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

I wish they'd have given more examples for traditional bikes than bottom brackets. Yeah, bike shops have to deal with lots of different BBs, but that's because they deal with bikes that might be 30 or 40 years old, from all over the planet. Some threaded, some press-fit, etc. Some high-end, some very cheap.

On the most part, bike manufacturers use standardized parts that can be replaced by and end-user with sufficient know-how and the tools to do it. There aren't that many companies making drivetrain parts, so you tend to see Shimano and SRAM just about everywhere, and maybe the odd Campagnolo-equipped bike every now and then. At least here in the US. (Unrelated, Shimano's product range is crazy - somehow their components come stock on bikes ranging from $250 up to $12k or more.)

Outside of < $200 Wal-Mart bikes, I've never had any trouble repairing or finding someone to do "normal" repairs or maintenance on a bicycle. I'd like to know what prompted the article, unless the real point was to complain about E-Bike batteries, which is not something I can really comment on.


I spent 6 months chasing down a new rear wheel because my freewheel hub had broke and I didn't want to spend twice the price of the used bike on a new wheel set from the manufacturer.

Visited many local bike shops, got a lot of bad advice there (same as online), then finally got told what to order exactly at one online shop because not even their supplier had it.

Maybe the blame is not on the manufacturer here because they wanted to make it less repairable, but if showed me a lesson in non-standard components. (it's an 11-speed Ultegra on a QR, which seems to be very, very uncommon).

But just the amount of "just do X" responses I got showed that there are too many fine details.


I agree that the message seems to mostly be about e-bikes, though I will say I was recently forced to get yet another Shimano BB / rotor tool recently, and I have quite an assortment of them now... My newest ride has a belt (not an ebike, but Gates + Alfine 11sp) which means more unique bits (but likely not for a while, I did have to get a fitting to oil the hub, but that's less work than keeping a chain lubed, so I can't complain).


Hub is way more work than the chain. I have to repack mine soon and its going to take me getting it open, replacing the ball bearings, packing them in new grease after cleaning out old, then an undetermined amount of fiddling getting them just a quarter turn or so too loose in the cones so the quick release will torque them perfectly when its installed. Thats assuming nothing fights me along the way like a seized locknut on the cassette.

Chain on the other hand if it starts getting noisy it gets a generous squirt of rock n roll gold and a rag for 5 mins then its fine for another long while.


Which Hub?

I know back in the day I could abuse the hell out of Nexus/Alfine 8 speeds, had one on a Chicago Schwinn that I rode in all sorts of weather, and another on a Diamondback Tandem that was ridden through tornado warning weather [0] on top of the overall abuse of being on a tandem with two experienced riders.

[0] - 10/10 would totally do again.


Some tiagra 10 speed set up with an 8 speed cassette but its a simple enough cup and cone loose bearing design like most rear hubs made. Still though if you leave it in the rain like me the bearings will dry out and howl when the wheel is spinning, requiring new grease and bearings (since they are cheap enough to replace instead of clean and inspect).


The 'red' Alfine 8 speed hubs had extra weather sealing, that's why it was great in the tornado weather lol.

TBH if you leave it in the rain just do a repack before you ride it again, unless it's a stupidly fancy hub, a lifetime bucket of grease suitable for hub bearings is probably still under 20-30 USD per person, and if you've got the wrenches and a stand [0] a repack can be done in 20 mins or less.

Edit: to be clear, I'm talking about normal hubs, with the Alfine we did an inspection but the seals looked good and a peek done inward indicated everything was fine so... yeah those red bands were gr8.

[0] - IDK if they still make it, but the Minoura FT-1 is... pretty good balance of quirks for the lack of dollars spent, both for truing and other wheel checks...


I had a belt drive not so long ago and when the belt broke, I found out that the entire system had been recalled by Continental with no repair or upgrade path possible. I couldn't even get a replacement belt.

The bike was a write-off as a result.

