refer to extensively, argues that “knowledge workers” use the physical space of the desktop to hold “ideas which they cannot yet categorize or even decide how they might use.” The messy desk is not necessarily a sign of disorganization. It may be a sign of complexity: those who deal with many unresolved ideas simultaneously cannot sort and file the papers on their desks, because they haven’t yet sorted and filed the ideas in their head. Kidd writes that many of the people she talked to use the papers on their desks as contextual cues to “recover a complex set of threads without difficulty and delay” when they come in on a Monday morning, or after their work has been interrupted by a phone call. What we see when we look at the piles on our desks is, in a sense, the contents of our brains.”

I have friends whose digital desktops have that look about them but this strategy doesn’t readily translate to the digital realm. The physicality of paper gave us version control and audit trails as a free byproduct.

Digital tools promote a focus on final product and divert attention from the work that goes into developing that product. “Track changes” and digital Post-It notes provide inadequate support to the process that proceeds the product. Project teams employ crude naming practices in lieu of substantive version control. Software developers and some research academics have given thought to the problems of how to manage the materials that go into digital knowledge artifacts. Average knowledge workers have yet to do the same.

Visibility is the starting point. Once you make the work observable, you can make it improvable. Concepts like working papers, and audit trails, and personal knowledge management can then come into play.

Tools – Every Day Carry and Well-Equipped Digital Workshops

Where craft matters, so do tools. That got lost in the industrial revolution. Tools were carved out and attached to minute pieces of process, not to the people who wielded them with skill. Meanwhile, the raw materials of knowledge work–words, numbers, and images–did not call for much in the way of tools other than pencil and paper. Mark Twain was an innovator in adopting the typewriter to improve the quantity and quality of his output. But the mechanical tools for aiding knowledge work came to be seen as beneath the dignity of important people.

There was a time when “computers” were women charged with carrying out the menial tasks of doing the calculations men designed and oversaw. It was not that long ago when executives thought it perfectly sensible to have their email printed out and prepare their responses by hand. These attitudes interfere with our abilities to be fully effective doing knowledge work in a digital world.

There’s the old saw that to a child with a hammer, everything looks like a nail. To a skilled cabinet maker, every problem suggests a matching hammer. A well-equipped workshop might contain dozens of different types of hammers, each suited to working with particular materials or in specific situations.

If our materials are digital, then our skill with digital tools becomes a manageable aspect of our working life.

There’s a useful distinction between basic and specialty tools. A basic tool in hand beats the perfect tool back in the shop or office. I’ve carried a pocket knife since my days as a stage manager in college. Courtesy of the TSA I have to remember to leave it behind when I fly or surrender it to the gods of security theater but every other day it’s in my pocket. There is, in fact, an entire subculture devoted to discussions of what constitutes an appropriate EDC—Every Day Carry—for various occupations and environments.

In the realm of knowledge work, Every Day Carry defaults to an email client, calendar, contact manager, word processor, and spreadsheet.  For most knowledge workers, tool thinking stops here. Other than software engineers and data scientists, few knowledge workers give much thought to their tools or their effective leverage. Organizations ignore the question of whether knowledge workers are proficient with their tools

If you are judged on the quality of the artifacts that you produce, you would do well to worry about your proficiency with tools. If you have control over your technology environment, set aside time to extend your toolset and learn to use it more effectively. Invest time and thought into how to design, organize, and take advantage of a knowledge workshop filled with the tools of your digital trade. Plan for a mix of EDC, heavy duty, and experimental knowledge work tools.

Practices – Design Effective Habits

Process thinking built the industrial economy. To deliver consistent quality product, variation is designed out and all the steps are locked down and controlled. If your goal is to craft unique outputs suitable to unique circumstances, industrial process is your enemy.

Where then are the management leverage points if industrial process is not the answer? McDonalds is not the only way to run  restaurant. In a knowledge work environment, both design and management responsibilities must be more widely distributed and shared. Peter Drucker captured this when he observed that the first question every knowledge worker must ask is “what is the task?”

Answering that question entails understanding the materials and tools available. From there, knowledge workers can design approaches to creating the necessary unique knowledge artifacts. Habits, routines, rituals, and practices replace rigid processes. In a fine restaurant, the day’s fresh ingredients set the menu and the menu guides which preparation and cooking techniques will be called for that evening. Line cooks, sous chefs, and chefs collaborate to create the evening’s dining experience.

The building blocks for constructing suitably unique final products are learned over years of practice and experimentation. They are passed on through observation and apprenticeship. In a volatile knowledge economy, they must also be subject to constant evolution, refinement, and innovation.

Learning as a craft practice

In the pre-industrial craft world, learning could be a simple process. Find a master and apprentice yourself to them. Time would suffice to transfer expertise and skill from master to apprentice.

We do not live in that world.

In an industrial world, learning was focused on fitting people to the work. Open-ended apprenticeship was replaced with narrow training programs to learn the specifics of where humans fit into a larger, engineered, process design.

We do not live in that world.

Integrating a craft point of view with the pace of the technological environment that now exists makes learning a craft practice to master.

We are all permanent apprentices. We are also all permanent masters of our craft. Apprenticeship must become conscious and designed. Mastery will always be temporary. Our understanding of materials, tools, and practices will always be dynamic. Learning and performing will be in constant tension.

The post The 80 IQ point move: knowledge work as craft appeared first on McGee's Musings.

It’s a case of the cobbler’s shoes – I rarely find the time to think about how to articulate my own process, but this hot-air balloon really helped me out…

How to create 10 visual ideas from a 1-hour conversation: https://visualthinkery.com/10-ideas/

The post 10 ideas appeared first on Visual Thinkery.

Honda maintains an iterative approach with its HondaLink infotainment system, but the 2019 version takes a few key steps in the right direction.

One thing automakers are slowly getting better at with their respective systems is usability, especially as it relates to aesthetics. The new-look HondaLink follows the trend, featuring a simpler layout that looks better because of improved graphics.

The current system isn’t in every Honda model just yet. I got to test it in a 2019 Accord Touring, but it also comes in the Insight, Pilot, Passport and Odyssey. There are worthwhile points in what Honda did here, and the room for improvement seems within reach.

The basics

When I previously tested HondaLink, the system had undergone a sea of change that included adopting a capacitive touchscreen as well as embracing Apple CarPlay and Android Auto. The user interface is one of the biggest differences this time, though it does go further than that.

One of the key changes was going back to a physical input for volume. When Honda overhauled things in 2015-16, it opted to go with a volume slider that both customers and reviewers maligned as being unintuitive. It’s gone now, replaced with a knob to more easily adjust how loud or low you want to hear. Steering wheel controls are still there.

Many of the foundational pieces remain, and for the most part, the system functions similarly. Some of the new elements, however, do add to the package in positive ways.

Connections and layout

At the centre of it all is the 8-inch display, which seems to have a sharper resolution than the previous one. Part of that may be a credit to the improved UI that feels cleaner and easier to navigate. Icons are larger and sub-menus shorter, making it a little easier to get from point A to B within the system.

Flanking the screen on both sides are physical buttons acting as shortcuts to various software features. I liked it because it negated having to touch the screen all the time, and reduced the number of steps to get to something. I would argue the biggest difference is the volume knob, which Honda had already returned to last year.

Interestingly, Honda didn’t include USB-C ports in the Accord. While some other manufacturers have already embraced the popular port, traditional USB-A reigns here. What’s neat is that both the dash and centre console each have the same active port. Either one can interface with the system to run CarPlay or Android Auto. I tried to see if I could somehow run both at once, but it forces you to choose one.

There are two USB-A ports in the rear-facing passengers there, though they’re only for charging. There is no aux-in jack, staying consistent with what the rest of the industry has been doing. The 12-volt socket is in front next to the dash USB port.

Honda embedded an NFC tag into the dash on the passenger side, and it proved to be a rapid way to pair a phone with the car. It didn’t work with an iPhone, as expected, yet was smooth with Android phones I tested.

I didn’t get a chance to test the in-car Wi-Fi hotspot for lack of a data bucket I could use, but it seems to offer the same thing others do. AT&T handles the connection and data, with the same pricing other cars have. It’s still 20GB for $200 prepaid over 12 months — still the best deal of them all, in my view. Like the others, you can also roam into the United States and not pay extra.

The wireless charging pad under the dash is a little recessed and not the easiest to access because of its proximity to the stick shift. It’s a limited feature in that it’s only standard in the Touring, Sport 2.0 and Touring 2.0 trims.

Voice and texting

Things get a little interesting with Bluetooth. Unfortunately for Android users, Google Assistant doesn’t operate that way — it only works through Android Auto. That’s not the case with Siri, which pipes up when long-pressing the steering wheel’s voice button.

A short press brings up Honda’s own voice assistant, which lives within the car, so offers typically limited depth. That includes calling and texting, the latter of which comes with an unusual twist.

I was surprised to see that I could actually read incoming text messages in full — even while the car was in drive. Other legible incoming text functions I’ve tested in other vehicles often limit it to when the car is in park. I asked Honda Canada to confirm this was, in fact, a feature and not a bug, and was told it was the former, not the latter.

The system can read out incoming texts with a simple tap on the screen when they pop up. The messages app also maintains constant access to messaging, including the ability to respond.

Responses, however, are more of the canned variety, but I did get away with verbalizing basic ones. Siri and Google Assistant are far better at it, yet this is one of the deeper voice integrations I’ve seen in an infotainment system independent of those platforms.

CarPlay and Android Auto

There weren’t many surprises here. With Honda having already supported both platforms going back to 2016, the Accord has included them since at least 2015 in some markets.

The big difference, at least for me, had little to do with Honda, and that was the new-look Android Auto. I had already opted to try it out prior to this test drive, and found it a significant improvement over the original UI. CarPlay will get its own little makeover when the next version of iOS goes live to the public.

Not all the physical buttons on the sides shortcut to anything on either platform. Instead, they leapfrog past them and go to HondaLink. For example, the ‘Home’ button goes to the HondaLink main menu, not the home screen for Apple or Google’s respective platforms. However, the back button seemed to backtrack within those platforms. Skipping or repeating tracks always works too.

Honda will ultimately benefit from what these two platforms will look and function like when they’re both out of beta. Since they live on a phone, the automaker doesn’t have to do anything, and the improved UI only makes HondaLink look better.

HondaLink app

I wasn’t set up with a way to test out the HondaLink app for iOS and Android, which allows for some remote access to the car. The Accord Touring can access a variety of features within the app, including remote start, remote lock/unlock, Find My Car and stolen vehicle tracking and assistance.

Not all these features come free. After an initial trial period of 12 months, subscribing after that ranges from $99 to $369 per year, depending on what package you go with.

I should also note that Honda has a unique section that deals with APK files. There is a way to upload them, but it’s not clear how far this goes when it comes to sideloading apps. Honda says the files are “for internal use to manage the system,” without elaborating further.

The post 2019 HondaLink Infotainment Review: Looking better all the time appeared first on MobileSyrup.

John Gruber has mentioned, on The Talk Show, that I’ve got some weird ideas about what beta means.

Here are my definitions:

development (d): everything is in progress and the app might be completely unusable.

alpha (a): the app is feature-complete and has no known bugs — but, importantly, it’s had very little testing.

beta (b): the app is feature-complete, has no known bugs, and has been tested — but further testing is still warranted. Every beta is a release candidate.

These are defined in a NetNewsWire Technote. It’s important to have definitions that everybody working on or testing the app understands.

But why these rather strict definitions?

It’s part of our commitment to quality. What matters is the end result — the shipping app — and these definitions make sure we don’t get to beta, or even alpha, with the app up on the table with wires sticking out and pieces missing.

This gives us a big space between development and shipping, and that space is all about making sure the bugs are all fixed.

This is a matter of ethics and pride in our work. Absolutely.

But it’s also pragmatic. This is an open source app, written by volunteers in their spare time, and having this rhythm baked-in to the process helps make sure we can uphold our standards even without full-time developers, managers, and testers.

* * *

And… it bugs me how little real attention our industry pays to quality these days. In some cases the consequences are disastrous; in other cases they’re merely expensive. It doesn’t have to be this way.

If it seems like I’m going too far with my definitions, well, I’m trying to bend the stick here.

It's hard to imagine not knowing what to do with a book, but that was in fact the case in the days before near universal literacy. Now we're in that position again, but with respect to digital technologies, says Dave Cormier. The comment comes in the context of presenting a dradft of "a model for preparing an education system for the internet" and Cormier suggests that "some people are never going to make it all the way to being ready to teach with or on the internet."The model offers four progressively more accomplished levels of digital literacy, from basic awareness, though use, creativity, and finally, teaching.

This paper describes "the progressive merging of invisible tools, working in the background (Ubiquitous Computing), integrated into every domain of life and connected to the Internet (Internet of Things), with independently thinking and acting actors (Artificial Intelligence), implemented both into environments (Ambient Intelligence), and into everyday objects (Smart Objects)." In so doing, it extracts "our dimensions of modern technologies: Connectivity, Invisibility, Awareness, and Agency." Interesting read.

The headline may seem to suggest it's a job title that make a person a learning technologist, but that isn't the intent. Rather, the point is that learning technologists may have many different job titles. "34%(13) of respondents stated that they had Learning Technologist in their title, whilst 66% (25) had a different title, e.g. education, blended, designer or other that includes duties of a Learning Technologist." That seems like an odd place to start a series on "What makes a learning technologist" but I imagine the remaining articles of this series will have more substance.

The number refers to colleges in the United States, which has a different system than most of the rest of the world. "They're going to close, they're going to merge, some will declare some form of bankruptcy to reinvent themselves. It's going to be brutal across American higher education." The article tells us only that the expert, Michael Horn, "studies education at Harvard University", but I would imagine it's this Michael Horn from the Christensen Institute, which may lead you to take the prediction with a grain of salt.