Hacker Newsnew | past | comments | ask | show | jobs | submit | more BiteCode_dev's commentslogin

Looks amazing. I've been using some hacky vscode setup to do GTD but this seems like exactly what i want. Too bad it's not open source, i would still pay their instance but i would feel better knowing i have an escape hatch if they close.

Will try it anyway.


It depends of how much control I have over the topic. If:

- I know the topic well. - I understand the domain well. - I can picture the technology, expectations and API well. - I have the mastery of my time and deadline.

Then yes, I do.

E.G: I'm currently making a proposal for a new API in an open source python lib called iohttp.

I have time. I have leeway. I have a good overview of the entire problem. So I can afford thinking about the API first:

https://github.com/aio-libs/aiohttp/issues/4346

Then I will write unit tests. Then I will write the code.

But that's a rare case.

Many times coding involves fiddling, trying things out, writing drafts and snippets until something seems to start to solve what you want.

Other times, you want a quick and dirty script or you have a big system, but you are ok with it to fail from time to time. The cost of testing is then not worth it. You'll be surprised of how well a company can be, and how satisfied customers says they are despite the website showing a 500 once in a while.

And you course, you have limited resources and deadlines. Unit tests are an upfront payment, and you may not be able to afford it. Like it's often the case, this means the total cost of the project will likely be higher, but the initial cost will be in your range of price. And you may very well need that.

One additional things few people talk about is how organizations can make that hard for you. You may be working in orgs where the tooling you need (CI, version system, specific testing lib...) may be refused to you. You may even be working in companies where you cannot get clear information about the domain, but get a vague specs, and the only way to design the software is to ship, see it break, and wait for customers to report to you because otherwise the marketing people don't let you talk to them. I'd say change job, but it's not the point.

At last, you have the problem of the less experienced devs. TDD is hard. It requires years of practice in the field to be done properly because you build a system in a completely abstract way. Dependency injection, inversion of control, mocking and all that stuff you need to make a properly testable system is not intuitive: you learn it on the way. Even harder is the fact you have to use it in your head since your are not coding it, but what wraps it, first. And even more terrible is the fact that baldly implemented, over used and over engineered, design patterns make problem worst, not better.


Plus, they lived their entire live in a world with:

- sugar everywhere

- deficit in many important nutriments

- attention depleting tech and env

- boomer culture messages in a post 2000 economy and social climate

- low but continuous stressing life style

- micro doses of thousands of pollutants in food, air and water

I have seen first hand how working those improved my quality of life, and I have not even lived my all life on those conditions.

I find it amazing how resilient they are, considering.


Can you explains why even Boomers were moody teens before all this happened?


Moody is too vague. I'm more interested in measurable symptoms. E.G: teens have menstruations sooner on average, and boys have a fully develop hormonal system later on average than boomers. They have more adhd, are more obeses, have more cases of diabetes, have less confidence in the future or trust in the institutions.

And yes, that all translates to attitude.


I'm completely ok with it being complicated to design and cost 10 euros more if it delivers.

And despite all its flaws, USBC does. It's small. It's reversible. I can charge all my devices on it: my samsung tablet, my oneplus phone, my logitech mouse, my jabra headphones, my switch and my XPS 13 laptop, which I can also dock to my screen/ethernet. It's amazing.

So yes, I had to take the time to find the proper cable, and I bought a few before finding brands I can trust to work with all my machines... and not fry them.

It sucks, but it will improve with time, like it did with early USB and HDMI (too many people forgot how terrible THAT was). It already has, in fact, improved a lot, and it will keep it up.

Meanwhile, I have only one cable and one dongle in my bag that do everything. I don't even have to get them out most of the time.

This is imperfect. This needs fixing.

But this is better than what we had before. Way better.


There's also a middle-ground that would have made people happier across the spectrum. Standard power cable, for phone, laptop, etc. Standard data cable, for your ereader, camera. Standard video-audio cable.

I think it's a good principle to have that if the engineers suffer, that will run downhill on consumers too. I think every engineer can find examples of that.