Oh... the Alfine hub that is supposedly nice and weather-proof get some rust in it that pitted some of the bearing races in the hub. Turns out those are not replaceable short of torching the hub.


Personally, I love my smart home stuff. Homeassistant has been utterly reliable, automating things in node-red is easy and fun. I can use IKEA's stuff or buy from any of dozens of other Zigbee manufacturers. My plant lights go on and off at sunrise/sunset, my phone notifies me when there's movement on my porch or driveway, living room lights offer a nice ambience when I come home at night, etc.

It's mostly inessential stuff but it's nice. I tinker with the system maybe an hour or two every year and it all just works without any real drama or troubleshooting.


I don't know much about the flute world, but there's a guy on youtube who's been extending his contrabassoon's range to absurd new lows for years now. Here he is seven years back:

https://www.youtube.com/watch?v=G8nRRwTnSbM

In the ensuing years he's been manufacturing a proper subcontrabassoon and here's a more recent video of its progress:

https://www.youtube.com/watch?v=LWQADWh3D8s


Zwift is great. I've been on it about three months and it really feels like the most productive training I've ever done. It's hard to get a decent ride in where I live but zwift is always available. They've managed to gamify real-world performance and it's surprisingly fun. There's a lot of different things to do in the game.

But be warned - it's a bit pricey to get up and running. All in, I'm sure I've spent at least $2k (USD) on various equipment and accessories and it's also $20/month for the service. To me it's worth it for sure, but running (for example) is very nearly free. Just make sure you think indoor cycling is something you'd be willing to stick with.


I have a hard time agreeing with the central premise here.

The article states that the Taliban had a protection scheme (possibly legitimate, possibly a racket) going on to effectively tax all trade through their regions and that made them a lot of money. I can believe that.

But calling that the "real economic prize" seems to miss the point. There was so much trade happening because of the tens of billions of US dollars flooding into the nation. If you take away all that money, you are left with a lot less economic activity to tax, and are probably dealing with customers and suppliers who can't afford nearly as much protection money.

If there was a true economic prize, it's the foreign aid coming in which will likely dry up soon. There are no major international shipping routes through Afghanistan (I think! see my note below), so without the foreign money they're left skimming off the top of trade through a poor developing nation, which doesn't seem like much of a prize at all.

However... the article states this : "Even before their blitz into the capital over the weekend, the Taliban had claimed the country’s real economic prize: the trade routes — comprising highways, bridges and footpaths — that serve as strategic choke points for trade across South Asia." I wish they'd elaborated more on this - are there some desirable routes going through Afghanistan that nations in the region might like to use for trade?


The thing is, you profit more from a trade route when more trade flows. This happens with the rule of law, and with infrastructure development. The Taliban could grow this golden goose, or they could strangle it. I'm leaning toward predicting "strangle".


I don’t see much trade opportunities from looking at a map- unless India or China have a desire for a route through there with someone else. Possibly for oil.

Another meme is they’re a resource for rare earths. What people forget is the US has plenty of rare earths - if people want it enough.


Combined import/export trade with India is ~1.5 billion USD per year. Not much but still significant for Afghanistan.

"Federation of India Export Organisation expressed concern that in coming days dry fruit prices may go up due to the turmoil in Afghanistan. India is importing around 85 per cent of dry fruits from Afghanistan."


Yeah sounds about right. How did billions in foreign aid get squandered? Well, tons of middlemen (among them the Taliban) took their cut until nothing was left.


The "OS as a service" era was also when they really ramped up the advertising and telemetry in windows. It would be nice if they decide whatever revenue that brings is not worth the loss of good will people had for Microsoft before all that began and give it up.


I'm not even a privacy nut, but this annoyance has me looking at Linux distros for the first time in 10 years.

Year of the Linu....


I don't want to over-romanticize WSB, but this all seems fairly wholesome and even old-fashioned.

