On Tuesday this week, awlnx and Krombel of Freifunk München (FFMUC) gave a virtual presentation online about their public Jitsi Server cluster installation and how they have set up the system. Currently, it serves well over 400 simultaneous participants in the evenings in 70 concurrent rooms. And with their 20 video bridge servers they say that they are prepared for over a thousand simultaneous users!
In their 30 minute talk (in German), they give an overview of how they have set up their Jitisi cluster with a front end server that communicates with 20 Jitsi bridge VMs to handle the overall load. If you are interested in the topic but don’t speak German, the slides they show in the video are still insightful even without the German commentary, and further information is available on this Wiki page in English.
Apart from feeling reassured that Jitsi can be scaled if required for my purposes, the other interesting piece of information I got out of the session is that Firefox is currently not working well at all with Jitsi. FFMUC has even gone so far as to block access to their system when Firefox is detected. They say this has become necessary as Firefox not only gives a bad experience to one person but also has significant negative consequences for all other people on a bridge, as videos are not shown and are not rate limited.
Initial mask wearing was good -- around 80 percent. By November cases were down, and public health officials recommended re-opening the city. Residents rushed to entertainment venues after having been denied this communal joy for months. The Mayor himself was fined by his own police chief after going to a show without a mask.
But a second wave surged in Dec 1918, and SF's Health Officer again urged people to wear masks voluntarily. [...]
San Francisco residents were fed up. This was the second wave of the pandemic, and they had already spent months between Sept and Nov being hassled, fined and even arrested for not having a mask on. Challenges of constitutionality were heard.
Christian Scientists objected, arguing that it was "subversive of personal liberty and constitutional rights." Civil libertarians argued that if health officials could force them to wear masks, then it could force them to inoculate "or any experiment or indignity." [...]
Over 2,000 people attended an event formed by San Franciscans called themselves 'THE ANTI-MASK LEAGUE,' denouncing the mandatory masking ordinance. [...] Moderates in the Anti-Mask League wanted to circulate an anti-mask petition. Extremists wanted to initiate recall proceedings of SF's Public Health Officer. [...]
We learn through this episode that various groups of Americans have been pushing back against public health measures for more than a hundred years -- and for similar reasons!
As of 8pm tonight, wearing a mask in public places around other people who are also wearing masks is both forbidden and required by New York law. I'm sure the police will enforce these contradictory mandates in a completely rational & nondiscriminatory fashion. [...]
Although @mjs_DC is undoubtedly right that many anti-mask laws were passed to fight the KKK, New York's anti-mask law actually dates to 1845 when it was passed to suppress the "anti-rent movement" of upstate tenant farmers.
NY's anti-mask law has been used against radicals ever since. In 2004, a state appeals court upheld the convictions of 11 anarchists who "covered their faces with bandanas while shouting epithets and political slogans during a May Day demonstration in Union Square Park."
Steffani Cameron, Social-Distancer@SnarkySteff
If you would like to try making this meal yourself, I estimate it cost me about $10 per portion, including the wine… twitter.com/i/web/status/1…
I’ve been using Pis as thin clients for ages, but it hadn’t really dawned on me yet that the 4 is indeed powerful enough to run WebRTC-based video conferencing software.
My oft-considered experiment of cleaning out my desk, moving everything with fans to a faraway closet and just having a Pi 4 with two monster 4K monitors is closer to becoming a reality, and this is another data data point to consider.
Apple has released design instructions and a manufacturing guide for its face shields to encourage others to develop their own versions.
Due to the ongoing COVID-19 pandemic, there has been a shortage of medical equipment across the world. Tech giants like Apple have contributed by developing their own versions of protective equipment and donating it to hospitals in need.
Apple has donated hundreds of thousands of face shields to date, and has now released instructions for other companies to do the same.
The manufacturing guide includes details about the materials that are needed to produce the shields, along with tips on how to properly pack them for transport.
Releasing its designs for the shields is a helpful move on Apple’s part because it could help other companies that are trying to aid public health efforts by allowing them to skip the design steps and go straight to the production steps.
The tech giant notes that the manufacturing instructions should only be used by an expert in a factory environment since the production requires professional level expertise.
Einige Bilder sehen in Retrospektive wie Photoshop aus, Straßenmusik ohne Publikum, offene Geschäfte ohne Kunden, Flughafen ohne Flugzeuge. 20 Minuten Momentaufnahme einer Geisterstadt ohne Menschen. Langsam füllen sich die Straßen wieder mit Menschen, mit diesem Kurzfilm wollten wir ein Bild von Berlin einfangen, als die Straßen noch leer waren, ein Bild von einer historischen Situation, die hoffentlich nicht wieder kommt.
Translink’s No. 1 job: Safety and cleanliness, so the public has confidence that taking transit is, if not entirely risk-free, safe to use when following the guidelines we will use for daily life. Financial support is necessary now, of course, but without the former, it’s not sustainable.
For lessons and guidance, I’d immediately turn to those places that have the same task, experience, and confidence of their community.
Like SMRT in Singapore:
Seoul Metro:
Taipei:
A quick glance at the pic above and you might think it’s SkyTrain. Later this year, let’s hope it is.
Zoom’s stock: Up and down and up and down and up and to the right.
As the coronavirus crisis built to pandemic levels in early March, a relatively unknown tech company confronted a defining opportunity. Zoom Video Communications, a fast-growing enterprise videoconferencing platform with roots in both Silicon Valley and China, had already seen its market cap grow from under $10 billion to nearly double that. As the coronavirus began dominating news reports in the western press, Zoom announced its first full fiscal year results as a public company. The company logged $622.7 million in revenue, up 88 percent from the year before. Zoom’s high growth rate and “software as a service” business model guaranteed fantastic future profits, and investors rewarded the company by driving its stock up even further. On March 5th, the day after Zoom announced its earnings, the company’s stock jumped to $125, more than double its price on the day of its public offering eleven months before. Market analysts began issuing bullish guidance, and company executives noted that as the coronavirus spread, more and more customers were flocking to Zoom’s easy-to-use video conferencing platform.
But as anyone paying attention to business news for the past month knows, it’s been a tumultuous ride for Zoom ever since. As the virus forced the world inside, demand for Zoom’s services skyrocketed, and the company became a household name nearly overnight. Zoom’s “freemium” model – which offers a basic version of its platform for free, with more robust features available for a modest monthly subscription fee – allowed tens of millions of new users to sample the company’s wares. Initially, Zoom was a hit with this new user base – stories of Zoom seders, Zoom cocktail parties, and even Zoom weddings gave the company a consumer-friendly vibe. Just like Google or Facebook before it, here was the story of a scrappy Valley startup with just the right product at just the right time. According to the company, Zoom’s monthly users leapt from 10 million to more than 200 million – an unimaginable increase of 2,000 percent in just one month.
Just as quickly, however, Zoom became the subject of controversy. Like Google and Facebook before it, Zoom’s success as a product comes from its unwavering focus on convenience. Zoom makes it as easy as possible to use its platform. Employing invisible technical tricks, Zoom engineers made the platform easy to install, easy to share, and … easy to hack. Press reports about “Zoom bombing” began dominating the headlines, and as reporters dug in, so did reports of significant (and long ignored) security failings. Large corporations, state governments, and school districts banned the company’s products. Media outlets began to investigate the company’s Chinese roots – only to discover that the young firm had mistakenly routed user sessions through its servers in mainland China. Zoom responded quickly, freezing product development and focusing entirely on fixing critical security issues. The company then updated its privacy policies, bowing to criticism that it might leverage user data in a manner similar to Google and Facebook (more on that below).
But with China and the United States entering a third year of an increasingly heated trade war, and blaming each other for the origin of the novel coronavirus, Zoom finds itself in an extraordinary position that no amount of crisis communications can overcome. Zoom’s founder and CEO, Eric Yuan, is a Chinese ex-pat and naturalized American citizen. More than 700 of his 2,500+ employees live and work in China. Until March of this year, Yuan was held up as an example of the best that global capitalism can offer – an ingenious immigrant who bootstrapped his way to America and leveraged hard work, smarts, and venture financing into a multi-billion dollar fortune.
Now Zoom’s brand – and its future – live under storm clouds of suspicion. In just four weeks, the company has inherited the full force of the American “techlash.” And the companies previously at the center of that storm – in particular the “Big Four” of Apple, Facebook, Google and Amazon – are happy to pass along that unpleasant mantle.
So what might Zoom do next?
* * *
As some readers know, I’ve been a student of the “Big Four” for more than two decades. For the past 18 months, that study has focused on the terms of service and privacy policies of the Big Four. Thanks to the work of researchers and faculty at Columbia’s School of International Public Affairs and Graduate School of Journalism, we’ve published a study of the underlying architecture of the Big Four’s core policies, a visualization we call “Mapping Data Flows.” This tool breaks down and compares each company’s privacy and data use policies, with a goal of giving both ordinary consumers and academic researchers insight into the architecture of control currently dominating our economy’s relationship to data.
Given its very recent and extraordinary rise as a consumer tool, we decided to apply this approach to Zoom’s terms of service and privacy policy as well. You can find our initial findings here.
As with every research project, our study of Zoom’s policies began with a working hypothesis. One of the most interesting findings from our initial study of the Big Four was how similar their policies were – they all collect vast sums of data, and their terms of service allow them nearly unlimited usage of that data. And of course, all four granted themselves the right to collect, process, and employ user data for the purpose of pursuing advertising businesses – a multi-hundred billion dollar industry driving what Harvard scholar Shoshana Zuboff calls “surveillance capitalism.” We therefore asked ourselves two questions: First, would Zoom’s current terms of service and privacy policies allow them to join the Big Four in the pursuit of an advertising business? And second, given Zoom is (or was) a business facing, as opposed to a consumer-facing platform, would its privacy policies and terms of use be markedly different from the Big Four?
The short answer to that first question is yes. And for the second? That’d be a no. As the first image below demonstrates, Zoom collects a ton of data, and its policies are quite similar to those of its Big Four cousins.
Figure 1 – Zoom’s Data Collection visualized
But exploring that first question – whether Zoom might become an advertising-driven business – yielded even more interesting insights:
Figure 2 – Zoom’s data collection for purposes of Advertising.
As the above illustration from our new visualization demonstrates, our research shows that nearly all data collected from Zoom user sessions may be used for the purpose of advertising. Despite the clarification of Zoom’s privacy policies posted on March 29th around usage of data from user sessions, nothing material changed in its actual policies. Indeed, the company writes that “We are not changing any of our practices. We are updating our privacy policy to be more clear, explicit, and transparent.”
To be clear, Zoom does not currently run an advertising business along the lines of Facebook, Google, Apple, or Amazon’s (and yes, both Apple and Amazon have significant data-driven advertising businesses, they just don’t like to talk about them). So why, in its own policies, does Zoom reserve the right to use all collected data for the purpose of “advertising”?
As any lawyer will tell you, words are slippery things. Certainly in the context of Zoom’s current business, the word “advertising” covers the company’s role as an advertiser – as a brand that uses data to market to current and potential customers using platforms like Google or Facebook. But a careful reading of the company’s policies reveal how easily the same words could allow the company to pivot from advertiser on other platforms to provider of platform advertising, should the company wish to. In other words, there’s nothing stopping Zoom from joining the Big Four as a major player in the provision of advertising services, should it wish.
How might Zoom do such a thing? And given its current privacy backlash, why would Zoom ever consider such a move?
Let’s start with the How, then we’ll cover the Why.
As I mentioned at the start of this piece, Zoom’s current business is based on what folks in the tech industry call a freemium SaaS (software as a service) model. The company makes a version of its platform available to anyone for free, and then “upsells” those free users to a paid version that has more bells and whistles, like the ability to record conferences, larger numbers of participants on a videoconference, and so on. Pricing starts at $15/month, scaling up to thousands a month for large customers. This model is most often employed for enterprise customers (Slack is a good example), but it’s also found success in consumer-facing applications, where more often than not users pay to avoid advertising (think YouTube or Hulu). Regardless of whether the service is enterprise or consumer focused, free users always outnumber paying ones by an order of magnitude or more.
One of the most difficult elements of a freemium SaaS model is luring those free users “down the funnel” into paying for a monthly subscription. So how might Zoom convince its bumper crop of roughly 190 million new consumers to start paying up?
By now you’ve probably figured out where I’m going with all this. Zoom could implement a free service that’s supported by advertising, then encourage users to pay for a version that’s ad free. Doing so would be ridiculously simple: Just as with YouTube, Zoom could force its users to watch a “skippable” pre-roll video ad before the start of each videoconference (and it could use its data trove to make those ads extremely targeted). Well aware that such an interruption would be an annoyance at best, Zoom could then offer to strip the ads out for customers who paid a small subscription service of, say, $5 a month. If just one quarter of its customer base decided to do so, Zoom’s revenues would jump by $250 million a month – adding a cool $3 billion a year to its top line revenue, nearly all of which would be pure profit. The resulting advertising business could easily add hundreds of millions, if not billions more. That’s five times more revenue than the company reported in its last fiscal year.
Which brings us to the “Why” of this admittedly speculative (but nevertheless quite reasonable) exercise. And that why comes down to capitalism. Zoom is a public company with a massive valuation – more than $40 billion at the time of this writing. That gives it an unsustainable price to earnings ratio of roughly 1,750 – 76 times larger than the S&P average. The pressure to “grow into” those outsized expectations is enormous. Zoom is staring at a multi-billion dollar business model just begging to be implemented. For its shareholders, board, and senior executives, the question isn’t why it should be adopting the business model that made Facebook, Google, Apple, and Amazon the most valuable companies in the world. Instead, the question is simply this: Why shouldn’t it?
In another post, we’ll explore answers to that question (and how Zoom, if it’s thoughtful, could help reimagine the core architecture of surveillance capitalism). For now, take a spin around our newest visualization, and give us input in the comments below. Thanks for reading, and take care of yourself – and others – out there.
###
The Mapping Data Flows project is seated at Columbia SIPA – we are grateful for the support of Dean Merit Janow, as well as the support of the Brown Institute at Columbia’s Graduate School of Journalism, the Omidyar Network, and faculty and staff including Mark Hansen, Juan Francisco Saldarriaga, Zoe Martin, Matthew Albasi, Natasha Bhuta, and Veronica Penney. Hat tip as well to Doc, who’s been focused on these issues for decades.
This week the Park Board placed big red jersey barriers to create S-curves on Park Drive on the steep downhill portion from Prospect Point. No notice, no explanation, but presumably to slow down cyclists, particularly where the Tatlow Trail crosses the road.
Well, bless their hearts, after a couple of days observing that the bike slalom course they’d created on the Stanley Park downhill was dangerous, the Park Department has made the chicanes much more rideable. On Wednesday, each of the five chicanes forced cyclists descending Stanley Park Road into a tight S curve to squeeze between a pair of offset plastic barriers. Today, Saturday, the barriers were more widely spaced, so the curve to slide between them was much more gentle. No more squealing brakes or casual riders on the edge of control.
But if the point is to slow riders on the descent, forcing a mix of riders aged eight to eighty, casuals to road warriors into narrow gaps, makes little sense.. The descending road is a wide two lanes. Signs at the top warning of the descent, and advising slow riders to keep to the right, are all that’s needed to reduce or avoid conflicts.
The more urgent need is for signs that keep riders from riding the wrong way on the parks one-way roads. At the bottom of the hill, I met a dozen tourists wandering the wrong way, unsure where to go, and grateful when I got them turned around. There were others, worriedly descending the climb to Prospect Point like salmon swimming upstream. How had they gotten that far without figuring out their mistake? Lack of signs, could it be?
Closing Stanley Park roads to cars, so that walkers, runners and cyclists can exercise and enjoy nature at social distance, is wonderful. Very early last Wednesday, we came upon crews putting up the first barriers. I pointed out that one barrier blocked cyclists from exitIng the park to Beach Drive. As they moved it, they said that the changes were a work in process, and that they expected to make other changes.
But why do this by trial and error? Why not start with experienced consultants to design the best signage and traffic controls? Cyclists and city planners have complained for decades that the Park Board wants cyclists to ride to parks, but not through them. There are examples throughout Vancouver.
I had posted about the way Tumblr uses time-series monitoring to alert on crash spikes in the Android and iOS applications. Since then, we’ve done a lot of work to reduce the overall volume of crashes. As a result, we created a new problem: it was possible for a handful of people, caught in crash cycles, to cause our stability alerts to trigger.
Once the stability alert is triggered, we typically start looking in the crash logging systems, like Crashlytics or Sentry, to find more information about the crash. We found an increasing number of occurrences where no particular crash could be easily identified as causing the spike.
Getting paged at 2am because of a stability alert is not great, but not finding a crash was even worse.
The problem was the way we were monitoring. Simply watching all crash events wasn’t good enough. We had to start normalizing the events across the users. Thankfully, we collect events and not simple ticks. We have rich data in the crash event, including a way to group events coming from the same device.
Here is an example of the two styles of monitoring over the last week:
raw count shows a large spike, but the unique device count shows a normal trend
That after-midnight Raw Count spike on Friday would have paged our on-call person if we hadn’t changed to alert on the Unique Device Count instead. We still use the Raw Counts to identify issues and investigate, but we don’t alert on them. We can use the high-cardinality events to zero-in on the cause of the spike. In this case, two (2) people were having a bad experience using their Cubot Echo devices.
Since moving to the new alerting metric, we’ve had far fewer after-hour pages, while still being able to focus on the stability of the applications across our user base.
Cumulative deaths by country, plotted against number of days since 10th death occurred; graphic by John Burn-Murdoch for UK FT.
This is the fourth in a series of articles about the current pandemic, drawing on the best available science and data we currently have at our disposal. The previous articles:
The Least We Can Do (what we’ve learned and what we can do, as at Apr 15, 2020)
This article updates the above by describing some new learning about this virus that might dramatically change what happens now, and the courses of sensible action open to us. Sources include former infectious disease expert Dr Michael Greger’s videos, and summaries from recent science journals.
So here we go: Ten things even most experts didn’t know a month ago:
The mortality rate for the virus, so far, looks to be more like 0.2% of those infected than the 1% experts initially expected. [EDIT: May 14th: Now it appears the mortality rate is closer to 1.0%; see follow-up here.] The reason for the large error is that we didn’t realize how many were actually, apparently, contracting the disease asymptomatically, and how many would recover without treatment or not even show symptoms, thanks mainly to a near-global lack of testing capacity that led to utter uncertainty about the mortality rate and who was infected. We now have a slightly better idea, principally from the number of deaths and hospitalizations (still a very rough number, but much more useful than positive tests when so few have been tested and so many are asymptomatic). The consequences of this 0.2% rate, if it’s correct (and some small-scale intensive testing studies suggest it is), is that this virus in its first wave will probably kill only about 1-2 million worldwide, which is about three times the annual number of deaths from “regular” influenzas in recent years. It also suggests that five times as many as suspected have actually contracted the disease, perhaps 10% of the global population, and perhaps twice that proportion in Europe and North America, without serious incident. That still leaves the vast majority vulnerable to infection when containment measures are relaxed, and to the next wave(s) of the pandemic, but still means five times as many people have acquired or have natural immunity than most experts expected.
There are six coronaviruses known to affect humans, of which four are endemic “common cold” viruses. You’ve almost certainly had them. The problem with these coronaviruses is that, unlike HxNx type flu viruses, these coronaviruses only confer immunity on those infected for an average of 45 weeks [EDIT May 3rd: This originally read 45 days, but it appears this was a misreading of the data by the data analyst], after which you can get them again. So the hopes for eventual permanent “herd immunity” to CoVid-19 may be wishful thinking; we might see cases recur and spike until a reliable vaccine has been developed and almost universally given to humans. This is a highly contagious disease, perhaps five times more than we’d thought, and a small minority un-inoculated could wreak havoc.
Infection with coronaviruses can have serious and lasting, debilitating effects on its victims, even those who seem to have fully recovered. These viruses, and our immune systems’ “cytokine storm” reactions to them, can leave tissues and organs severely damaged and susceptible to later organ failure or other infections, and to neural, psychological and brain damage. We may not know for years.
There has not been a single confirmed case of anyone getting CoVid-19 from delivered groceries or other packaging. That’s not to say there is no risk, but compared to the risks of shaking hands or being near a cough or sneeze, they are hardly worth worrying about.
Masks work, but not so much to prevent the virus from being transmitted to the wearer; they do significantly reduce the risk that you, if you are unknowingly infected, will transmit it to others, or that you will touch your eyes, nose or mouth with an infected hand, and hence give the virus its needed passage to your own throat and lungs. For the same reason, gloves are less important than masks in preventing disease spread (they can carry the virus much as your hands do).
You are most infectious before you show any symptoms. That’s why our lack of preparedness, and hence shortage of test kits, allowed the disease to spread so far so fast. In most countries, you couldn’t even get a test, and were assumed uninfected, unless you showed symptoms.
Although the science is not yet certain, so far it seems unlikely that you can contract the virus through skin lesions (or otherwise through your skin) or through sexual or anal transmission. However, since the virus can pass through to feces and end up in aerosol spray from flushed toilets, you should close the lid before you flush public toilets.
Pangolins, a kind of anteater and the likely intermediary for the coVi-19 virus from bats to humans, are the most trafficked animal in the world. They have been hunted nearly to extinction both as a ‘delicacy’ for humans and as a source for ‘cures’ in ‘traditional Chinese medicine’. China has now banned their consumption but not their ‘medicinal’ use.
There is no need to use hot water when washing your hands for 20 seconds or more with soap.
It bears repeating that the second wave of the 1918 influenza, in some reckonings the biggest killer of humans in history, resulted from a fast and dramatic mutation that shifted the most vulnerable victims from the old and immune-suppressed, to the young with the strongest immune systems (who mostly died due to ‘cytokine storm’ responses from their own immune systems). Average age of victims shifted from about 60 (first wave) to about 30 (second and most virulent wave).
It should also be noted that vaccines are not panaceas — as I’ve written elsewhere, viruses are not our enemies, but essential parts of every ecosystem. And vaccines don’t guarantee 100% immunity. As we continue to crowd farmed and exotic animals into ever-more confined and crowded spaces, and then eat them, the recent explosion of new viruses and mutations is very likely to outstrip our capacity to develop vaccines (as is already starting to happen with ‘regular’ influenza vaccines).
What’s more, there will always be a horde of rabid anti-vaxxers ready to sabotage vaccine programs any way they can, and refuse to inoculate themselves, to everyone’s peril (read Eula Biss’ brilliant book On Immunity for the full extent of this challenge). Viruses are particularly adept at mutation, and we could soon find ourselves in a never-ending and hopeless race to try to keep ahead of the mutations, precisely when they are getting cleverer, as we provide them more fertile breeding grounds, at adapting to our vulnerabilities.
For a brief but comprehensive summary of everything we know right now, please read this.
My friends Darcie and Peter cycled to the New Years Levees, so they’ve got me beat. But at almost two thirds of the year, that’s still a pretty long season, and it’s a lot longer than the “well, for six months of the year, you can’t bicycle” reasoning that automobile enthusiasts bring out.
For the first time this year, I did some basic spring maintenance myself: MacQueen’s has a backlog of 70 bicycles to get to before they could get to mine, and I wasn’t patient enough to wait. I oiled the chain, tightened the bolts, and even trued the back wheel, with the help of a spoke-tightening tool loaned to me by my friend Cynthia.
Oliver’s bike, and my bike, are now ready for the season. We’re about to head off on our first ride.
For the past month or so, I've been thinking about the longer-term technology, policy and business trends that might emerge in the wake of the current pandemic. I'm especially interested in those that could directly or indirectly affect the use and deployment of networks and communications.
I wrote up my initial scenarios for what might lie ahead for the telecoms industry in the recent STL Partners report on COVID-19: Now, Next & After (link), and also discussed them on the STL webinar on the same topic (link). The next update webinar is on May 6th - link.
I've also participated in other client webinars and podcasts on campus networks (link), private 4G/5G (link) and Wi-Fi6E (link) recently - and I always include a section considering the pandemic's impact. (Any market analysis or opinion formed more than 2 months ago now needs to be reconsidered in the light of the pandemic and coming economic recession).
With that in mind, one area I've started thinking about is that of in-building wireless and smart buildings, especially relating to business locations. (Residential coverage of cellular and better home broadband / Wi-Fi is also top-of-mind, but I'll tackle that separately another time).
Obviously, offices and shopping malls are currently empty in much of the world, but eventually they will return to regular use, to some level at least. Even buildings sadly vacated by companies that cannot survive the economic impact will likely gain new tenants and uses.
Making buildings pandemic-proof
We already have building codes and regulations to protect us against fire risks, and even earthquakes. For fires, we have sensors, alarms, fire escapes, drills, signage and so on. In parts of the world there are specific rules governing indoor coverage for public safety radios, and they are being updated as agencies upgrade from P25 / TETRA systems to 4G / 5G critical-communications cellular alternatives.
So what else would it take to make a building "pandemic-proof"? I'm especially interested how we manage social distancing - both during the next phase of recovery and a gradual return to near-normal when a vaccine becomes available, but also during possible future waves of COVID or new entirely outbreaks.
In the wake of the 2008 financial crisis there was a big focus on banks' transparency, financial stability and regulatory "stress tests". I'd be very surprised if equivalent changes don't take place over the next few years - especially as many coronavirus infections are understood to occur indoors.
I've found various articles about smart buildings and the pandemic already, where the main focus seems to be on general hygiene and infection control. Using thermal cameras (and perhaps facial recognition) can automate detection of people with fevers. LED lights can provide disinfection in some cases, and bathroom sensors can help enforce hand-washing. Remote access to building-management systems allow facilities personnel to work from home. Better management of temperature and humidity may reduce the survival time of viruses and bacteria.
I can imagine a range of strategies being adopted in coming years:
Temperature-detection and hygiene management, as above.
Ability for remote building-management wherever possible
Design guidelines for wider corridors and stairways, better ventilation, virus-unfriendly surfaces, automated doors rather than handles, and so on
Ways to impose, measure and enforce social-distancing rules in emergencies - for example by dynamically lowering maximum numbers of permitted people in enclosed spaces, or digital signs for making corridors or aisles into one-way systems.
Use of sensors to measure occupancy, density and flow of people, and control entry/exit better
Automated disinfection systems or processes (maybe using robots)
Use of occupants' / visitors' phones or other devices to help them navigate / work more safely
Ability for authorities to use cameras, admission-control and other data for contact-tracing purposes (subject to emergency laws on privacy etc).
Clearly, not all of these can apply to all buildings - and there is obviously a huge spectrum of venue types with different requirements. A supermarket is different to an office block, a corner-shop, a factory or warehouse full of robots. Older buildings are not likely to be able to widen corridors, while a "cube farm" has more flexibility.
But what that means is that in a future outbreak, a government could say: "Workplaces certified to standard PNDMC-A can remain open, if they reduce occupancy to X, Y & Z metrics. PNDMC-B locations must comply with emergency rules A, B & C. All others must close."
Clearly, those type of rules will incentivise building owners and developers to upgrade their sites wherever possible. While it is too early to guess exactly how the specific regulations might be formulated, there are nonetheless some initial ideas and steps to think through.
The role of networks
Given my own focus on mobile and wireless systems, a key theme immediately leaps to my mind: many of these techniques and practices will require better and wider indoor connectivity than is common today in many places.
While some building-management systems will be based on wired connections (not least as they'll need cables for power anyway), I expect wireless networks to be extremely important for much of this.
I see wireless networks being employed both indirectly (for connection of sensors, cameras or other devices, such as smartphones used for distancing apps) and directly by using the network itself as a sensing and measurement tool. Indoor mapping and positioning will be needed in tandem with wireless for various use-cases.
There are particular challenges and opportunities for indoor wireless systems here:
There will be a need to support both public networks (for indoor use of nationwide MNO networks and services) and localised private wireless, for the building or company's own needs.
Almost inevitably, both 3GPP cellular (4G/5G) and Wi-Fi (5/6/7) will be essential for different use-cases and device types, plus public-safety wireless such as TETRA. In many some cases additional technologies such as Bluetooth low-energy, ZigBee or proprietary systems will also be required as well.
All of this will occur while major transitions to 5G (at different frequencies) and private cellular networks are ongoing in coming years.
Any real-time mobile app, whether it is giving alerts, or uploading updates on location, will be dependent on good wireless connectivity, either via Wi-Fi or in-building cellular connections
Proximity-based apps (for instance using Bluetooth) will risk false-positives if they are not integrated with building location and indoor-mapping systems. You can safely stay 2 metres from someone infected, if there is a wall or floor/ceiling between you.
IoT systems such as disinfectant robots will also need access to indoor maps and granular positioning technology.
Next-generation networks such as private/campus 5G and also recent Wi-Fi meshes have improved wireless-positioning abilities. This could allow both real-time and reported proximity-monitoring - as well as enabling remote working & even "lights out" full automation in industrial settings
Both Wi-Fi and cellular networks can work out how many devices/users are not just connected, but detected, even if they do not attempt - or are not permitted - to connect to a given system. That could yield good data on user-density, especially if they are personal devices such as smartphones.
Wi-Fi enhancements already enable motion-detection - which can be considerably more accurate than traditional infra-red, and also work through walls. One technology innovator here is Cognitive Systems (link) but there are others as well. I've also seen suggestions that future 5G variants may be able to do something similar, if deployed with small cells. (I'm not sure how it would work with other in-building shared networks, though).
Potential to use localised cell-broadcast messaging, or Wi-Fi hotspot captive-portal pages, to distribute public health information and advice
There may be a growing need to align the indoor wireless network(s) with nearby outdoors connectivity, or link multiple buildings together well. Campus networks are already growing in importance for multiple reasons (link) and social-distancing and control adds another set of use-cases. (Consider private/public spaces such as courtyards, rooftop bars, parking lots and so on).
The use of virtualised radio networks (or specific variants such as OpenRAN) could also prove valuable here - for instance to enable operators to scale up/down capacity dedicated to indoor 5G wireless systems, or switch radio VNFs between indoor and outdoor coverage. (This goes far beyond pandemic-proofing and I will write about it another time).
Neutral-host indoor wireless systems will be able to onboard new tenant networks such as public safety, or private building management networks, depending on future requirements and spectrum licensing policy.
There may be edge-computing requirements driven by pandemic-proofting, although that doesn't necessarily imply either on-prem or very granular nearby edge facilities, rather than metro-level.
This is still just a very rough draft of my ideas - and clearly there are various policy / regulatory hypotheses here as well as technology direction. I'm not a specialist on building regulations, so it's quite possible I've made unreasonable assumptions. But this is intended as the start of a discussion, rather than a definitive forecast. I expect this topic and more detailed discussion to surface in coming months and years.
Your comments are very welcome - and if you want to get in touch with me directly, please connect on my LinkedIn, or send me a Twitter DM. If you're hosting any webinars, or holding internal brainstorms on this, I'd be very interested in participating.
This Barclays-funded bobbins by Farage’s former press officer is remarkable for one reason. It never once acknowledges that the man whose alleged ‘leadership’ is being praised personally & directly appointed every single minister being criticised. Maybe the author didn’t realise. pic.twitter.com/JvReGhz2ps
Radio.garden is an amazing and fun discovery, perfect for infinite distraction during life in quarantine. (James Vincent in The Vergecalls it “Google Earth for Radio.”) Here’s a list of just some discoveries I’ve made while mining that Earth with Shazam open on my phone:
CIAU/103.1 in … not sure where this is, except in the vast nowhere east of Hudson Bay. Just played Rock’n Me, by Steve Miller. Now it’s Light my fire by the Doors.
Chanso Du Berceau, by Georg Gabler on (can’t say, it’s in Cyrillic), in Plotina, Russia.
Magic, by One Direction, on FM Trölli, somewhere in Iceland.
No More sad Songs, by Little Mix Feat. Machine Gun Kelly on Ice FM, Nuuk, Greenland.
Espère, by Joe Bel, on CFRT/107.3 in Iqaluit, Nunavuk.
Everything played on CJUC/92.5, Community Radio in Whitehorse, Yukon. My fave by far. Just put it on my Sonos.
If I can’t Have You, by Etta James, and now Got My Mojo Working, by Muddy Waters on kohala Radio.
KNKR/96.1 on the Big Island somewhere. Also liking Kaua’i Community Radio KKCR/90.9 in Hanalei. Alas, Shazam knows nothing they play, it seems.
Another thing Shazam doesn’t know, on Radio Kiribati AM 1440 in Tarawa.
Walking on a Dream, by Empire of the Sun, on Cruize FM 105.2 in New Plymouth, New Zealand.
Some kind of bottleneck slide guitar, with a guy playing “My baby says she loves me.” On Spellbound Radio FM 106.8 in Gisbourne, NZ. Followed by Ry Cooder’s One Meatball.
One Fine Day, by the Chiffons on 101.5 Moreton Bay’s Own, Moreton Bay, Australia.
Tie a Yellow Ribbon Round the Old Oak Tree by Dawn, followed by Woo Hoo, by the Rock-aTeens, on 88.9 Richmond Valley Radio, Far North Coast, New South Wales, Australia.
You Got To Me, by the Wolfe Brothers, on Ten FM in Tenterfield, Australia
Liar Cry, by Pigram Brothers on 2Cuz FM 107.7 in Bourke, running 99fm, in Brisbane I think.
Winds of Change by Airborne, on The Lounge FM 106.3 in Port Douglas, Australia.
Adies Meres Adies Nihtes, by Christina Maragozi, on Radio Vereniki 89.5 lerapetra, Crete.
Per Tu (Joan), by Amadeu Casas, on Formentera Ràdio, El Pilar de la Mola, Spain.
Eu Gosto De Ti, by Elas, on Rádio Graciosa FM 107.9, Santa Cruz Da Graciosa, Azores.
Hm. I had some from South America and then WWOZ in New Orleans, but those disappeared. Grr.
Souly Creole, by Joe Sample, on The Jazz Groove in San Francisco.
Nothing Else Mattrs, by Metallica, on Radio 1 100.0 in Papa’ete, Tahiti.
Some Girls, b Racey, on 88 FM in Avarua Distrct, Cook Islands. The voices are clearly from Australia.
I know you, by Craig David Feat. Basille, on Отличное Радио in Birobidzhan, Russia.
I remember, by Claude Diniel, from Radio Trassa, Blagoveshchensk, Russia.
So Good to Me (Extended Mix), by Chris Malinchak, on Radio STV in Yatusk, Russia.
Tusi Sam, by Mari Kraymbreri, on Radio Sigma in Novy Urengoy, Russia
Одинокая Луна by Артём Качер on Sever FM in Naryan-Mar, Russia. Followed by If I’m Lucky, by Jeson Derulo.
I wanna Sex You Up, by Color Me Badd, on SAMS in Jamestown, Saint Helena.
I Go Alone, by Stephen clair and the Pushbacks, on Jive Radio KJIV Madras Oregon.
Jungle Love, by the Stever Miller Band, on WOYS FM 100.5 Oyster Radio, Apalachicola FL, United States (This follows a very long invitation to please not visit “the forgotten coast” now, because everything is closed.)
Angie McMahon on KMXT-FM 100.1, Kodiak AK, United States, playing NPR’s World Café
Everything through #21 was on Monday, April 13, during which I learned some things, such as copying and pasting station names and locations from the lower right panel there. The rest were listed today, a few minutes before I posted this.
Most of the stations here are in very very outlying places, which are easiest to find and grab.
I could go on (it’s very tempting… for example noting now much English-language music is all over extremely rural Russian radio). I could also go back and stick some links in there. But I’ll leave the rest up to you. Have fun.
And big thanks to @ccarfi, who turned me on to this thing.
Steffani Cameron, Social-Distancer@SnarkySteff
Omigod, guys. This is so good. Wine-braised boneless beef short ribs and carrots with mashed potatoes, garlic Swiss… twitter.com/i/web/status/1…
We held our second “Pen Night in the Clouds” tonight, in Zoom. We maxed out at 15 participants: 14 with video, and one person who dialed in by telephone.
For a Zoom meeting it was pleasantly non-chaotic, as was last month’s inaugural edition.
We had nothing on the agenda, which was convenient because “everyone introduce yourself and tell us what you’ve been up to” took 90 minutes. Lots of pen and ink and paper talk, but useful for many of us, I think, as a social-isolation-buster.
Some things for me to follow up on:
Step Forward-brand tree-free paper was highly reviewed.
Melt wax (candle or sealing wax) into a Chapstick tube for a handy “prevent envelopes addresses with a fountain pen from running if they get wet” tool.
Our next meeting, again on Zoom, will be May 23, 2020, at an earlier start time of 4:00 p.m. Atlantic to provide interested Europeans with an easier way of attending.
Recientemente I wrote a Twitter thread on the value of annotated bibliographies, which I wanted to turn into a blog post because I wanted to really strengthen my argument. As most of you who read my blog post know, I HAVE written about annotated bibliographies, literature reviews, systematic reviews, etc.
In my view, an annotated bibliography, a literature review and a systematic review are three important exercises in systematizing all the material we read. In my view, ABs, LRs and SRs have increasing degrees of complexity, as I show below.
Source: My own construction.
This thread started at the request of one of my students, as the tweets below show.
He was looking for a way to show everything he had read. Personally, I don’t think a student should write a literature review BEFORE writing an annotated bibliography. Allow me to explain why. https://t.co/IhXJhBALzo
A literature review requires the student to digest it all.
For me, producing an annotated bibliography and/or gnerating a bank (or set) of rhetorical precis (or a bank of article and book chapter summaries written in index cards) is the prerequisite step BEFORE writing the literature review.Often times, my students simply prefer to show me their Excel Dump.
IF a student feels more comfortable with a more detailed document than CSED, then I recommend that they do a series of rhetorical precis and THEN dump those notes either in a Word document, or that they write an AB.
Personally, I very strongly believe that Annotated Bibliographies are legitimate scholarly outputs. For one of my projects on water conflicts, I asked one of my research assistants to produce an annotated bibliography on the sociological concept of “framing” and “Frame Analysis”
… as a public service. I am overwhemed with deadlines, but I think that over the summer, I am going to produce an Annotated Bibliography on Comparative Ethnography.
What’s the value of ABs for other scholars? Anybody who downloads and reads my AB now has a STARTING POINT.
Anybody who reads my AB on Comparative Ethnography can say “oh wow, there’s X author and Y author, and Z book and W chapter, and those are probably the first ones I should read”. It saves time for other scholars (including my own students).
I love the journal WIRES Water (Wiley Interdisciplinary Reviews) because that’s what they publish, exactly (systematic literature reviews). So does Geography Compass. So does Progress in Human Geography. So do the Annual Reviews of Political Science, etc. https://t.co/6ODRV09ibb
The intermediate step between writing an annotated bibliography (or building a deck of index cards on a topic, or a bank of rhetorical precis, or a set of Synthetic Notes) and producing the Literature Review is the DIGESTION, ANALYSIS AND SYNTHESIS of everything they’ve read.
“… in their CSED or annotated bibliography” and writing the literature review.
I have to finish 3 journal articles literally within the next week so I may not get to write that blog post (from AP to LR) soon, but at least this thread should help people make the transition.
Overall, I strongly believe in the value of teaching our students to do Annotated Bibliographies, THEN Literature Reviews, THEN Systematic Reviews. My website has resources on each one of these, and the links above in the tweets shown should also work. Hopefully this post will be of use to instructors and students alike!
Loving Mohit Bhoite's circuit sculptures. where resistors, LEDs and brass rods take on structural elements within the circuits to beautiful effect. Just gorgeous.
"COBOL is often a source of amusement for programmers because it is seen as old, verbose, clunky, and difficult to maintain. And it’s often the case that people making the jokes have never actually written any COBOL. We plan to give them a chance: COBOL can now be used to write code for Cloudflare’s serverless platform Workers."
Not an April Fool; instead, a deep dive for newcomers to COBOL, a platform to make it on, and some movie trivia. Great blogging all around.
Flickr Galleries help you curate your favorite public photos and videos from other Flickr members into a virtual showcase. Unlike Faves, Galleries can be organized around a theme, an idea, a visual style, and pretty much any other way you can think of.
Looking for photo inspiration for your next trip after quarantine is over? Create a gallery like this one, inspired by the USA West Coast! See an exciting topic you like? Try something fun, like this street photography gallery. In need for more inspiration? Explore this page of galleries curated by Flickr staff.
Introduced for the first time in 2009 and redesigned for better UX in 2018, Flickr Galleries are a great way to engage with other members of the Flickr community and to show them you appreciate their work. They are also an excellent way for photographers to stand out as talented curators and get some additional exposure. So here are a few tips to help you make the most out of this beloved Flickr feature:
Choose a theme and title for your gallery.
Look for a cohesive theme to make your gallery feel more like a visual story and less like a random collection of photos you like. Common themes include inspiring images around a particular topic or event (birds, architecture, supermoon shots), bucket list locations, highlights from a Flickr Group challenge, best of Explore, etc.
Give context.
When you add a new image to a gallery, the photo owner gets notified, which might lead to some new connections and comments. Whenever you create a new gallery, make sure to add a description to it. Gallery names and descriptions help your followers understand what your gallery is all about.
Personalize the look and feel.
Update your cover photo at any time by selecting an image from the gallery. You can order the photos in whatever way you like and as often as you wish. You have the ability to add up to a maximum of 500 public photos, as long as the photo owner has selected the setting that allows their work to be added to galleries.
Share.
Share your curated selections with your nearest and dearest. You can share galleries outside of Flickr with our intuitive sharing options that will improve the presentation of your galleries on other social networks.
If you’re ready to get started but want to learn more, check out this Galleries Overview article in our Help Center and learn how to create, add, and delete a gallery on Flickr. We’re also looking to promote interesting galleries curated by Flickr members in our next Flickr Newsletter! If you have a gallery you’re particularly proud of please add the link to this discussion thread in our Flickr at Home group.
Over the years I moved from QuickSilver to Spotlight (avoiding Alfred because I saw no real added value in it), and am now at a point where Spotlight is slow as molasses, barely usable during the first half hour after system boot, and, in general, far less effective than it was in the very beginning.
Having a really fast launcher on Windows would be another point in its favor from a productivity standpoint, and I have been bemusedly watching how the Cortana search bar has followed mostly the same trajectory as Spotlight in terms of bloat and unresponsiveness. So it is something to actually look forward to.
My suggestion is to call it Searchlight, in honor of the little flashlight icon that popped up as a spinner back in Windows 95…
Factor is a programming language I've written about before and in the early days of Factor development I wrote a number of libraries and contributed to development. It's been a while since I've contributed but I still use Factor. The development environment has a very Smalltalk-like feel to it and it includes full documentation and browseable source code of libraries.
This post isn't about Factor the language, but is about some of the neat fun libraries that people have written that shows off the graphical development system a bit.
Minesweeper
The first example is an implementation of the game Minesweeper in Factor. A blog post by the author explains the implementation. To run it inside Factor, do the following:
"minesweeper" run
A new window will open showing the game. Help can be shown with:
If it seems like all the examples I'm using are from the excellent re-factor blog - well, most of them are. This blog post from re-factor shows pulling historical facts from Wikipedia:
USE: wikipedia
USE: calendar
today historical-events.
...a list of historical events from wikipedia for today...
yesterday historical-births.
...a list of historical births from wikipedia for yesterday...
5 weeks ago historical-deaths.
...a list of historical deaths from wikipedia for five weeks ago...
The items in the list are graphical elements that can be manipulated. Left clicking on the coloured words will open a URL in the default web browser. Right clicking allows you to push the element on the Factor stack and manipulate it.
The calendar vocab has a lot of interesting words that allow doing calculations like "5 weeks ago".
Hacker News
There's a hacker-news vocabulary that provides words to list current articles on the Hacker News website. Like the previous Wikipedia example, the graphical elements are clickable objects:
USE: hacker-news
hacker-news-top.
...list of top articles...
hacker-news-show.
...list articles related to showing projects...
CPU 8080 Emulator
A number of years ago I wrote a CPU 8080 emulator in Factor and used this to implement a Space Invaders Emulator and then emulators for a couple of other 8080 arcade games, Balloon Bomber and Lunar Rescue. These examples require the original arcade ROMs and instructions for using them are in the online help:
"rom.space-invaders" run
...opens in a new window...
"rom.balloon-bomber" run
...opens in a new window...
"rom.lunar-rescue" run
...opens in a new window...
"rom.space-invaders" help
...displays help...
Gopher Implementation
Another magical implementation from the re-factor blog, a Gopher server and a graphical Gopher Client. This is a video I made of the gopher client on YouTube:
I also did a video that shows some Factor development tools on the running Gopher client to show how everything is live in Factor:
And More
There's much more buried inside Factor. The list of articles and list of vocabularies from the online help is a good way to explore. This help system is also available offline in a Factor install. By default many libraries aren't loaded when Factor starts but you can force loading everything using load-all:
load-all
...all vocabs are loaded - prepare to wait for a while...
save
...saves the image so when factor is restarted the vocabs remain loaded...
The benefit of doing this while developing is all the online help, source and words are available via the inbuilt tools like "apropos", "usage", etc.
Ich wage mal eine Vorhersage, was ich selten mache. Alles wird anders. Es gibt ein Vorher und ein Nachher. Wir werden nicht einfach in unser altes Dasein zurückfallen. Und das macht mir Hoffung, denn wir waren auf dem Weg, die Lebensgrundlagen der nächsten Generationen zu zerstören.
Die Phase zwischen vorher und nachher wird so lang, dass wir in dieser Zeit lernen können, unser Leben besser zu gestalten.
On Tuesday this week, awlnx and Krombel of Freifunk München (FFMUC) gave a virtual presentation online about their public Jitsi Server cluster installation and how they have set up the system. Currently, it serves well over 400 simultaneous participants in the evenings in 70 concurrent rooms. And with their 20 video bridge servers they say that they are prepared for over a thousand simultaneous users!
jwz.
srfeld:
Das Kraftfuttermischwerk.
mrjamesob on Twitter.
wtyppod on Twitter.
internetofshit on Twitter.