I like plugging my laptop into the USB-C hub at home, and that giving me keyboard, mouse, display, ethernet and power without having to plug in multiple cables.


USB-C allows me to hook up my laptop with one cable. Your middle-ground standard wouldn't.


I probably misused the "across the spectrum" idiom. What I meant is that it would make developers much happier than USB-C and it would make consumers much happier than USB 3. My idea was that it's a more wholesome situation than content consumer and frustrated developer.


As a consumer I want one cable to do all of it. Sure, like anything else, it will be difficult and slow progress to get there. As developer you sould see it as challenge.


It's good that you know your ideal user-experience (developers are aware of it too), but the trade-offs will depend on what's happening on the developer's end, so dismissing developer difficulties is not conducive to a good consumer experience.

Developer difficulty is a symptom of a more systemic problem, and it will usually affect the end-user through: higher-cost, lower-reliability, accumulation of technical debt, and/or resource-inefficiency.


Not dismissing, I understand it is a hard problem. I'm simply saying that is the developer job to solve the issue.

Similar thing happen in web development, there are tons of device, screen size, javascript framework, browser, os, all come with its own quirks. Is it suck ? Yes it is but it is my job as as a competent developer to manage all those complexity and solve the issue. From the user prescriptive, all they need to know is the site is working with whatever device they use, that is the goal.


Yes, but because of "the suck" of web it's in some ways in a dismal state and that affects users. We can have only a small number of browsers that are well maintained; browsing is resource-intensive; web is extremely inaccessible to people with disabilities; web is notorious for security problems; fingerprinting can be seen as a side-effect of bad design; web requires a lot of man-hours to develop and maintain, like you mentioned; etc. The list seems endless.


Sure you I wish that the state is in whatever it is that I prefer but thats unrealistic. As developer you have to be ready with whatever state you are in.

Sure, If there are only small number of browser or all of my user is not disabled, I would prefer that and my job will be easier but that is not realistic.

I have to be ready with whatever situation I've been dealt with. As developer my job is to solve problem and manage complexity. As a developer dealing with complex and hard problem is given.

In the end is the developer job to handle those issue so that in the perspective of the user, everything just works.

If I can't do that then I'm a failed developer.


I can sympathize with wanting to solve as many problems as possible and as well as possible and to provide a great user experience; however, our ability as devs is necessarily limited and that should be accounted for. Both the developer and the user are in this together.


Good thing that developer/engineer are specifically trained for problem solving and managing complexity.


Where I work many people have a docking station with a USB-C plug and a power plug molded into one connector. One cable to plug in, can safety deliver more power than USB-C can alone. With a bit of work engineers could have standardized an expandable cable. You have your basic USB-c cable with one fast (display port?) lane, one usb-3 data lane, the - basic cable in the article, enough for phones and light use. Then the larger connector that can supply more power for laptops and has a few more high speed data lanes for more monitors. It is a bit of work to make the connector work for either mode, but it is a solvable problem.


What happens when you get a laptop with a slightly different port arrangement? Just throw out all the old docks? Seems like quite a solution from a manufacturer's perspective!


The reality is every time someone gets a new laptop they get a new docking station. Docking stations seem to last about 3 years. It is very nice when traveling to get to a remote office and hear "Joe is out this week, you can use his office - he even has a docking station that will fit your laptop". However it is questionable if his docking station will fit your laptop.

If connector N (usb-c or whatever) has enough bandwidth and a standard connector it would be really nice if we could give everybody that docking station. Even if we have to upgrade docking stations every few years, so long as I can mix and match I'd be happy.


And the manufacturers would have never agreed on which ones to use. Starting with the zoo of power connections out there.

It's a miracle I can charge my Dell XPS15 laptop using the work-provided Apple USB-C charger. Really.


It's just a question of regulation. I was under the impression, correct me if I'm wrong, that USB-C is the result of regulation as well.


Looking into it, it was more of a campaign. Letters of intent, etc. And it was only for mobile phone chargers.


I'm soon switching a lot of things to USB-C. What brands of cables would you recommend ?


I just use a MacBook charger and usb-c cable, even if i don't have any apple devices.

The reason is that I move a lot and forget my chargers one or twice a year in a customer office, a hotel or at the airport.

The MacBook pro charger is very expensive, but charges everything I have and I can buy it pretty much everywhere, including at the last minute in an airport.

It's also rated "best cable on the planet" by Nathan K: https://docs.google.com/spreadsheets/d/1vnpEXfo2HCGADdd9G2x9...

Any cable recommended by Benson Leung or Nathan K will do, anyway.

A good compilation of those: https://www.reddit.com/r/GooglePixel/wiki/officialguide/usbc...


Anker. They're consistently reliable, and accessories are their whole business, so a problem would hurt their reputation a lot more than one of the big hardware companies.


Well, if you look at the spreadsheet linked below https://docs.google.com/spreadsheets/d/1vnpEXfo2HCGADdd9G2x9... Anker did have one model of cable that was recalled because it was dangerous :)


Another vote for Anker, they're consistently better quality than the other "random character name" brands found on Amazon.


I cast another vote for Anker.

Also, for USB-C to USB-B, DJ Tech Tools' Chroma Cables [0] are very good.

[0] https://store.djtechtools.com/products/chroma-cables-usb-c


something with Kevlar in it like powerline.


Isn't already the case though?

If yes, it should not prevent us from tying out term limit, only push us to also add some mechanism to deal with bureaucrats.


It also further empowers the lobbyists, which adds to the risks. A government full of ignorant people will tend to fall back on those who seem to know what they're talking about.


Lobbying is a problem on its own, and should be dealt with anyway, and independently of other changes to the system.

There is no way to preserve democracy when there is a legal path for bribing those in charge of power.


The lottery system has several variants to take that in consideration.

In one of them, everybody votes for x (e.g: 3) people around them. Then you remove the top 10% to avoid people campaigning, and the bottom 10% to avoid psychopathic candidates.

The ones that remain can then register to the lottery.

In some variants, you have limits to the wealth you have, jobs you hold, geographic min and max, number of times you can try, etc.

The lottery concept is not about embracing chaos, but smoothing the odds, playing with the Gauss curve in our favor, limiting the network+money factor and avoiding turning politics into a job.


Tactical voting would have a field day with that.


Tactical voting has a field day with almost any strategy. We love playing games with out system.

But my intuition is that this system would lower the incentive to do so by nature, because of the culture it promotes.


The alternative is to write SansI/O code (https://sans-io.readthedocs.io/), so that your program don't have to think about that.

Besides, you don't have to put async/await everywhere: if your code is not performing IO, it completely ignore this concern.

The problem is that most of your code is mixing I/O and non I/O code, and people just don't think about it. E.G: a django website is not just a web server, but has also plenty of calls to the session store, the cache backend, the ORM, etc.

Now you could argue that the compiler/interpreter is supposed to hide the sync/async choice to the code user. Unfortunately, this hides where the concurrency happens, and things have dependencies on each others. Some are exclusive, some must follow each others, some can be parallel but must all finish together at some points, some access concurrent resources...

You must have control over all that, and for that to happen, you can either:

- have guard code around each place you expect concurrency. This is what we do with threads, and it sucks. Locking is hard, and you always miss some race condition because it can switch anywhere. - have implicit but official switch points and silos you must know by heart. This is what gevent does. It's great for small systems, not so much at scale. - have explicit switch points and silos: async/await, promises, go-routine. This is tedious to write, but the dangerous spots are very clear and it forces you to think about concurrency upfront.

The last one is the least worse system we managed to write.


What's the benefit over the forked SingleFile ?

I use it regularly, and it creates a single pure HTML file with inline medias. Easy to read anywhere.

So appart from the compression gain, why the need for SingleFileZ ?



But not supported by iOS Safari. And APNG not supported by edge. That's why sites like imgur use mp4 videos as an alternative to GIFs.

It's not that a bad deal though, one minute of video is roughly 7Mo for something close to DVD quality. I had GIFS lasting a few seconds weighting way more, with terrible quality.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: