Non-obvious remote work techniques | Hacker News Reader

Non-obvious remote work techniques

queue.acm.org

128 pointsrbanffy84 comments

> makes the mute button easier to find

It's odd this is still an issue. I'm not sure why the mute button is not featured more prominently on videochat apps. It's almost the primary feature during a call that makes things go smoothly.

Also, there's always one guy who has to say every bloody time, "are you there, have you forgot to unmute?" if you fail to unmute and start talking <100ms after your name has been mentioned.

With Zoom (as long as Zoom has focus) you can hold down spacebar to temporarily unmute yourself. It's been a real lifesaver

Back in my day we called that, push-to-talk. Everything old is new again.

Teamspeak guild, assemble!

Yea, gaming is miles ahead of everything in this area. It obviously doesn't focus on video too much, but audio and text are on point.

I've heard of a few companies preferring Discord over Slack or Teams right now exactly because gaming figured out so much of this decades ago and Discord prioritizing it works well for those companies.

Hangouts is the worst in this regard. It hides in a bar off the bottom of the window that only stays visible if your mouse is in the exact correct position. It takes 5 seconds of chasing it to be able to catch it and click it. Why can't it just stay visible??

It's funny because for me, it's always on, and it's terrible because it shows "are you trying to talk" every time my kid screams (very often). The worst part of it, is that it covers the bottom of the screen, even during screensharing.

What is the most important part of the screen when using a terminal?

Just use a headset with a physical mute button. I've been using this pair[0] for years. It works great.

[0] https://www.logitech.com/en-us/product/wireless-headset-h800

I do this and it has downsides:

* muting yourself is not clearly visible to the conference app (Your face doesn't get a mic with a strikethrough on the screen). Apps that warn you that your audio device is receiving audio but you are muted on the application level no longer work correctly.

* some apps (Looking at you, Teams) will actually complain that you seem to have deactivated your audio device (of course you did - that's the whole point).

This is only a problem with headsets that have poor drivers.

Better devices like Jabra headsets integrate directly with Teams and will actually synchronize mute status between the headset and Teams!

Cool!

YMMV but 0 of my WFH colleagues own business-grade headsets though. Some have Airpods, and most use Gaming headsets (like I do) or Bose QuietComfort (which IIRC don't have dedicated mute buttons).

Until we find a unified API around all those devices, a prescribed unified usage pattern IMO doesn't make sense yet.

Eventually companies will realize they should spend or reimburse to upgrade "home office" equipment to "business-grade".

Some videoconference solutions that are pushed on businesses aren't mature at all: in that Cisco thing when you mute someone it mutes them for everybody (not just for you), so it's hard to have a mixed conference because you either hear people in your physical room twice or remotes obviously don't hear them at all if you use the 'mute participant' button.

I can’t think of a situation where someone would need to be muted and it shouldn’t mute them for everyone. That is not a scenario that I think anyone would expect to work that way.

If you have some people in the same room and some not, this is a very useful feature.

As someone pointed out, all these things already happened in gaming.

I appreciate apps that provide a universal keyboard shortcut like Zoom that works even if I’m focused on another app. Next best is Google Meet but it only works when you’re focused on the same tab. Worst is Teams, WebEx etc that don’t provide any keyboard shortcut

My mute preference in videoconference tools:

Universal keyboard shortcut > app focused keyboard shortcut > no keyboard shortcut

I set up a mute pedal going to my microphone so there's no dependency on software or even clicking with my hands. I just toggle it with my foot. But then, it's a full audio setup with a shure sm7b, preprocessor and audio interface, instead of your typical conference call headset.

We use Hangouts a lot at my employer, and it inevitably slides away the mute/video off/screenshare buttons if you move the mouse away, so when we do the polite "mute until your turn to speak" standups, it tends to 10 seconds of "you're muted" while we frantically fumble with the mouse and repeat our statements.

This drove me crazy in Google Meet. Also had issues with trying to show the Meet bottom toolbar sometimes where it just refused to unhide itself. Mercifully, there's a keyboard shortcut for toggling mute state. Try out `⌘/Ctrl + d` next time.

I feel like "Accurate Chat Status" only works if it's automatic (as in, a calendar integration that switches your status to "in a meeting" or "it's after 6pm, probably afk.)

Otherwise you get:

4pm: "Ah, bob's status says he's in a meeting. I could message him, but I'd like this to be a quick, synchronous, chat discussion."

5pm: "Guess bob's still in a meeting..."

7pm: "Let me check - yep, still 'in a meeting'"

10pm: "Yeah, he forgot to update his status"

And then you never trust those statuses again.

You can hook up Google Calendar to Slack. It does very well. I agree with you on auto-status.

For me automatic idle+meeting detection is not good enough. There are lots of times when I can't chat immediately, but I'm using my computer and not in a scheduled meeting. Pair programming, someone else at my desk (pre covid), programming in the zone, watching a presentation, currently chatting with someone else, etc. etc.

I want the expectation that my Slack icon is green to mean that I'm at work and I'll reply sometime today, not that I'm ready for you to interrupt what I'm doing right now.

Simply don't replay immediately. You'll build the expectations as people learn to work with you. Some people on my team respond immediately 95% of the time. Some people on my team turn off all slack notifications and check it throughout the day. We all work together pretty well.

I completely agree that there shouldn't be an expectation of an immediate reply and rather expect the exchage to be asynchronous. However, that renders to whole "presence indicator" moot, which is the subject of this particular thread.

If I try to talk to Bob and the expectation is that he might take a while to get back to me, it doesn't make any difference to know that he's "available" or "in a meeting". If he's available and coding he might not anwer for an hour. If he's in a meeting and bored he might answer right away.

Several of those situations you could use existing "meeting detection": full screen apps signal to most of these detection systems that you are busy. If you are watching a presentation or even programming in the zone, if you have it full screen you should get automatic signalling that you are busy.

(VS and VS Code have nice fullscreen views that are great because they also cut down on other distractions even inside of themselves. VS Code also has Zen Mode which pairs really nicely with fullscreen and cuts down on distractions about as much as possible.)

Hey that's a good tip, I didn't know about fullscreen status being a signal for busy detection. This is a useful tip for me personally that I'll start using, thanks!

That doesn't quite resolve what I'm talking about. Keeping apps full screen isn't realistic for me, since I'm app switching all the time and often need multiple apps side-by-side, especially when coding. And again, my main goal commenting was just to challenge the very idea that chat status should be minute-to-minute accurate. I don't want that idea to spread, I want people to expect to wait for a response, I want chat to be default asynchronous, and for people to just be happy when they occasionally don't have to wait.

By and large this just isn't a big problem for me, it's a totally minor issue, but would be nice if we collectively have the same expectations. At my work, people mostly do use chat asyncronously and don't start chats with just "hi" waiting for a response. I'm personally relatively happy with my work interactions, but I'd prefer if attitudes didn't shift in the direction the article is advocating.

> Accurate Chat Status > In meatspace you don't walk up to someone's desk and just start talking.

This list seems to assume or suggest that chat should only be used for synchronous conversation... Are there reasons to avoid using chat asyncronously?

My assumption is that chat can be asynchronous, but I recognize many people don’t want to ask their question until I’m actively responding. I can imagine security concerns, like someone doesn’t want their question to be read by people near me or someone wandering by my desk while I’m away...

But, my status as to whether I can chat synchronously changes dozens of times per day, and I don’t particularly want to have to manually toggle my status when I’m on the phone or when I’m chatting with someone else, or when I’m in the middle of coding and prefer to answer your “quick question” later, since it will inevitably require a 10 minute conversation. Quick to ask is often not quick to answer.

If you're doing incident response or on-call work, I think it needs to be more synchronous and more accurate. However, chat messages are more natural to me to indicate availability than setting the availability itself. Too many people don't set accurate statuses and it can get changed by the programs own rules, so it's a noisy signal IMO.

If you're doing incident response or on-call people don't contact you based on your availability, you guarantee your availability so people can contact you.

Do people expect immediate responses to chat messages? if I ping a coworker and don't hear an immediate response, I just assume they're busy and use the time to ask my question in detail. am I am outlier here?

I don’t know that people necessarily expect immediate responses, I don’t think you’re an outlier. But the article is explicitly calling for accurate chat status for the purpose of being able to chat synchronously and immediately. And there are some behaviors that indicate, if not the expectation to get immediate chat responses, at least the signal of being unwilling to use chat until responses will be immediate. I’m referring to how some people start their conversation with “hi”, and wait to hear back before telling you what they want to talk about.

#5 - IMHO you just can’t emulate a physical social event, the biggest problem being that you cannot establish smaller groups which are natural in physical events.

Agree, but that’s better than nothing. I spent several months working with a remote team. I didn’t really know my teammates for six lonely months, they were just names in a screen, until we started scheduling social meetings. Of course it’s nothing like the real thing, but it helps.

In my dnd group I just switch Discord channels when we need synchro conversations when the group is split. Because as DM you sort of have to participate everywhere this is harder but I like the multi-channel method.

I wonder how much time and effort has gone into solving the hallway effect. Or the cocktail party effect.

Maybe we just assume it can't be solved but haven't really tried that hard, or at least that recently?

I recently participated in C++ on Sea (https://cpponsea.uk/) - online, of course. The conference setup used https://remo.co/conference/ which automatically groups people into tables, where each table is essentially a video conference call. You can move around freely between tables but each table can only hold 8 people.

I felt it was possibly better than a physical conference because it's so simple to join new conversations - no walking around looking for interesting groups (you can see who is currently sitting at each table) and none of the nonverbal social awkwardness in joining a conversation. Plus you always have the names and profiles right there. Never been easier to meet people imo.

I'd contradict your claim based on this experience. You clearly can establish small groups, and it seems to work really well when you do.

"the biggest problem being that you cannot establish smaller groups"

Zoon has breakout rooms.

Rally.video has tables.

I love the quick chat protocol. It's a concrete example of what to do, instead of just saying "please don't just say hi".

I think there is one good strong counterpoint to it -- it is if there is a chance that the words "tomato incident" will be screencast to a meeting containing execs and/or external clients. Some people operate in "hello protocol" all day, and for them, I find it easier to just let the extra round trips happen.

Similarly, if I have access to a colleague's calendar, I usually check it before I initiate a quick chat protocol message, to avoid the same failure mode.

#3 - Holy hell I hate it when someone sends me a message with just "Hi". Just tell me what you want!

What I do is just not respond until they say what they want from me.

And I also do this when someone just pastes an error message without a question.

Same here. I have a “No Hello/Hi/Hey” policy. I just ignore it

Well, that sounds logical. I do the same, too.

1. Current work

2. Today work

3. Review for when work

4. Far future work

Everything in 1 and 2 are lined up for work. Anything without clear action items requires follow-up and clearly signals non-urgency so it's in 3 which is the end of the day. And a 4 is just intentionally forgotten.

If you stick that in your github.com/username/me/README.md then it's documented behaviour.

Same. Or I reply "nohello.com".

It's easier to reply "sorry, busy right now" in response to "hi". Once they say what they want I feel more obliged to make time for them.

Can't really determine if I actually have time for something before I know what it is. And unfortunately, I don't have the luxury of just ignoring things I don't want to deal with.

There's a huge space to occupy between 'hi' and launching into your life story before you know if the other person is even reading.

So you want them to use the chat program like e-mail?

The article explains perfectly what the parent would want instead.

So an e-mail with the subject tomato incident and a link to a zoom?

Chat is for back and forth.

And to pretend that this "conversation preparation" somehow takes (and interrupts) a worthy chunk of your day is either laughable or pompous.

plus, didn't we agree that chat is async and thus this "conversation setup" never happens? I am forced to respond to "how are you" in real time?

async convo = UDP "convo setup" = TCP, and it doesn't happen

> hi

...

>> hi what's up

...

> there's a bug, it does abc when you do xyz

...

> ok I fixed it

versus

> hi, there's a bug, it does abc when you do xyz

...

> ok I fixed it

pleasantries are fine, just include the point along with them

I think you might be missing the point that the back-and-forth is actually terrible for work communication, considering that workdays are quite limited.

Um, yes?

Email WORKS JUST FINE. You know what else works fine? A PHONE.

I know that actually calling someone on the phone seems to be traumatic for anyone under 35, but hey.

If it's really important--call. If it's not that important--why am I being interrupted? Send me email.

Chat--all the expressiveness of text with the interruptiveness of a phone call.

If something's broken or if someone needs some info/action to be unblocked, I want a notification. I'm less likely to check email for something time sensitive, but I sure as hell don't want 20 phone calls a day.

> I know that actually calling someone on the phone seems to be traumatic for anyone under 35, but hey.

First off, this is needlessly hostile and ageist. Please stop it. I am over 35 and we do not need to keep on with this "well if you're young you are too stupid to understand the One True Ways of doing things." Age discrimination is unfair regardless of direction.

> Chat--all the expressiveness of text with the interruptiveness of a phone call.

For me, chat is far less intrusive than a phone call yet still has the back-and-forth that doesn't easily happen with an e-mail. If you send me a chat message, that's more than just "hi," I can look at it and formulate a brief reply and we can keep going back and forth as needed. We may also migrate the conversation to voice (and video!) or we may drop to a rate more like e-mail.

Either way, I do not need a ringing phone 15 times per day any more than I need 300 e-mails or daily chats that start with "hi" and nothing more.

What industry do you work in, exactly? I think it would be a very odd thing if I were to ask my co-workers for their phone numbers. They'd probably ask why I needed it, and I'd say "so that I can call you if there's a problem". They'd say "why don't we just use Google Meet, like we do for everything else? Is there something wrong with your mic? I don't understand."

> They'd say "why don't we just use Google Meet, like we do for everything else? Is there something wrong with your mic? I don't understand."

Okay, that's the first real, valid argument I've seen against a phone call. I had no idea that these services had penetrated so far that they are normalized.

However, not all of us operate in "developer-only" land. People external to the company (at least, prior to Covid) would have phones but almost certainly not a "conferencing app". So, giving out your number would be a fairly normal thing to do in the course of business so giving it to colleagues would not be considered unusual--at least in the US.

You say "Hi!", I'm in the middle of wrapping up something. 8 minutes later, I have time, I say "Hello, what's up?". Unfortunately, during these 8 minutes you moved on to doing something after waiting for 30 seconds. 12 minutes later you see my reply, and say "Can I ask you something about this Foo module"". Again, during these 12 minutes I moved on to something else, blablablabla.

Had your first message been "Hi! In the Foo module, do we really need the function Bar to be public?" you would have been all set after 8 minutes.

A chat, like email, is just a tool to get shit done.

Plenty more of that here: https://news.ycombinator.com/item?id=24239880

also https://news.ycombinator.com/item?id=24259156

I normally just write back "hello", and know that I'll be receiving a question from them in the next 10 minutes.

I've tried ignoring it, but it's the sort of thing that sits simmering in the back of my mind. It's only one guy that does it. He's also the sort of guy who wouldn't understand what I meant if I tried to explain why I didn't want him to do it.

Does anyone have any suggestions on a platform to facilitate Tip #4: Idling in a Videoconference Room? Zoom is the obvious choice, but wonder if anyone has built something better for this use case?

I have friends who ran a startup full remote. They sat on Discord together all day. Video off most of the time.

I don't know that non-gamers would understand idling in a discord/mumble/ventrilo/teamspeak server, but it's for sure the next big trend in remote work. I've been hanging out this way with friends for a long time.

I won't be surprised to see Discord pivot to b2b in a few years and completely destroy Slack.

Kevin Kwok agrees: https://kwokchain.com/2019/08/16/the-arc-of-collaboration/

I think this will work - but only for people (groups) that can communicate GREAT.

I'm a big fan of https://pragli.com/ for this. Been using this successfully with my team.

If you want to host your own, Jitsi Meet Server. It doesn't take much CPU or RAM. Bandwidth is king. Uses browser or ios/android apps for clients.

I am light on with webtech and linux but I followed their "few commands and a shell script" install notes to get something working with LetsEncrypt https.

If you host the video centre, and everything is https out from there, it should be somewhat secure.

For business security you may want to go over it more carefully than my home bodgy job though.

This is cool, thanks for the tip! How is performance compared to meet.jit.si?

I used meet.jit.si once but my own server mostly and I couldn't tell you the difference in user experience. Seemed exactly the same.

Video: The default install sets up for 720p. It seems to start notably worse and then scale up to 720p within a minute, often much less. Some congestion control thing I expect.

You can set 1080p or other (lower too) resolution limits if you like. I tried 1080p but at ~6-7Mbit per connection outgoing my home internet 18MBit upload got saturated pretty quick! The default 720p has worked ok for up to 6 people for a few hours at a time. 1 of the people was Europe-to-Australia and seemed like anyone else. No huge lag times but some lag as you'd expect from that distance.

Audio: I note this separately as it seems to be handled separately by Jitsi. We've noticed it seems to tick along pretty well the same the whole time, whether the video is going up or down in resolution due to congestion or not. Audio quality seems fine. The mic makes more difference in quality than the audio codec.

Apps: Have a requirement to go into setting and put the URL for your own server. Otherwise it defaults to the public one. Bit awkward. It seems to suck down a lot of battery but I haven't benchmarked it against anything else. Maybe this is normal for camera/compression/decompression/display/network all at the same time?!

Other niceties: can share desktop/application, has a text chat side option too, has "hand up" indicator if someone wants to indicate they wish to speak - not that we've ever used it.

All in all, it works well for home personal use with 6 people so far. Definitely handy to have and I intend to keep one running for the future. Being able to send a link to my friends and show them my desktop, or vice versa, is handy if nothing else.

Edit - added App section

Thanks for the detailed rundown, definitely gonna give it a go now!

Missed the edit window of my other post. Just got it setup and it was a breeze, only took ~30 minutes! It's running with Docker (nginx w/ Let's Encrypt, Jitsi on http, port 8000) and it was a bit unclear what specifically needed to be configured in the `.env` file[0] but managed to get it up and running on the first try.

So far I've only tried it with 2 people but the CPU barely flinched (a 2 core basic VPS) - most of the heavy lifting seems to be happening client-side. Uses ~500 MB of RAM.

Initial impression: 5/5

[0] https://jitsi.github.io/handbook/docs/devops-guide/devops-gu...

Cool. Glad you got it going and are happy with it! You've made my day. :-)

Two points:

- My only hitch you likely won't come across. The LetsEncrypt will need to update the cert every now and again(months?). On my ubuntu VM install there is a cron job for this. But I leave my VM turned off unless I need it, so every now and again I have to "leave it on for longer" and snapshot after it updates the cert.

- two people use might be a special case where it sets up Peer to peer. So you may wish to test with at least three people before using it in anger.

https://jitsi.org/blog/p2p4121/

I can tell you it won't work with Zoom. For some reason it takes over all of your sound, so if I play quiet music I can't hear it.

This would also be incredibly wasteful of bandwidth and CPU. Bandwidth is not unlimited, and if everyone in the Zoomy world kept a silent conference on in the background all day, the internet pipes [Zoom's especially] would be quite a bit more full, and my CPU would continue to churn unnecessarily.

Then there's the etiquette. What if people just start water-cooler-chatting? Maybe at a low volume it's OK, but I could see people leaving just because of too much laughter [yes laughter is annoying and yes I am a horrible person].

One of the worst Zoom things is that it doesn’t allow you to control the volume of meetings...

My wife's ex colleagues built this - https://www.remotion.com/. It's more like always on room

There was this Show HN for watercooler.work a few months ago.

https://news.ycombinator.com/item?id=23434271

I'm working on a product to help with this exact type of use.

I would love to find anyone interested in trying out a tool for this -- contact me at carss.w@gmail.com or drop a name/email into the form at https://heysync.chat if that's you! I will have a simple version ready to try within a few weeks.

The idea is make asynchronous audio chat easy, in a context similar to slack. You can send audio easily in place of text and whoever you sent it to can listen to it live or later. Then they can get back to you the same way. Imagine a chat room with 5 people talking, where you can replay the whole thing and chime in later if you were busy. It's chat across time.

Again, if you're interested, please reach out!

I haven't used it personally, but https://sneek.io/ will do ad-hoc video chat. Those that aren't "live" on the call are still up on the wall of faces, with a screen capture from their webcam every few minutes.

Avoids the waste of unused video over the network, but still gives a sense of your teammates being live people going about their day.

Is there a good reason why ACM can't make their site readable for the largest group of computers in the world - mobile devices?

Works fine for me on Firefox, and I can even use Reader mode.

For me, with Firefox on Android, it delivered what appeared to be a desktop site (tiny text, clearly non-mobile layout). I almost never encounter that on websites, so the problem is not with me nor my phone. My phone is a Note 8 (still reasonably high performance and with a very large screen).

It's nice that it works for you, and sure I could try reader mode, but I should not have to switch to reader mode to view a website.

stopped reading after tip #1. so lame. not practical at all.

Me: Hi!

You: Hi.

Absolutely not. Just say what you want.

The article lists that as an example of what _not_ to do

Reply on news.ycombinator.com