Hacker News Reader

Comment by gorgoiler | Hacker News Reader

IPv6 is great if you control lots of components that need to talk to one another. Life is much simpler when you can trivially have thousands of globally routable addresses. It provides great value without having to be 100% deployed and for that, I am grateful.

I liken it to a commercial kitchen. Large numbers of identical steel pans, everything measured in grams, Cambros, walk-in refrigeration. Ideas and tech that are unlikely to catch on in the home kitchen, but when you operate at a larger scale, invaluable.

View on HN

> ...everything measured in grams...

> Ideas and tech that are unlikely to catch on in the home kitchen

Given the posting time I’d guess you were likely not US based so I found this surprising.

Even if that is so, then I continue to be surprised every time I rediscover the fact that the US hasn’t moved towards metrification.

Edit: rereading this the wording is harsher than intended. So to add: Either way the wording tickled me for some reason. Whether intentional or not, thank you (genuinely) for the smile this morning.

Metric units are a mouthful. Centiliter <-> ounce. Kilometer <-> mile. Centimeter <-> inch.

I think having easy-to-pronounce names for the metric units would help. Kliks are much better than kilometers.

Completely agreed. And being in the UK, and of a certain age, I still use inches, miles, pounds and stone but I think that’s due to familiarity during childhood. I don’t use Fahrenheit like my parents did though when describing the weather. Oddly, I also tend to use metres instead of feet/miles for distances that aren’t drivable, or human-height related.

It’s also interesting that in speech I often hear (and find myself), contracting to the SI prefix itself when the unit can be derived from context.

For example: “It’s 25 millis” [long]

The argument for Fahrenheit over Celsius is actually relatively strong: http://lethalletham.com/posts/fahrenheit.html

I also have a soft spot for fractional units and different bases (e.g. 12, 60, etc) over decimal, but I can admit the practical case is comparatively weak and inconsequential.

The author states that the best system is the one you're used to, in his conclusion. Earlier he says that he doesn't like the other system because it uses negative numbers for temperatures that aren't that cold and that also it's good to be able to avoid using negative numbers. Those seem like very subjective reasons to me.

very subjective indeed. i find that knowing whether water will feeeze or snow will melt without having to memorize a number for that rather useful.

well, actually memorizing a number is probably not a problem. i realized that the real advantage is to be able to look whether the number is positive or negative without taking a close look at the actual value.

> aren't that cold

Something very relative to where the author lived and developed the preference. This changes from person to person, and region to region. "Not that cold" could be -25C in Siberia (well... not with the recent heatwave) or +25C in Quatar.

It makes sense for the whole world to use a common measuring system (and not just for temperature), perhaps even one where 1 degree is equal to 1 degree Kelvin since this would fit better with the science crowd. But whatever point of reference is chosen it will definitely fall short of expectation in different parts of the world. The freezing point of water as the 0 is as good as any. Body temperature (~37C) could also be a universal 0 with plenty of advantages and disadvantages that I never considered.

37C as the universal would put these parts permanently in negative temperatures. Freezing point of water is a good one especially when you think of the effect ice has on so many things

True, one of the disadvantages I was thinking of is truly high magnitude lows and low highs. The scale would reasonably go from -100 to +15. It would be a universally familiar 0 point. I do agree that water freezing is a very practical point or reference if there's ever been one.

I don't think body temperature is very familiar in any other context. Nothing feels special about that temperature when I walk outside; it's just somewhere in the middle of "too hot".

Room temperature has benefits as a basis point, water freezing does too, everything else is much more arbitrary.

> Nothing feels special about that temperature when I walk outside; it's just somewhere in the middle of "too hot".

It's probably worse then that since "real feel" is strongly influenced by humidity or wind. What I meant is that body temperature is more or less universally applicable to all humans, something relatable but with no benefit over the freezing point of water.

Room temperature isn't that much better. It's a bit more useful than body temperature but far less precise, could be anything between 18 and 24 C depending on who you ask.

I don't agree that is quite strong. He starts with a number of purely subjective requirements that make Fahrenheit look good and ignores all other use cases. Such as centering the zero around the freezing point of water is quite useful, since it makes it obvious when one has to deal with frozen water in various variants.

Further, the melting point of water can readily be reproduced with astonishing precision, which is often sufficient for the calibration of temperature sensors.

People cite this as a reason so often, but is it really so hard to remember the number 32? The boiling point of water is something I think more people tend to forget, but no one ever cites that as a reason to use SI... It seems like the crux of the argument is just that the US drools and Europe rules and haha aren't Americans so stupid?

Fahrenheit is honestly more useful, because 0 is uncomfortably cold and 100 is uncomfortably hot. Trying to figure out what temperature I need a jacket or AC for in Centigrade requires a lot more memorization -- and it's something I deal with far more on a day to day basis...

The argument for using SI is simply that you are a) using the same units as everyone else b) using a set of units that connect neatly together so one can actually do meaningful calculations with them. And °C is not part of SI either, but is only a constant offset away from Kelvin - which is.

And why is 0 is uncomfortably cold and 100 is uncomfortably hot more useful? Same thing as in the article.. avoiding negative numbers - why?

Because it tells me when I need a jacket or I need to turn the AC on... Things that I actually care about in my day-to-day living.

The rest of the world can do the same just fine with Celsius? And didn't really answer my question either..

When I go to buy wood I buy 2" by 4" in 4.8 metre lengths. Was buying screws the other day and I was asking for 3 and a half inch and he was saying that's 90 or something that. I had only written everything in imperial because another guy there uses that. We seem to be stuck in a half and half place.

Fasteners are categorized using a higher-level system because there are multiple, independent dimensions--thread pitch, thread width, shank width, unthread shank length, threaded shank length, head type, tip type, etc--at play, and the relationships are relatively arbitrary (AFAIU--maybe there's some sophisticated, deliberate geometry at play I'm not aware of). There's no such thing as a 3 1/2" screw--there are many different types of screws that are 3 1/2" long, and they each have a particular label. So you can't really directly compare the qualities of metric or American standard fastener sizing systems.

I must put you on to my carpenter friend then, seems he's wrong. Not in America btw.

If you do a particular type of work, such as building decks, then 3 1/2" will mean a very specific type of screw. But if you go into a good hardware store (Ace, a smaller, neighborhood chain, is often better than big-box Home Depot or Lowes in terms of variety) and tell the clerk you want a 3 1/2" long screw, things can get frustrating rather quickly unless you can provide some context. When you're working with wood, it's easier to fudge things, but when you're working with machine screws, details like pitch and shank profile often cannot be ignored.

> AFAIU--maybe there's some sophisticated, deliberate geometry at play I'm not aware of

Some earlier fastener standards were somewhat arbitrary, but mechanical engineers have calculated "optimal" values for these parameters, and e.g. the ISO metric screw thread parameters are based on such calculations.

Hah yes when I buy timber I ask for a four by two inch quote that is labeled as 50mm by 100mm planed all round and measures 45mm by 90mm. The length is usually correct.

Yeah, I live in Europe. When I go to a restaurant in North America, they have beer in 335,341,474 ml cans, draught beer in pints, 12oz, 16oz etc. Very confusing. If I ask the waiters, they also have have no cloue about what gets me the me most beer for my money. (I did check it all with units(1) now, but will probably have forgotten it next time I need it)

In my experience kilometre is pronounced “k”, centimetre is pronounced “ten mil”, and centilitre is “hundred mil”.

Yup, I'm an Aussie, born when inches and miles ruled and I live ninety two "k" from my office and I drink beer from three seventy "mil" stubbies (glass bottles). Not hard at all

In more worldly/sciencey/engineering parts of the US, I tried shortening metric like that and then I ran into this fucking abomination of a measurement, which is also abbreviated to "mil" in the US:

https://en.wikipedia.org/wiki/Thousandth_of_an_inch

I'm 100% American and I despise the imperial system so much. Maybe our plummeting stature in world politics and power will force us to convert. Not exactly the way I wanted it to go but a silver lining is still a silver lining.

I've never heard anyone say Centiliter. It's milliliters or liters. You find cl or dl on bottles but I've never encountered it in spoken language.

Which is nice because having fewer units make it easier to compare, even if converting in metric system is easy.

In some countries like Slovakia, they say deci for decilitre.

Shot glasses are 2cl around here.

I believe they're 2cl in all of Europe, never heard anyone use it outside of refering to it.

This is rather common in France. We use centiliters quite often when speaking.

Aren't all imperial units standardized on metric values nowadays? So technically Americans are using metric, they just use a different notation?

I mean in theory a foot is defined as xyz centimeters but i don't think that matters for any practical purpose.

1 foot is exactly 30.48 cm, thanks to Carl Edvard Johansson.

Yes, I think you're right.

Product specs are often in inches like 3.9 inches (which is ~ 100mm)

Although sometimes it's the opposite - computer pins are commonly 2.54 mm apart (0.1 inches)

> the US hasn’t moved towards metrification.

I couldn't care less about which system is used, i just want there to be only one. I'm not a fan of owning two sets of tools, and I'm not a fan of working on a car and coming across something I need my second tool set for because people can't figure out what system to use.

> IPv6 is great if you control lots of components that need to talk to one another.

And that's the fallacy. Those things don't need to talk to each other (my toaster doesn't need to talk to your fridge), they do talk to a well defined peer (gateway, server, whatever). There might be someday more than a few billion networked things out there, but there won't be even remotely a billion gateways on the Internet in the foreseeable future.

Enterprises generally don't like outside elements access elements within their network at will. That's not going to change, it will always be through some form of gateway.

There are still benefits to using unique address space on all networks though. The fact that my fridge doesn't need to talk to your toaster isn't a reason for us to not do v6.

If you've never had to deal with clashing RFC1918 blocks when doing a VPN or network merger, consider yourself lucky. It's a giant headache that can never be properly fixed, yet v6 completely avoids it.

I've only witnessed it on the sidelines (while the internationally distributed company I worked for was acquired).

> It's a giant headache that can never be properly fixed, yet v6 completely avoids it.

It's a headache, sure. But that it can't be fixed is hyperbole. All it takes is reassignment of IP addresses, there are plenty after all. The headache comes only from IP addresses stored in places they don't belong. If they were only stored in the DHCP and DNS servers, a script could fix this in no time. Part of the problem I witnessed was, that people are hesitant to use DHCP for static IP addresses (servers) and some devices (switches) are incapable of using DHCP.

I also observed (much earlier) misguided sysadmins using a scheme encoding location of a device in its name. That of course let to horrible long, unpronounceable names and people remembering and using rather its IP address (that danger of course is greatly reduced with IPv6 ;-}

You haven't seen the full depths of it then.

I originally wrote "a giant headache that never goes away", but changed it because it is possible to renumber the networks. But... that doesn't properly fix the problem. You'll hit the exact same issue the next time you go through a merger.

Worse, RFC1918 isn't actually that big. There are plenty of companies out there that have either exhausted it or have to be very very careful with their use of it to avoid running out. At some point you run out of space to renumber into. You're also going to have issues with e.g. VPNs to people's home networks, where renumbering isn't viable.

Reply on news.ycombinator.com