A hoard of like-minded investors put a lot of money into AMC because they felt it was under-valued and with a little push could ride out the pandemic. Now that it's on the rebound, they're not immediately cashing out with their short-term gains, instead they're encouraging the company to avoid wall street games like stock issuance and instead focus on organic growth as the economy recovers.

We were warned they were going to take down whole markets, but at this point they seem like agents of stability.

(Yes, I understand I may be giving the lads way too much credit here, but one can hope.)


This had been my feeling watching all this unfold, but I'm starting to reconsider.

I was there in the early days of 4chan. Back then, brigading/mob justice felt the same: a force for good, a bunch of snappy underdogs thumbing their noses at jerkoffs like the Scientology crowd, or animal abusers. Who could find fault with that?

But the problem with mob justice is that it never stops there. Image board campaigns became increasingly dire and cruel: hacking, ddoss-ing, harassing people that maybe didn't deserve it as much, or at all. And eventually the idea hit the broader culture, mixed with radical left ideology, and mutated into what we now call Cancel Culture.

It's satisfying on a primal level to see institutional investors lose even a little control. But it's just that primal satisfaction that makes mob justice so dangerous. Maybe this is a blip and won't go anywhere. But if it's indicative of a trend, the implications are sinister. The Mob has just gained the ability not just to end people's careers and reputation, but to end whole companies, and potentially even destabilize the whole economy. It's great right now, because we're still in the "harass animal abusers" stage. What happens when we hit the "someone started a rumor about this company, so let's run it into the ground" stage?

Rule by tyrants is bad. Rule by the mob is even worse. Which would you rather live under: Henry the 8th, or the French Revolution?


At this point it's really hard to tell who's a smart trader acting like idiots, or if they're all just novices acting like idiots, along for the ride.

I call them "idiots" as it's a somewhat nicer term than what they call themselves (presumably jokingly).

I love it either way.


I think what was not expected was for so many to keep the momentum going after the initial GME and friends buying craze at the beginning of the year. So many pundits dismissed it. Of course those with the most providence in AMC were rewarded since the stock didn't really spike until June.


I've been primarily a linux user for some time, so I've gotten pretty desensitized to various applications looking vastly different on the same system. But in the last several years that's mostly gone away - Gnome knows how to change the theme for QT apps and KDE can configure GTK themes so things are actually remarkably uniform.

But then when I fire up Windows, it is just this bizarre mix of all these different UI frameworks. I guess Windows isn't the big cash cow it once was for Microsoft and maybe it's not the highest development priority anymore, but still it's remarkably off-putting and I have no idea why they would let it stay in this state.

I read they were introducing a new UI framework to unify everything (Project Reunion), but I can't help but think that will mean it's just another different-looking toolkit thrown into the mix.

With all that said, Microsoft's commitment to backwards compatibility is fairly legendary so there could be an actual technical reason some things can't be redesigned with a newer toolkit.


As far as I know Project Reunion is not really a UI framework - it's more about refactoring Windows APIs into individual libraries which can be used across different OS versions. This should enable apps targeting all versions of Windows 10 to use the latest stuff, in a similar fashion to polyfills in browsers.

This is probably to encourage developers to build software using the new Windows 10 APIs. Which is something they definitely have a problem with - I've been using Windows 10 quite heavily since the first beta and I don't think I've ever seen or used a third party Windows 10 (fluent / modern / metro) application.


It is a bit more than that, and I bet with Windows 11 announcement we will get a more clear picture on it.

It is also a way to bring into Win32 side most of the API that really matter on the UWP side of the fence, most likely (although they are yet to confirm this) as migration path to UWP developers.

.NET Native, C++/CX are pretty much dead, and C++/WinRT keeps missing many of the tooling available on C++/CX for XAML applications.

So I expect them to migrate the APIs into an OS version independent layer, and then give the mercy shot on UWP apps.


I don't understand how this would work. You would tax corporations based on actual income? So money-losing businesses would have to pay taxes on their income even though they operated at a net loss? And low-volume/high-revenue businesses would pay less in taxes than high-volume/low-revenue would?


One of the problems is that net loss is fictional. Companies use well meaning tax credits and deductions to offset income. His point is fair but the specifics need a little tweaking.


I am agnostic on if it should be based off of income or profit and leave those sorts of details up to congress.


> leave those sorts of details up to congress.

Yep, let's leave it to the "experts".

Policy "details" tend to have drastic unforeseen consequences, and our congress-critters tend to defer responsibility of those details to lobbyists.

We the Constituents\People\Voters have a responsibility to do our own thorough research and not allow Congress to figure out the details.

Precision matters and details can make all the difference between a good policy and a destructive policy.


How can one possibly do this? The amount of regulations/laws produced is so enormous, it is impossible to keep up.


The good news is that in my plan, you don't have to care. Unless you are the cfo of a company, or an accountant.


This is an oddly unconvincing little write-up.

Firstly, learning gradle doesn't mean you are learning Maven simply because most dependencies use pom files for their metadata. I just means there's a little XML document embedded in your dependencies that 99% of developers never pay attention to, whether they're using Maven or ant or any other build tool. It's entirely irrelevant.

Secondly, Gradle does not require you to learn Groovy. Use Kotlin if that's your thing - I get the impression that's the preferred DSL nowadays anyways.

Sadly I think those are the only two reasons really presented. Neither is especially convincing. IDE support seems just fine for both platforms.

I actually _like_ Gradle and I think I could come up with a longer list of things about it that aggravate me. But either way, use whatever build system you want, and if, as stated, their projects have fairly humble requirements build-wise, I'm sure either system will get the job done.


Since I mentioned Gradle aggravations, I'll share a couple:

1. Bizarre and inscrutable syntax. They say things like "build scripts are just Groovy code", but some things, like how you define sourcesets or configurations, don't look like anything that should parse, much less compile, in any C-like language. They're doing some fancy meta-programming stuff behind the scenes to keep things light and simple, but it really isolates (and possibly alienates) developers from exactly what's happening under the covers. The Kotlin DSL seems to help make it more clear what's going on, but I haven't used it much yet.

Having used Gradle for years, it is strangely off-putting how unsure I often am about whether to put a colon in a particular place or a comma, or if something should be in parenthesis.

2. If you want to do anything beyond what's documented you can get into a bit of trouble, and if you find answers more than a couple of years old there's a good chance they often won't work. Making sense of the DSL documentation is sometimes a little tricky.

Beyond that, I would dare say it's the least awful option for building for the JVM. Their documentation is very comprehensive, it's fast, they are very fair about providing deprecation warnings before features are removed. It's rare I find a case where I'd like to use some external library that has a Maven plugin but no Gradle support.


It's regular Groovy. Look at this unrelated library that is doing the same thing. Toggle from Groovy to see what it is doing behind the scenes. https://spreadsheet.dsl.builders/#_outlines

The first parameter in a Groovy closure it will be treated as "this" and any function you call can applied to "this" first.

sourceSets is just a function that accepts a closure whose first parameter is sourceSets (this is implementation specific). If you call main inside the closure then groovy will call sourceSets.main and set that property.

What you are probably complaining about is that parenthesis are optional in two contexts. Your function fits on a single line or your last parameter is a closure (curly braces)

You're also confused by the fact that you can pass a map as a parameter by simply doing "function key1: name1, key2: name2".


I have used both and like maven better. I do not like the idea of using a turing complete language to define your builds. Every Gradle project I worked on eventually devolved into build code insanity.

Give me any maven project and I can figure out how to build it and modify dependencies and stuff within minutes.

With Gradle, it can be an hour of messing with code if the project maintainers got creative.

I've never had a project that maven, with it's mostly fixed build setup, couldn't handle. Having a turing complete build script is like turing complete cooking recipes. Pointless and definitely going to create more problems than it solves


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

Search: