Souk's Notes - Unofficial Summaries

2021-08-06 · open the original ↗

• Build extensions on PJSIP, never chan_sip. FreePBX now labels chan_sip "Legacy"; the host says it is deprecated and that newer Asterisk versions do not even compile it in without a rebuild. PJSIP runs on port 5060, and an extension does not exist for a phone to register against until Apply Config is clicked.
• Choose extension numbers that cannot collide with feature codes. The host wants at least three digits and none starting with 1: extensions such as 101 collide with feature codes: *1 toggles on-demand recording, while transferring to someone's voicemail is * followed by their extension, so the system cannot tell which is meant. He starts at 200.
• Set an emergency caller ID for anyone working away from the office. E911 must send the phone's actual physical location, and a home worker dialing 911 with the office caller ID would send responders to the office, possibly in another state. Outbound caller ID, by contrast, can be left blank to inherit the trunk or route default.
• Voicemail defaults trade convenience against security and lost messages. A PIN equal to the extension forces the setup wizard on first access, useful when an extension passes to a new employee; "require from same extension" makes *97 ask for a PIN. Deleting voicemail after emailing it is tidy, but FreePBX deletes it even if the email never arrives.
• Raise max contacts for multiple devices, and bulk-load by CSV. The default of one contact allows a single device per extension; he raises it to five so desk phones and softphones ring together. On-demand recording (*1) is off per extension by default, and above about 10 extensions Crosstalk uses Bulk Handler CSV import. Sponsor voip.ms is one of three SIP providers Crosstalk recommends to clients.

2021-08-04 · open the original ↗

• The UDM Pro's WAN must hold a routable public IP for client VPN. The host says the ISP connection should be in bridge mode so the real static or dynamic address sits on the UDM Pro's outside interface, avoiding double NAT; otherwise you must port-forward through to it. With a dynamic address, he suggests a dynamic DNS name in the client instead of the IP.
• The series covers only L2TP client VPN, the sole remote-access option offered. Site-to-site VPN is left out because UniFi-to-UniFi, EdgeRouter or Cisco variants all differ. Setup is a new network set to remote access, with a pre-shared key, the hosting WAN, the default RADIUS profile and a user list; his VPN subnet was 192.168.2.0/24.
• Sending all traffic through the tunnel is the simple default. On the iPhone he leaves "send all traffic" on, because sending only UDM Pro-bound traffic needs extra routing rules. Over LTE the phone took 192.168.2.6, showed his static WAN address ending .29 as its public IP, and reached the UDM Pro by direct rather than cloud connection.
• The default client VPN is fairly wide open. He notes that firewall rules are needed to stop VPN users from seeing other networks, beyond this basic setup. The video carries a sponsor segment for voip.ms, one of three SIP providers Crosstalk recommends to clients.

2021-08-04 · open the original ↗

• Companies open-source code for outside ideas, reuse and fresh eyes. Faris, technical community manager at Rocket Mortgage, says open projects draw innovative ideas and stop teams reinventing the wheel internally, while outside users catch missing features and bugs, so it is "mutually beneficial". She says Rocket surveyed its technology division on attitudes to open source, interviewed staff who proposed projects, and is now working out how to release and promote them under a regulated financial company's restrictions.
• Running an open source project is mainly managing communication. At her previous job, a Python-based SecOps/DevOps project with a paid enterprise version, she was hired because open source contributors felt unheard. She describes triage and testing teams, engineers answering contributor questions on Slack, weekly meetings and Twitch streams walking people through testing and triage, and a wiki of getting-started information; managing the community "is not really about the code".
• Contributors need not be developers; documentation is the usual gap. Faris says the biggest complaint about every open source project is poor documentation, which Scrum Masters and BAs she worked with have contributed to, and which she says pushes would-be contributors away. For finding projects she suggests GitHub search and company open source pages such as Spotify's and Netflix's, then reading the readme, wiki and code of conduct before picking up issues.
• Rocket open-sources only what reveals none of its "secret sauce". Faris says anything released must clear legal and public relations, and the company uses the MIT license. Its Spark Design System, a consistent UI across company websites worked on by developer advocate Chris DeMars, qualifies because it is design-only; she also mentions a podcast management platform and analytics and test-reporting tools.
• To launch your own project, start with people rather than resources. Her step one is putting it on GitHub, then asking individuals you know to contribute and sharing it on social media; even without a broad network, a co-worker on a lunch break is a start. She adds that open source contributes to learning and to jobs that weigh contributions, and that she has found the community mostly friendly and collaborative, while acknowledging not everyone's experience has been positive.

2021-08-03 · open the original ↗

• The UDM Pro offers dual-WAN failover only, not load balancing. The load-balancing setting has just one choice, "failover only", which the host reads as a hopeful sign that Ubiquiti may add balancing later in software.
• The second WAN plugs into the SFP+ port. That port is 10 Gbps; he used a plain SFP module with LC-to-LC fiber because he had no spare SFP+ modules, and notes a DAC cable or an RJ45 SFP module also works. WAN2's advanced settings cover a separate DNS server, an ISP VLAN ID, MAC cloning, smart queues, static addressing and expected speeds.
• Failover and failback each dropped only a ping or two. Pinging 1.1.1.1 at about 14 ms, pulling WAN1 produced two timeouts before traffic resumed on WAN2, and reconnecting it switched back within a few seconds with one drop. His lab used NAT'd addresses on both WANs; in production they would ideally be routable.
• The interface misreports which WAN is active during failover. Rather than flagging WAN2 as in use, the dashboard showed WAN2's address under the primary WAN, which the host calls misleading.

2021-08-02 · open the original ↗

• The new camera is a ceiling-hung fisheye with digital pan-tilt-zoom. Ubiquiti's announcement video presents the AI 360 as a 360-degree digital pan-tilt-zoom camera with enhanced person and vehicle detection, microphone and speaker, infrared, PoE power and two interchangeable bubble lenses. $299 is the early-access price, which the host notes may change; he has not used one and says early-access status would bar him commenting on its function anyway.
• Leaving it zoomed in could miss events elsewhere in the room. No Ubiquiti AI camera yet auto-zooms onto detected people, the host says, so a camera left in PTZ mode records only where it points. He would like it to revert to 360 view after inactivity, or always record 360 and allow PTZ on playback, but does not know whether it does.
• He doubts the speaker and the second lens are useful. Previous Ubiquiti two-way cameras had weak speakers, and this one would sit 8 to 15 feet up or more. He cannot see why a non-weatherproof lens is included when the tamper- and weather-resistant one would seem good enough, though suspects the larger clear dome may affect the view.
• Comparable 360 cameras all cost more. The AI 360 is 5 MP and 1920x1920 in 360 mode; the Axis M3067-P is 6 MP, 2016x2016 in 360 mode and up to 60 fps for $469. Hikvision's 6 MP models run $380 to $550 at 30 fps, and a 5 MP Dahua with IP67 and IK10 ratings costs just over $400, each before any NVR.
• Protect's polish is the trade for proprietary hardware. Compared with NVRs such as Blue Iris, which he runs on a test server, the host finds UniFi Protect a more seamless experience. The cost is that it cannot be installed on your own server, which he says is also what lets Ubiquiti control and refine it.

2021-08-02 · open the original ↗

• F# aims at simplicity without trading away correctness or speed. Syme, F#'s designer and architect, argues that only strongly typed functional languages deliver succinctness, robustness and performance together: in his framing Python gives succinctness without the other two, and C# gives robustness and performance but not succinctness. He plays down the "functional" label; "functional first" means only data plus functions that transform it, with immutable defaults.
• Correctness comes from code that stays tight under change. Syme calls nullness "absolutely a disaster" for correct code, and contrasts C# silently discarding an unused result with F# requiring an explicit ignore. Records, discriminated unions and tuples force every producer and consumer to be updated when the domain model grows; he cites anecdotal enterprise deliveries with very low or zero bug counts. Hanselman relays Yaron Minsky's line that F#'s type system makes "illegal states unrepresentable", which Syme endorses, recommending Scott Wlaschin's Domain Modeling Made Functional.
• Immutable by default, mutable wherever performance needs it. Syme says most web, UI and data-pipeline bookkeeping suits immutable data, while cores such as one huge machine-learning array can drop to mutation. F# 6, then in planning for .NET 6, adds task { } built on a general "resumable code" feature, compiling to essentially the same high-performance, debuggable code as C# async; F# 5 added struct records and other struct work.
• F#'s async model came first and inspired C#'s and Python's. Syme credits F# async with inspiring async/await in C#, then Python and others, but concedes the early F# implementation had performance and debugging issues that C# ironed out, which the new task support addresses. Resumable code also covers async sequences, which he says F# can do in a library where the C# team must build each into the compiler.
• F# is pitched as a cloud and JavaScript language, not only .NET. Syme says .NET for Apache Spark has "solved the big data problem for .NET", a space Java dominated, and points to F# templates for Azure Functions. Through the community's Fable and WebSharper compilers F# targets JavaScript and npm, playing TypeScript's role; at .NET Conf Focus Day, recorded before 29 July 2021, Guido van Rossum used it on both sides and quickly grasped list expressions, which Syme says subsume three or four C# features.

2021-07-22 · open the original ↗

• FreePBX ships with no default backup, so you must build one. The host calls Sangoma's removal of default backups astounding; setup takes two parts, a File Store location (local, FTP, Dropbox, SSH or Amazon S3) and then a job in the Backup and Restore module. The SSH store exists for primary/warm-spare setups, which Crosstalk offers to set up for clients.
• A local backup alone is not enough for production. If the drive crashes, the local copies go with it. For both hosted and on-premises clients, Crosstalk keeps a local backup plus an off-site FTP copy on its own servers, and keeps three monthly and three weekly backups.
• Path variables take double underscores, and backup names should become subfolders. The local store points at the spool variable plus /backup (/var/spool/asterisk/backup). Turning on "append backup name as a directory" keeps monthly and weekly jobs in separate subfolders. Retention is set by "delete after runs" (3 keeps three months; 0 never deletes) or by age in days.
• Email on both outcomes at first, then switch to failure only. Inline logs put the backup log in the email body. Once backups have proven reliable, the host only wants to hear about problems. Leave all modules selected unless you know otherwise; add custom folders such as custom sounds or TFTP phone configs and firmware if needed.
• The rewritten module restores backups made on older versions. The module was rewritten in FreePBX 15; earlier releases could only restore onto the same version; the host says cross-version restores from 12, 13 or 14 work pretty well. His near-empty test server produced a 53 MB tar.gz, and backups in the configured File Store appear automatically under the Restore tab.

2021-07-20 · open the original ↗

• Gaudet owns the publishing process, not the technology behind it. She works on Microsoft's Commercial Marketplace enablement team (formerly operations) covering infrastructure offers — virtual machines, Azure Apps (solution templates, a subset of ARM templates) and container images from Azure Container Registry, with IoT Edge about to be added — and owns metadata certification, checking titles, descriptions and links against hundreds of policies.
• Publishing runs through Partner Center and two parallel certification workflows. She describes getting an MPN ID, being vetted (a physical business location is required, as she remembers it), choosing Commercial Marketplace and one of 16 offer types, entering metadata and uploading; after automated checks such as malware scanning, content and technical validation run in parallel, about 95% automated and taking a couple of hours, and a passing offer goes to preview and is in Marketplace within 24 hours. Sovereign clouds such as China are handled separately, and VM extensions still have to be submitted by email.
• Certification is being automated, but multi-image offers made it painful. Until about six months before recording, VM certification was manual, and a publisher with 20 to 100 images on one offer had to publish them one at a time while she chased the certification team for each approval; VMs now go through in a couple of hours, and the aim is to automate as much certification as possible within the next year. Subjective checks remain, such as blurry logos, which she attributes to one upload being resized into three logo sizes without a set DPI.
• Listings must not read as if the product is Microsoft's. She says "monkey application on Windows" passes where "monkey application Windows 8" fails, for Microsoft IP and legal reasons, and logo use is restricted; the publisher agreement lets Microsoft remove insecure or policy-violating offers, and apps may not collect customer data without telling the customer. Her closing advice is to read all the policies and terms before starting.
• SolarWinds showed how much the process depended on her. She spent a weekend on the incident checking whether any Marketplace offer was affected and waking people across teams because she lacked access to parts of the infrastructure; the offer in question was not affected because its publisher had fixed it over the summer. She wants a process that works without her as the bottleneck, and asks publishers to use Marketplace support (aka.ms/marketplacesupport) rather than email her, since her inbox starts at about 300 messages.

2021-07-20 · open the original ↗

• IoT devices belong on their own network, away from the main LAN. They phone home to vendor clouds and could be compromised, so isolating them both controls what reaches out and protects the computers holding photos and documents. The host puts every IoT device in his house there, and suggests a separate network for cameras on the same pattern.
• Firewall rules needed the classic UniFi interface at recording time. The IoT network is a corporate network (not guest) on VLAN 20, 192.168.20.0/24, with DHCP widened to .2–.254 and DNS 1.1.1.1 and 9.9.9.9. Its SSID is WPA2 only, since most IoT devices cannot do WPA3, and optionally 2.4 GHz only to spare the 5 GHz band.
• Ordered LAN In rules block inter-VLAN traffic but spare the LAN. First allow established/related; then allow the main LAN to an RFC 1918 address group (192.168.0.0/16, 172.16.0.0/12, 10.0.0.0/8); then drop RFC 1918 to RFC 1918. An existing ping kept working until the 30-second ICMP state timeout expired, because of the established/related rule.
• LAN In rules do not stop access to the gateway itself. The IoT network could still ping and open the UDM Pro's UniFi OS at 192.168.1.1. LAN Local drop rules block IoT and guest networks from the other networks' gateway addresses, using per-network address groups.
• Block only the management ports on a network's own gateway. Blocking 192.168.20.1 outright would kill IoT internet access, so the host drops IoT traffic to an all-gateways group on a port group of 80, 443 and 22. He leaves ICMP open for testing; allowing access to a Pi-hole on the main LAN is left out of scope.

2021-07-20 · open the original ↗

• A simple guest network works entirely in the new UniFi interface. The host creates it on VLAN 10 as 192.168.10.0/24, turning off auto-scale network so he can set the subnet and start DHCP at .2. Optional manual DNS (1.1.1.1 and Quad9, or a subscribed service such as OpenDNS) adds content filtering for guests.
• Device isolation is what actually walls guests off. Before it was enabled, his phone on the guest Wi-Fi could ping his main-LAN computer at 192.168.1.144 and the UDM Pro at 192.168.1.1. Enabling it adds a restricted-access list of all RFC 1918 ranges, after which guests reached only the internet, not even their own gateway at 192.168.10.1.
• UniFi bandwidth profiles cap each wireless client, not the network. A "crosstalk guest" profile of 10 Mbps down and 10 Mbps up applies per client on the guest SSID and does not affect wired clients; a speed test showed the cap working. The host wishes UniFi offered an aggregate cap for the whole guest network instead.
• Businesses should schedule guest Wi-Fi to business hours. He recommends it so nobody parked outside can use the connection after hours, especially for passwordless guest networks. Security is set to WPA2/WPA3 mixed mode because his access point is Wi-Fi 6 capable.

2021-07-15 · open the original ↗

• A well-designed tech pouch is the bag's best item. The host's Peak Design Tech Pouch is water resistant, with hand straps underneath for one-handed access and an internal pass-through so a power bank inside can charge a phone in the side pocket. He used that more than expected on his first business trip in over a year, a training in Appleton, Wisconsin.
• He loves one travel router and refuses to recommend another. His GL.iNet Slate rides in a hard case made for the TP-Link AC750 travel router, which he calls absolutely horrendous; he scrapped a planned review of it. Retractable Ethernet cables (a two-pack for under $10) and tough nylon USB cables suit a bag that gets knocked around.
• One brand supplies all his power banks and chargers. It is Anker: the PowerCore 10000 costs $21.99 and the 20000 mAh PowerCore Essential $42.97; the 20 W Nano charger is $16.99, and the 60 W PowerPort Atom III is $37.99 in white or $39.99 in black, as of recording. He says the banks charged his iPad without trouble.
• Bring a streaming stick, plus a flexible HDMI jumper. Both the Amazon Fire TV Stick ($47.49) and the Roku Streaming Stick+ ($49) got past hotel Wi-Fi captive portals and keep your logged-in apps; he prefers Roku from familiarity. Some hotel TVs physically cannot take a stick in the HDMI port, which has happened to him.

2021-07-14 · open the original ↗

• D3 generates interactive SVGs and does the placement math for you. Dowden explains that SVG is function-based rather than raster, so it scales without limit and can carry clickable, animated elements. D3 is a JavaScript library for building SVGs on the fly. It runs anywhere JavaScript does, including an Angular project she is working on.
• D3 is usually overkill but pays off for complex visualizations. For ordinary charts she finds Charts.js perfectly fine. For a tabletop-RPG project she gives D3 a list of nodes and connections, and it draws a "spider web" of rooms, traps, enemies and treasure without overlaps; she swapped its circles for her brand's hexagons. That took "an afternoon of fiddling," against a lot of work without D3. She also built a spinning globe of the countries she has visited and spoken in from a JSON map of the world.
• SVG's fixed positions make data refresh an implementation decision. Unlike HTML, an added element does not push its neighbours aside, so the developer must work out how much to recalculate. She redraws her whole node web on refresh. In a nested-circle view (castle, levels, rooms) she redraws only the circle that changed.
• D3 wants pre-shaped JSON in memory; data volume governs speed. Her Angular services turn Firebase data into the JSON a given D3 function expects, much like old jQuery UI plugins. She expects about 5,000 nodes to lag, because rendering is client-side and the DOM has limits. Speed can also come from tighter data transformation, such as looping once instead of three times. She computes her statistics before handing the data to D3, though she does not say D3 cannot do it.
• Getting started is importing D3 and writing DOM-style JavaScript. It is available from npm or a CDN, and using it feels like "back to the good old DOM manipulation days." She recommends the D3 Graph Gallery for examples with their code, and warns that it assumes you already know JavaScript well.

2021-07-13 · open the original ↗

• Several FreePBX services need a valid certificate on a real FQDN. The host lists WebRTC, Sangoma's Zulu and Connect, and Clearly Anywhere. He uses Let's Encrypt because it is built into FreePBX, though your own certificates also work.
• Move the plain-HTTP admin GUI so Let's Encrypt can take its port. He moves plain-HTTP admin to 8080 (which can then be kept closed at the firewall) and assigns port 80 to Let's Encrypt. He argues port 80 open this way is not a risk: it serves only the files Let's Encrypt checks to verify the server.
• FreePBX's firewall opens the challenge port only during renewals. The responsive Let's Encrypt rules allow temporary access to the ACME challenge folder during an update. Let's Encrypt always challenges on port 80, so requests fail if an upstream firewall blocks it without redirecting; a forward can stay open permanently.
• Generating the certificate is not enough; install it. After creating it under Certificate Management, where it renews automatically about every two months, make it the default. Then select it under System Admin's HTTPS Setup and click install before the server actually uses it.

2021-07-13 · open the original ↗

• FreePBX has two update types, each runnable from GUI or SSH. System updates patch the base Linux OS; module updates patch the FreePBX software. The host's scheduler applies only security updates automatically and leaves everything else notification-only.
• System updates can take the PBX offline; module updates usually do not. In the GUI his check found 18 system packages, and module updates are typically safe to run in production.
• The host prefers SSH because the GUI handles dependencies poorly. He says the GUI gives little feedback and does not resolve dependencies between modules, while the command line does. Over SSH as root, yum update -y handles system packages (under a minute for him) and fwconsole ma upgradeall handles modules (about 11 minutes for 66).
• fwconsole restart drops calls, so use a maintenance window. He restarts services after a major update, which brings Asterisk down and disconnects calls in progress. Re-running both commands should then report nothing to update, followed by Apply Config in the dashboard.

2021-07-09 · open the original ↗

• High code coverage does not prove the tests are good. Bock, a developer advocate at Rocket Mortgage, notes that tests with no assertions still raise coverage. Good tests naturally produce high coverage, but the reverse does not hold. Mutation testing answers "who watches the watchers": it alters the code, for example changing an add's plus to a minus, and expects some test to fail. Each failure "kills the mutant." He says the idea dates from the 1960s or 70s and spread only recently because it is computationally expensive.
• Stryker automates mutation testing; installing it and running dotnet stryker is enough. It is open source and available for .NET, Scala and JavaScript; Giard found its site, stryker-mutator.io, and that the .NET version arrived in 2018. Bock believes it mutates the source in memory rather than the bytecode. A run is slow, like BenchmarkDotNet, and ends in a report that links to each file and marks which mutants survived. He asks companies that use open-source tools like it to contribute back.
• A surviving mutant flags a possible problem; a human must decide. Stryker applies a fixed list of mutations, including arithmetic operators, string literals and LINQ (Take becomes Skip). In his own code, changing < to <= broke no test, and he judged that acceptable. Blanking an exception message also broke no test. He changed his tests to assert the message, on the principle that code exists for a reason and should be tested.
• Adopt Stryker early, not years into a codebase. He sympathizes with teams under deadline pressure. Adding it six months to two years in can be overwhelming, so he advises bringing it in as soon as code is written. A Rocket colleague is looking at building it into their continuous integration.
• A fresh explanation can make a familiar concept click. Bock had heard of mutation testing for years, but it only clicked in a lightning talk on Stryker at a group run by Dan Clark in Oxford, as mocking had once clicked for him through one framework. So he urges people to blog or record videos on topics others have already covered.

2021-07-09 · open the original ↗

• A looping system automates what a type 1 diabetic otherwise does by hand. Withee's device takes the last 15 minutes or so of readings from her continuous glucose monitor, which reports every five minutes over Bluetooth, factors in recent food and insulin, predicts whether she will need more or less, and sets temporary rates on her insulin pump over the 400 or 900 MHz bands. She says it still needs occasional intervention but cuts a constant, mentally taxing workload of carb counting and dose correction down to a periodic glance at her phone.
• The open-source stack costs almost nothing beyond the device. She runs OpenAPS (github.com/openaps) on an Intel Edison, or alternatively a Raspberry Pi Zero, with an open-source Explorer board, running Debian Linux; LoopKit (github.com/loopkit) does the calculations on an iPhone through a RileyLink bridge but must be compiled and installed by the user. Nightscout collects the data on a free Heroku instance backed by a free MongoDB Atlas cluster; after three years she is still under the 500 MB limit. The glucose sensors and insulin pumps themselves are not open source.
• Withee trusts the code because of its research and its many eyes. She credits Dana Lewis with the original research, testing dosing algorithms against collected blood-sugar data, combined with others' reverse engineering of pumps and sensors; OpenAPS, LoopKit and AndroidAPS share much of the same technology. She says the repos are so actively maintained that she had to turn off notifications, and estimates at least 2,000 people run such systems.
• The FDA leaves self-built systems alone but not store apps. She says the FDA treats a do-it-yourself pancreas as a single-person experiment it does not regulate, but would have to review an app sold through the Play Store or App Store, a hurdle she thinks med tech companies avoided until the open-source community's success pushed them to build closed systems. Tidepool Loop, a non-profit bridging open-source software and closed-source devices, filed with the FDA in May of the previous year and, last she heard, was still pending.
• She argues hacking the device would cost far more than it gains. An attacker would have to find and break into her phone's hotspot, then the device's Linux, and the pump beeps at every message and doses slowly through thin tubing, so alarms would let her correct a low; altering Nightscout values would be overwritten by new readings. Giard adds the principle that a breach should cost more than it is worth.

2021-07-08 · open the original ↗

• A replacement PoE injector did not fix the link-speed fault. The dish kept falling back from gigabit to 10/100, capping throughput at 100 Mbps. The injector RMA was placed May 27, 2021 and arrived June 11; the problem returned after 5–10 hours instead of minutes.
• Starlink's slow RMA shipping would hurt users relying on it. The host notes 15 days for the injector and wants an option to pay for expedited shipping. Starlink did approve a new dish without question on June 25, and it arrived July 2, a week later.
• The new dish appears to have fixed the problem, so far. The next day his port showed a gigabit link, with about 200+ Mbps through his Starlink VLAN to an EdgeRouter X and about 246 Mbps over Starlink's router Wi-Fi; he will keep watching it. The new unit has a smooth face and grey cable and stand, and its cable still cannot be detached.
• The host is skeptical Starlink's economics can work. Citing estimates of $2,000+ per dish sold at $500 and 42,000 satellites at $250,000–500,000 each, he figures his two dishes leave about a $3,000 loss after about $1,000 paid ($500 plus $100 a month since February). He admits this oversimplifies, and worries more users could cut speeds and raise latency.

2021-07-06 · open the original ↗

• WAN settings need changing only for unusual connections. Plain DHCP from the ISP needs nothing; the advanced settings are for a static IP or a block of statics, PPPoE credentials, or internet delivered on a VLAN. The host says the series applies broadly to a USG or a cloud- or PC-hosted controller, not just the UDM Pro.
• Threat management trades throughput for protection. IDS only alerts; IPS blocks and costs more CPU and RAM. Sensitivity levels run from 6 of 35 categories at level 1 to 35 at level 5 (17 at level 3). His roughly 700 Mbps connection should cope, but he advises speed-testing after enabling it.
• Turn off auto-scale network to control the LAN subnet. With it on, the new dashboard hides the gateway IP, subnet and DHCP pool. He sets DHCP to 192.168.1.50–254, keeping .2–.49 for static addresses; the classic interface has no auto-scale setting at all.
• Leave automatic firmware updates off for networks managed for others. He waits to see whether a firmware release is stable and updates in maintenance windows; home users who care less might enable automatic updates.
• Use WPA2/WPA3 mixed mode and keep 2.4 GHz enabled. Not every client or access point does WPA3, and older IoT devices need 2.4 GHz. For the open guest Wi-Fi he set up at real-estate offices, he combined bandwidth limits with a Wi-Fi schedule so it ran only during business hours.

2021-07-02 · open the original ↗

• Good feedback beats precise parts. Fried says manufacturing needs about 99.9% accuracy at each of four or five steps (board design, stencil paste, pick-and-place, oven), because errors multiply and one bad part means hand rework. In her account, the error-prone stencil printer is rescued by a vision check that flags missing paste or a needed cleaning cycle, and the former Westinghouse inductor factory's thick concrete floor lets the pick-and-place machines hold precision several storeys up in Manhattan.
• Shortening the feedback loop is how imperfect people get better. Fried contrasts punch cards and UV-erasing EPROMs (five to ten minutes) with CircuitPython, which Hanselman shows running on a PyPortal: save in Visual Studio Code and it reloads. Hanselman adds his own analogy: a finger-stick reading shows his blood sugar from 15 minutes ago and insulin takes an hour and a half to act, while real-time readings let him make small corrections instead of big swings.
• Technically wrong choices can be the right engineering decision. Fried says CircuitPython deliberately exposes a FAT file system on flash without wear leveling, against the textbook answer of littlefs or a journaling file system, because mass storage is the one driver every computer has and wear-out is a long way off. She credits Microsoft Research's UF2 bootloader with padding files to twice their size so it no longer matters in what order each operating system writes them.
• Teach top-down and get something working first. Hanselman relays a colleague's view that he teaches "from the metal up" while newcomers work from the screen down; Fried sides with the top-down approach, arguing beginners need not know cache from RAM and premature optimization is the root of all evil. She says experienced engineers get "engineer's paralysis" from knowing the right way, and she breaks the work into single steps because "all code gets refactored".
• Deadlines and spite get projects finished. Fried says management's job is getting people over the finish line, and she uses each quarterly AdaBox as a deadline to force development, such as shipping a low-power Wi-Fi e-ink project on a new Espressif chip. She describes herself as "motivated by hate" of bad electronics; applied to Hanselman's son, who stalls on screenwriting, she suggests voice notes, starting at the second paragraph, and rewriting a show he dislikes as fanfic.

2021-07-02 · open the original ↗

• Firebase is a whole app platform, not just a database. Dowden runs Andromeda, which builds its own and clients' products mostly on Firebase. It covers database, serverless functions, analytics, crash reporting, ads and A/B testing, under one SDK and one web console. His team adopted it for a database that needed no servers, since operations was not a skill the team had, and they used more of it as the platform grew. Their parking-management app has run on it for about six years.
• Firebase offers two databases, and Firestore suits real-time apps at scale. The original Realtime Database is a pure JSON store read by path, and it can import and export JSON files at any location. Firestore is a document database of collections and sub-collections, and collections with the same name can be queried as one group. He says Firestore supports up to 1 million concurrent listeners and pushes changes to all of them within milliseconds over a WebSocket SDK, so a live page needs no extra code.
• Cloud Functions' strength is event triggers, not HTTP endpoints. Functions created in Firebase also appear in Google Cloud Platform, where more customization is available. He writes them in TypeScript; Firebase supports Node.js with TypeScript or JavaScript. Triggers include database changes, file uploads, new users, schedules and Pub/Sub topics, and Google scales instances out under load.
• Its operational tools are mobile-first, and web has nearly caught up. Analytics is Google Analytics underneath. Crashlytics is mobile-only, so his team uses TrackJS for web crash reporting. Cloud Messaging (FCM) handles push notifications, AdMob serves ads, and Remote Config acts as segmentable feature flags that combine with A/B testing.
• He recommends starting with Firebase static hosting. He calls it "ridiculously fast": firebase deploy publishes atomically with no downtime, versions automatically, and allows one-click rollback in the console. It adds a CDN and HTTPS certificates, and the free plan covers a lot of traffic. He uses it instead of GitHub hosting. For server-side rendering such as Angular Universal, the CLI can deploy to a cloud function instead. To learn, he points to the docs at firebase.google.com/docs and Firebase's YouTube series.

2021-06-30 · open the original ↗

• The centred taskbar is a default, not a requirement. Hanselman shows that taskbar settings move Start back to the left; live tiles are gone, which he calls "a bummer", replaced by pinned apps and a "recommended" list of recently added apps and recently opened documents.
• Window management gets snap layouts and named virtual desktops. Hovering the maximize button offers snap assist layouts (he notes it does not yet appear in every app, such as Windows Terminal); task view creates named desktops ("work stuff", "home stuff") that he switches with a four-finger swipe, and a widgets panel slides out with calendar, traffic, to-dos and stocks.
• Developer tooling is built in rather than bolted on. Command Prompt opened inside Windows Terminal, which he says can be set as the default "probably later this fall"; wsl --install sets up WSL and Ubuntu in one step, and Linux GUI apps, including Chrome on Ubuntu, show up in the Start menu. He pins Gmail and Outlook as PWAs installed from Edge, which he notes also works on Windows 10 today.
• The rebuilt Settings app is, in his view, the best part. He calls it "really cohesive" and searchable, and demonstrates Swift Pair for a Surface mouse, per-app network data usage, per-app battery usage "like on an iPhone", editable grouped quick-settings buttons and a per-app input/output volume mixer that used to be buried in Windows 10.
• This first public preview is rough; use a spare machine. He points out old dialogs in new chrome (Run, Explorer options, Control Panel), a "show more options" context menu that hides third-party entries, square buttons among rounded ones and a white flash in dark mode. He advises against putting it on a daily driver; his 13-year-old's Call of Duty has "weird things" on it.

2021-06-29 · open the original ↗

• Unless eth0 is in the Internet zone, the firewall does nothing. Out of the box eth0 sits in "trusted, excluded from firewall", so every rule on the Networks tab is moot. The host then lists each allowed LAN (in CIDR form), VPN subnet and remote WAN IP under Networks, with a description for each ("Bob's home IP"). He prefers the Local zone, which opens only FreePBX's ports, over Trusted, which opens everything, though FreePBX wants at least one trusted entry.
• A double reboot opens a window to undo a firewall lockout. Rebooting twice within five minutes is the safety net: if a mistake on the Networks tab locks you out, the second boot comes up with the FreePBX firewall disabled for five minutes so you can fix it. On a hosted server that means forcing two restarts from the provider's console; on bare metal, pulling the power twice. The host says this has happened to him.
• The host treats the responsive firewall as a last resort. Remote staff on dynamic home IPs can be allowed by dynamic DNS name, but DNS time-to-live can leave FreePBX an hour or more behind an address change. Responsive firewall lets every non-blacklisted address send a few packets, so valid phones authenticate, and jails brute-force attempts for growing periods. He calls it effective, but leaving it off is more secure.
• Sysadmin Pro is mandatory on every FreePBX system Crosstalk installs. It costs $25 once, on what the host thinks is a 25-year licence (his showed expiry in 2046), and is tied to the deployment ID. It adds the built-in VPN server, email/SMTP setup in the GUI and dynamic DNS. It is bought from Module Admin with a Sangoma portal login.
• Whitelist your own addresses in intrusion detection. Intrusion detection works alongside the firewall and had already banned an attacker's IP on his test server, since scripts scan WAN addresses constantly. He also sets a real FQDN hostname instead of freepbx.sangoma.local and disables PnP, which doesn't work hosted. A time-zone change, which time conditions depend on, needs a reboot.

2021-06-29 · open the original ↗

• Set module updates to email-only but security updates to automatic. A module update that breaks something, such as one not properly QA'd, can leave you facing a dead PBX in the morning with no idea why. Email-only lets you install in a maintenance window, while a zero-day security fix should go on at once. Update checks run at a random time between midnight and 4 a.m. so millions of servers don't hit Sangoma at once.
• Activate the server, because commercial modules attach to its deployment ID. An unactivated FreePBX cannot install commercial modules, and activation needs a free Sangoma portal account. The deployment ID can be moved to a new server or hosting provider, so modules already bought, even Sysadmin Pro, are not lost.
• The host updates through SSH, not the GUI, because SSH resolves dependencies. In the GUI, a module that depends on another forces repeated check-and-update rounds until everything is current; a single SSH run handles it in one pass. He therefore skips the pile of updates offered right after activation.
• On a hosted server, trust only your own IP. The firewall wizard offers to trust the client and then its whole network. Trusting the /24 of a WAN address would let neighbours on the same range reach the login GUI. On a local install, trusting the LAN subnet is usually fine because the phones live there too. He also avoids "admin" as the GUI username, since countless devices use it.
• One Apply Config click covers many changes. The red button appears after almost every edit, but it can be clicked once after a batch of work.

2021-06-28 · open the original ↗

• InlineData couples test data to the test and accepts only primitives. Giard says the coupling matters little for simple cases but can for complex ones, and that InlineData offers no way to pass a whole object as a parameter. ClassData is xUnit's answer to both.
• The test data lives in its own class implementing IEnumerable<object[]>. His class holds a read-only Data property, a list of object arrays (15, 30, 45; -15, -30, -45; -15, 30, 15). The generic GetEnumerator returns this.Data.GetEnumerator(), and the non-generic one returns this.GetEnumerator(). Each array must match the test method's parameters in number, data type and meaning. He notes the entries could be full objects with properties and sub-objects.
• [ClassData(typeof(...))] replaces InlineData on the test method. The class's type is passed to the attribute, and each array becomes one run, with its three values as the method's three parameters. He describes this as "three times three". In Test Explorer the new class shows one method with three tests under it, all passing.
• Debugging shows the runs do not follow the data's listed order. With a breakpoint set and that one test debugged, the first pass received -15, -30, -45, the second -15, 30, 15, and the third 15, 30, 45.

2021-06-24 · open the original ↗

• SonarQube combines a server, a code scanner and an IDE linter. Bauman, a few months into using it, describes the SonarQube server that receives scans, SonarScanner that analyses the code base, and SonarLint, a free IDE extension (he uses it in Visual Studio) that flags code smells and security vulnerabilities as you type and links to a plain-English explanation, such as socket exhaustion from an HttpClient held as a singleton in .NET. The server can be bought as a hosted service or self-hosted on an Azure or AWS VM, supports over 25 languages (the host later checks: 27, including COBOL), and SonarLint can run in a connected mode that pulls the team's own rules from it.
• Quality gates let a tech lead decide what code gets in. The default rules are rigid and follow each language's governing body, linking to Microsoft docs for C# and to PEPs for Python; a lead can set thresholds such as the number of code smells allowed per commit, or tolerate old code's faults while holding new code clean. Repositories and new commits get A–F grades, which Bauman says has helped on a mixed client team where some developers come from Java and bring its casing conventions.
• A failed Sonar scan fails the pipeline and blocks the merge. His Azure DevOps pipeline runs build, unit tests and a Sonar scan on each pull request, with SonarScanner run as a CLI in a bash task and flags used to, for example, skip JavaScript files. By default the scanner passes and the repository's health simply declines; it only fails the pipeline if told to wait on the quality gate.
• The free Community edition breaks down on larger teams. The team uses Community, which covers CI integration, bugs and vulnerabilities but only the main branch and a subset of languages; Developer edition adds multi-branch and pull-request analysis. The drawback he has hit is that the dashboard shows only the most recent scan, so a teammate's run hides why your own pipeline failed; he advises starting on Community and upgrading later.
• Teams need some automated quality gate, not necessarily SonarQube. Bauman's take-home is that larger teams should put a barrier between the programmer and the code base to validate quality. He says it could equally be StyleCop or a GitHub Super-Linter if the connected server isn't wanted.

2021-06-24 · open the original ↗

• The UDM-Pro's built-in switch has no PoE. Powering a UniFi access point therefore needs either a PoE injector on a UDM-Pro port or a PoE switch; the host uses a USW-Pro-24-PoE, adopted with one click and then updated, which takes about five minutes including the reboot.
• A solid blue ring means the device belongs to another controller. His U6 Lite showed solid blue, meaning it had already been adopted elsewhere, so he held reset with a paper clip for about 10 seconds. Once reset it showed a solid white ring and could be adopted, after which the ring turned blue again.
• The host avoids leaning on the classic dashboard. At the time of recording, the new UniFi Network dashboard lacked settings he needs, so he expects to switch between the two. He prefers the classic one but tries to use the new one wherever it works.

2021-06-24 · open the original ↗

• The host rates the UDM-Pro for homes, not businesses. For a business he would rather have more robust gear with more redundancy built in. At $379 it is a UniFi OS console that can run UniFi Network, Protect, Access and Talk as separate applications, whereas the standard UDM runs Network only.
• The UDM-Pro cannot be set up without a cloud login. The wizard demands a ui.com account and allows no local login, a point the host calls contentious, so he strongly urges two-factor authentication (he uses a hardware key). Before turning off remote access via unifi.ui.com, he says, create a local-access-only super admin, as he does at the end.
• The host disables automatic updates and turns on SSH. SSH is his way back in if an application update crashes an app and it has to be rolled back to an earlier version. As of recording he was on UniFi OS 1.9.3, UniFi Network 6.1.71 and UniFi Protect 1.17, updated to 1.18.1.
• Scheduled backups survive a factory reset. The wizard offered him backups taken before he reset the unit, so he recommends keeping scheduled backups on; he moves his from Monday 12 a.m. to Sunday 1 a.m.
• Choosing "personal" or "business" setup changes nothing the host could find. He installed both ways and saw no difference beyond extra questions about the business. For Protect storage he recommends a surveillance drive such as a Seagate SkyHawk or WD Purple, sized to the cameras and recording quality.

2021-06-22 · open the original ↗

• Vultr's ISO library gives a one-click FreePBX install. The host picks the newer of two FreePBX images there (version 2011-5), older than the freepbx.org download but equivalent once updated, and a location close to the users. He asks viewers to use his Vultr referral code, which he says gives up to $100 of credit; DigitalOcean and AWS are mentioned as alternatives.
• The host sizes a 20-extension office at the $20-a-month server. The $5 or $10 a month tiers are enough for testing; he demonstrates on $10. Vultr's auto backups, about $2 a month more, are worth turning on in production, but they are snapshots on the same host and are lost with it, so separate off-site backups come later in the series.
• Point a DNS A record at the server before going further. A valid FQDN resolving to the server's IP matters for the Let's Encrypt certificate, WebRTC and remote softphones, and he names the Vultr server after its FQDN to keep many servers straight.
• Remove the ISO in Vultr instead of pressing Reboot. After the 30–40 minute automated install, the virtual ISO is still mounted like a thumb drive. Removing it under the server's settings triggers the reboot itself, and the server then comes up at the initial setup wizard over HTTPS with a certificate warning.

2021-06-22 · open the original ↗

• Any Linux-capable PC will run FreePBX for learning, but not production. For production the host wants something more robust, and installs on the Crosstalk 710 appliance his firm sells. He says it handles up to 100 phones or 30 concurrent calls, is fanless, and comes like all Crosstalk servers with a 2-year warranty and FreePBX pre-installed.
• Start the ISO download early because Sangoma's servers are slow. The full ISO, the 64-bit Sangoma 7 PBX build released in April 2021 at recording, can take a couple of hours to come down from freepbx.org.
• Triple-check the target drive when flashing the USB stick. He writes the ISO with balenaEtcher to a 32 GB stick (he thinks at least 8 GB is needed), warning that picking the wrong device, such as his 12 TB backup drive, would erase it.
• The installer asks only for a root password, then runs itself. With a network cable attached it also updates FreePBX, taking about 30–40 minutes. Pull the stick at reboot, then browse to the DHCP-assigned address shown at the console login to reach the setup wizard.

2021-06-21 · open the original ↗

• A Theory replaces near-duplicate Fact tests with one parameterized method. The three tests from the previous episode (positive, negative and mixed inputs to an add function) repeat most of their code. Marking one method [Theory] instead of [Fact] lets it take int firstNumber, int secondNumber and int expectedResult as parameters and assert the expected result against the actual one.
• Each [InlineData] attribute supplies one test case's values. The number of arguments must equal the number of method parameters, so here each holds three numbers: 15, 30, 45; -15, -30, -45; and -15, 30, 15. He keeps the empty "arrange" comment as a reminder that the inputs arrive as arguments.
• Test Explorer adds a level showing each case's result. Under the class and method sit the three individual runs, each with its own green check mark. Giard recommends running all tests often, and says that is an incentive to keep tests short in running time.
• InlineData works only for primitive types like strings and numbers. He says passing objects needs another approach, covered in the next episode.

2021-06-19 · open the original ↗

• var is syntactic sugar, not loose typing. Hanselman shows that var fixes the type at first assignment, so a var holding a string cannot later take 99; its scope is the same as any other variable. Some dislike it because a reader must hover to learn what the type is.
• The compiler widens mixed arithmetic to avoid losing data. int plus decimal yields a decimal, and anything plus a string yields a string; Cherita's deliberate probing exposed that adding a double to a decimal is an error. The M, D and F suffixes force the literal's type, and display turned out to be a notebook helper for Console.WriteLine, not the source of the conversions.
• VS Code is an editor; Visual Studio 2019 is an IDE. Hanselman generalises that back-end developers use Visual Studio for profilers and memory diagnostics while front-end work suits VS Code. He opens the notebook in Notepad to show it is plain text, and says to switch to the full IDE "when it feeds your spirit".
• Keep side effects out of display calls. Writing apples += 10 inside a display changes the value on every run, so he stores results in variables first, and contrasts = for assignment with == for comparison. Cherita likes the ternary operator; Hanselman finds a well-named function easier to read.
• Whitespace doesn't matter in C#, but braces do. A brace-less if or else covers only the next line, so a second indented line runs regardless, unlike Python; braces "give it a hug". Brace placement is style, and a team's inherited style wins.

2021-06-19 · open the original ↗

• A new coding pack installs everything a beginner needs at once. Hanselman presents the .NET Coding Pack at dotnet.microsoft.com/learntocode, launched on this stream, which installs VS Code, the .NET SDK and .NET Interactive notebooks on Windows or Mac. For Ubuntu there is no pack installer; viewers install .NET, VS Code and the notebook extensions separately.
• Notebooks teach the syntax without the boilerplate. A .dib notebook mixes Markdown and runnable code cells, so learners skip the usings and namespaces they are normally told to ignore; Cherita, a TypeScript developer in the M365 admin center group, says beginners otherwise have to go back later and work out what mattered. Hanselman notes a notebook is not yet an app.
• C# is strongly typed, and its errors are help, not threats. Assigning 99 to a string fails, where JavaScript would allow it and TypeScript would not; an explicit cast of 10.5 to an int silently drops the 0.5. Hanselman urges reading the error message, while conceding some messages are genuinely bad.
• String interpolation needs a marker to protect existing code. Hanselman explains the $ prefix exists so that strings already containing curly braces keep their old behaviour. Strings are objects with methods and properties; ToUpper depends on a culture taken silently from the computer, so a Turkish user would capitalise a dotted "i" differently and stored strings might later fail to match.
• Know a method ought to exist; don't memorise it. He compares it to knowing Spanish has a word for hunger without recalling it, and lets IntelliSense find Trim or Replace. He favours storing return values in variables over clever one-liners because they can be inspected while debugging, and says naming case style is the team's choice.

2021-06-19 · open the original ↗

• Switch statements beat long if chains when cases share outcomes. Hanselman shows a day-of-week switch where several cases fall through to one branch, which as an if would be "pretty hairy"; break must be intentional, and DayOfWeek is an enumeration rather than a special case syntax. He also explains short-circuit evaluation: with ||, the second test never runs once the first is true.
• A for loop is setup, a question, and an after-step. He breaks the one-line header into its three parts and warns that off-by-one errors live there; <= against an array's length throws an out-of-bounds index error because arrays start at zero. Cherita was "on the fence" about foreach but called it easier to read after seeing both versions side by side.
• The boilerplate the notebook hid is a namespace, class and entry point. After dotnet new console, Hanselman explains using System as a shorthand, the namespace as "a space to make room for names", and static void Main as the program's front door. The notebook's display was a helper wrapping Console.WriteLine, which he rewrites by hand.
• Visual Studio does the same thing the command line does, hidden. He takes input with Console.ReadLine, runs it with dotnet run, sets breakpoints to watch variables change, then pastes the code into a Visual Studio 2019 console project to show the GUI route. He describes console.log versus interactive debugging as a cultural choice; Cherita says her team uses browser debuggers all the time.
• One language reaches desktop, web and more. He creates a Windows Forms app and a dotnet new web app on .NET 5, and points to the 19-part .NET 101 video series and Microsoft Learn. He asks viewers whether to build 200- and 300-level notebook courseware, described as an idea rather than a plan.

2021-06-18 · open the original ↗

• A shebang line turns a script into an executable-like tool. Shobokshi says a #! line naming the interpreter at the top of a file means it no longer has to be run through the bash keyword or carry a .sh extension, so a file called foo can be invoked like any executable; he uses this for small helper utilities and calls one of the shebang forms more portable than the others.
• Split long scripts into files and source them as libraries. His team's install script for a custom framework grew to 3,000 lines as it had to behave differently per cloud and work stream, and he now breaks even small scripts into files pulled in with the source keyword so the pieces can be reused. He points to Gruntwork's bash-commons and to an argument-parsing library of his own on GitHub (the captions render it as "sh arg"), written because parsing input is something he and others look up every time.
• Bash scripts deserve unit tests too, and ShellSpec makes that workable. ShellSpec is a BDD-style framework with a Gherkin-like DSL that exercises real bash, supports setup and teardown, and can mock bash functions. He mocks a function that fetches an API token so tests use a fake value, and says the tests have found edge cases in the team's more advanced scripts.
• Structure scripts as functions, with main invoked at the bottom. Shobokshi says this makes scripts easier to test and read than a flat list of commands. The catch is that return only gives back an integer status code, so strings are returned by echoing them inside the function and capturing the output into a variable at the call site.
• Turn on set -e and set -u in every script. With -e the script stops when a subcommand such as Terraform exits non-zero, where bash would otherwise carry on; he allows there are edge cases where that isn't wanted. With -u an undeclared variable, typically a typo that bash would silently create, causes an exit; the host likens it to Visual Basic's Option Explicit.

2021-06-18 · open the original ↗

• Pick on-premises for one building and hosted for scattered sites. The host breaks every phone system into server, service and phones. On-prem suits a single 50-person office, keeps extension-to-extension calls off the internet and takes POTS or PRI cards, but needs IT upkeep and replacing every three to five years. He recommends against virtualized FreePBX unless you really know VMs, because misallocated resources or networking show up as call-quality problems.
• Hosted PBXs double internal-call bandwidth but survive a site outage. Each extension-to-extension call goes out to the server and back, and every site should ideally have a static IP. If a site's internet drops, the hosted server can still answer and send calls to voicemail. The host recommends Vultr, which Crosstalk's own paid PBX hosting runs on, with QA'd monthly updates, snapshots, off-site backups and no contract.
• The host keeps POTS mainly as backup for schools and hospitals. POTS costs about $25–40 or more a month per line before usage, and caller ID follows whichever line a call used. A PRI carries 23 voice channels (an E1 elsewhere, 30) for roughly $500–750 a month. Both are often delivered over an internet circuit now, so ask the provider whether outages take them down.
• SIP trunking is where the host saves customers the most money. It is metered, per-channel or enterprise with SLAs, and fast to deploy with flexible caller ID, but it depends on internet quality and good provider documentation. He ranks first his own Crosstalk SIP: 5 channels with about 4,000 minutes or 10 with about 8,500. Next come ClearlyIP for complex sites needing Kari's Law and RAY BAUM'S Act E911 compliance, then voip.ms for small metered use.
• Standardize on FreePBX-compatible phones and skip over-priced brands. Phone apps for call forwarding, queues and visual voicemail need FreePBX support, and he calls Cisco and Digium phones overpriced at up to $400 a handset. Crosstalk's own phones, sold through its store, are the 270 (receptionist, sidecar, Wi-Fi, Bluetooth), the 250 (standard users) and the 230 (break room), with colour screens, gigabit pass-through and 2-year warranties.

2021-06-16 · open the original ↗

• Wang's cancer returned fast and spread everywhere. Diagnosed with colorectal cancer about three years earlier and declared cancer-free after a year to a year and a half of treatment, he felt ill from around November, ignored it, and in March scans showed spread to liver, lungs, lymph nodes, bladder and pelvis. He was told over a video call that he had two months without treatment and about a 14% chance with intensive treatment.
• You cannot wait until life stops being hard to be happy. Wang says he is "pretty selfish these days", his emotions spike and his mind goes to dark places he does not share on social media; but everyone is going through something, so he does what makes him happy now. Hanselman says he catches himself planning for "when I'm 55".
• He is throwing "the kitchen sink" at it, with partial success. The latest chemo, six sessions, was worse than any before and "almost broke me"; he has also cut back meat, eats mostly vegetables and fruit, and takes Eastern remedies and mushrooms his mother sends. He reports the cancer pushed back, with the largest remaining spot on the left lobe of his liver not currently growing, and does not know how many more rounds he could take.
• He chose to return to work because it brings him joy. Wang says it is not for financial reasons and he would advise most people the opposite; he wore a portable chemo pump hidden under a hoodie for his cameo in the Build developer keynote, his second keynote with Hanselman while on chemo. He takes up Hanselman's "don't accept the defaults" only as "figure out what makes you happy", pointing to his daughter's work-life balance as equally valid.
• Stage fear never leaves; rote repetition is how he controls it. He traces this to his mother making him play violin pieces three times in a row without a mistake before moving on, and says the nervousness, which he links to imposter syndrome, drives him to practise more. Hanselman says after 20-plus years only his recovery afterwards has got faster.

2021-06-15 · open the original ↗

• Ubiquiti sells the airFiber 60 LR as a long-range bridge. The point-to-point radio, model AF60-LR-US lists at $399 a radio, about $800 for the pair. Ubiquiti rates it to 12 km and 1.8 Gbps full duplex, with GPS sync and Bluetooth setup. A Ubiquiti contest's longest link stood at 24.69 km at the time of recording.
• The host measured near full speed across his own test link. At 27 m, which he calls overkill for this radio, the built-in speed test gave 854 Mbps receive and 853 Mbps transmit, just over 1.7 Gbps total, with 0.85 ms ping. A laptop iPerf test through three switches and a few hundred feet of cable gave 681 Mbps single-threaded and 869 Mbps with five threads. A client's 5.8 km link showed 86.5% link potential and 1.75 Gbps capacity.
• Crosstalk typically backs a 60 GHz link with a 5 GHz one. 60 GHz suffers in rain and snow, and the LR has no backup radio. So Crosstalk mounts a second 5 GHz point-to-point on the same poles and runs OSPF between them for failover or load balancing, as on that 5.8 km client link. This also allows upgrading one pair at a time.
• The radios have almost no settings to tune. Wave AI adjusts the link automatically, and there is no power-output setting. The host sets both ends up over Bluetooth from the UISP app. A static IP without DNS servers left the radio unable to reach UISP until he added DNS, after which GPS suggested the nearby UISP site. A precision alignment mount helps aim the dishes.
• For shorter links, the host points to the standard AF60. At $299 a radio and rated to 2 km, it has an integrated 5 GHz backup radio that handles failover itself, with no OSPF needed.

2021-06-15 · open the original ↗

• Visual Studio's xUnit template sets up the test project; no download needed. xUnit.net is a free, open-source framework for testing .NET applications, found at xunit.net. Giard picks the C# template, names the project DemoXunit.Tests (one test project per project tested, and one test class per class), and uses .NET 3.1, which he calls the current long-term-support version. The test project needs a reference to the project under test and a using for its namespace.
• Giard structures each test as arrange, act, assert. Arrange sets up variables or objects, act calls the method under test, and assert checks the outcome; he types these as comments first and favours long, descriptive test-method names. xUnit's Assert class offers checks for equality, true or false, and not null, and Assert.Equal takes the expected value first, then the actual.
• The [Fact] attribute is what makes a method a test. Once it is added, the test appears in Test Explorer (Test > Test Explorer), listed under its class, project and solution. Tests can be run or debugged at any of those levels, or from the code: right-clicking inside a method runs that test, and inside the class but outside a method runs the whole class. Debugging adds breakpoints and stepping.
• A deliberately broken function shows a failing test's report. With the add method changed to a bug, the test failed and reported it expected 45 but got -15. Mid-demo, unexplained build errors cleared after he closed and reopened Visual Studio, cleaned and rebuilt, which he puts down to cached results.
• Test edge cases beyond the happy path. He copies the test for negative inputs (-15 + -30 = -45) and mixed inputs (-15 + 30 = 15), and suggests zero or unexpected formats as other cases worth checking. A blue exclamation point in Test Explorer marks a test never yet run.

2021-06-10 · open the original ↗

• Wi-Fi phones roam on existing access points, unlike DECT handsets. Crosstalk normally sells DECT cordless systems (Snom, Yealink W60P and Yealink's new multi-cell), which only work within range of their bases. The host's earlier Wi-Fi phone, from another maker, kept losing Wi-Fi. Spectralink is said to be robust, but the first handset with its required support contract, charger, battery and cables comes to over $1,200.
• The host, usually no Grandstream fan, will stock both phones. He says the video is not sponsored. Grandstream phones are normally not fully FreePBX-compatible and its PBXs are convoluted, so Crosstalk doesn't sell them. But he likes the WP810 and WP820 enough to recommend them and add them to the Crosstalk store. Both auto-provision from Crosstalk Cloud or FreePBX Endpoint Manager.
• The host recommends the WP820 over the cheaper WP810. At $145 against $114.95, the extra $30 buys a larger 2.4-inch screen, Bluetooth 4.2, Android 7.0 and video calling. Both have dual-band Wi-Fi, 1,500 mAh batteries, two SIP accounts and a programmable push-to-talk/panic button. The WP820's swivelling belt clip juts out enough to stop him holding it on his shoulder, so he would drop it.
• Roaming between access points worked in the host's home test. On one call he hopped access points five or six times, each hop beeping and cutting about a second of audio, yet the conversation stayed usable. He stresses this depends wholly on the Wi-Fi. Firewall or access-point timeouts on the registration can stop the phone ringing on inbound calls.
• The WP820 shows incoming video but sends none. It has no camera. It displayed 30 frames-per-second video from a Yealink T58A, useful for a video door phone. Calls from the WP820 to the Yealink never started video, and the host could not tell whether that was a limitation or his configuration.

2021-06-09 · open the original ↗

• Renaming a project in a .NET solution is non-trivial. Other projects reference it, and Giard's own conventions mean the namespace and the project's folder should change to match. Renaming the folder breaks the solution's path to it. He warns that some files, such as JavaScript or Blazor files, sometimes have trouble being renamed automatically.
• Rename in Visual Studio, then close it and rename the folder. His steps: rename the project (F2), replace every occurrence of the old name, close Visual Studio, rename the folder, then fix the project's path in the .sln file in a text editor. He knows of no way to change the project's location from inside Visual Studio. In the demo a project became VideoManager.Lib.
• Replace occurrences one at a time, not globally. He uses Ctrl+H across the entire solution with match case and whole word on, but says even that is not enough, so he reviews each hit and replaces it (Alt+R) or skips it. That way he avoids renaming similarly named libraries he has not yet handled.
• After reopening, clean the solution before rebuilding. He waits for background restore tasks to finish, then cleans the solution, because cached build output can get in the way after radical changes. The rebuild is the first clue that something broke.
• A successful build does not prove nothing broke. Giard runs a manual smoke test of the app and says a change this large warrants full regression: integration, automated, unit and manual tests. He notes the rename makes source-control history look odd, which he says cannot be avoided, and advises making a backup and committing to source control first.

2021-06-08 · open the original ↗

• Scherotter coaches technologists who want creativity in their work. Trained as an architect at the University of Arizona before a software career, he began mentoring informally, partly through Microsoft programs, and this year put up techcreativecoaching.com, a WordPress site whose only cost was the domain. He coaches people in India, Ireland and Africa from his home office, does not charge, and is exploring it as a future possibility while staying at Microsoft.
• Creativity is a muscle he exercises by keeping sketchbooks. He has filled leather-bound journals for 30 years with drawings, paintings and anything flat, and says this "paintbrush creativity" keeps him in shape for solving problems creatively as a program manager. Much of that is spotting metaphors, since in his view few problems are absolutely new.
• A career is a series of jobs converging on what you love. He frames a tech career on three axes: the technology (blockchain, machine learning, creative tools), the type of work (coding, UI design, program management, testing) and the domain (financial services, public sector, his own media and communications). Each job should bring what you are good at and what you like closer together, and he says early-career people get frustrated expecting to have all of it at once.
• Show your creativity openly, in a way that is authentically yours. He tells people to use their own imagery in presentations instead of stock photos or clip art, and points to the 3D printer behind him, which he actually builds with. Most people keep it hidden for fear of critique, which he counts as part of growth: without feedback it is "just a little art project on the side."
• He gets new ideas by combining things not combined before. When the pandemic stopped his travel and his journaling stalled, he flew Microsoft's new Flight Simulator to New York, Sydney Harbour, Giza and San Francisco, paused it and painted the scene, which he calls virtual flight sketching. On a freelance project he built a virtual gallery of Hebru Brantley's work, called Solus, in Babylon.js for an auction house show in Hong Kong, adding the navigation and interactivity to a supplied model over a couple of weeks.

2021-06-08 · open the original ↗

• A cheap Sonoff RF bridge brings RF remote devices into Home Assistant. The host's build is the Sonoff RF Bridge ($15.99 on Amazon), a Sonoff PIR motion sensor ($17.19), any RF LED strip lights and Home Assistant 2021.6.2 on a Raspberry Pi 4, with HACS and a file editor installed. He calls the Pi great for learning but expects heavier users to move to a VM on a Synology NAS or an Intel NUC.
• Pair the bridge in eWeLink on a 2.4 GHz IoT network. His hardware version of the bridge does not support 5 GHz. It signals pairing mode with two short blinks and one long for 300 seconds. Each remote button and the motion sensor (added as an "alarm") is learned within a 60-second window. He turns off the app's alarm notifications, which fire on every motion event.
• The SonoffLAN integration links eWeLink to Home Assistant. Installed through HACS, it defaults to local-and-cloud mode, with cloud-only and local-only as options. The eWeLink email and password go in configuration.yaml. Because YAML is finicky about spacing, he suggests copying from his GitHub repository instead of the web page.
• A notification automation reveals each learned RF command's number. Commands are numbered from 0 in the order learned: his on button was 0, off was 1, the motion sensor 2. Scripts sending commands 0 and 1 then become dashboard buttons. An automation needs a unique 14-digit ID to be editable or switchable in the UI. He admits his blog post had wrong command numbers he meant to fix.
• His motion lights turn off 90 seconds after motion stops. The sensor is declared with device class motion and a 60-second timeout that restarts on each detection. A second automation waits a further 30 seconds after the sensor goes off before sending the off command. He calls both timings arbitrary and suggests tuning them.

2021-06-04 · open the original ↗

• The host's Starlink VLAN exists only on the switches, not the firewall. VLAN 574 ("STA" in leet speak) carries Starlink from his garage to an EdgeRouter X in his office. It is created in UniFi as a "VLAN only" network with that tag and does not exist on his Netgate SG-1100 pfSense firewall, since it needs no DHCP or routing. His other networks are the main LAN, guest VLAN 10 and IoT VLAN 107.
• An untagged port at each end puts Starlink on the VLAN. The Starlink PoE injector's data port plugs into port 5 of a USW Lite 16 PoE, set to a Starlink-only profile. All VLANs then ride trunk ports ("All" profile) through the attic and rack switches. Port 5 of a US-8-150W, again Starlink-only, feeds the EdgeRouter X's WAN. The link still showed 100 Mbps full duplex because of his separate Starlink fault.
• A trunk port carries one untagged network plus tagged VLANs. Anything plugged in lands on the untagged network, and VLAN-aware devices can also use the tagged ones. That is how one access point serves the main, guest and IoT Wi-Fi. A port untagged for a single VLAN gives its device nothing else, even if that device understands VLANs.
• Best practice keeps the Starlink VLAN off every non-trunk port. The host builds a "non-Starlink VLANs" profile (native LAN plus IoT and guest) and shows it applied to his U6-LR's port. That keeps a device on an ordinary port from joining VLAN 574. At home he leaves most ports on "All" because he controls all the equipment, but in an enterprise he would segregate further.