Hacker News new | past | comments | ask | show | jobs | submit | callum85's comments login

That article is bullshit. You can't judge gov.uk by its homepage, because that's almost never where you land.

Now that we have gov.uk, whenever I google something about doing my tax, or registering to vote at a new address, or any other arduous governmental admin shit, I usually end up on a beautifully simple and focused gov.uk page, and I'm in and out in 2 minutes. That's what a government website is about.

I agree the homepage looks like a parked domain but whatever. Take a random page like this: https://www.gov.uk/alter-a-will-after-a-death

Look at how brief and clear the writing is. Look at how straightforward its URL is. Have a look at some more pages, look how consistent the design language is. This is exactly what a government website should be like.


"I guess I need to apologize to my parents, friends, and the staff at Chuck E. Cheese's for all the times I called the cops on them."

— title text from today's XCKD http://xkcd.com/1581/


Not a lawyer, but here's how I see it:

Lawsuit: "You approached me and gave me false information that has cost me money. Pay me to cover all my losses resulting from this false information."

How (or even whether) the defendent benefited from the losses doesn't even need to come into it, as far as I can tell.


It's a bit self conscious, but it's good natured and sincere, certainly not pretentious.


Getting slightly off topic, but I'm confused about an aspect of server push (maybe you can help). Even on HTTP/1, the client typically only downloads assets on the first page load for a given website, then it caches them. Subsequent requests for other pages are much quicker because the assets don't need to be downloaded again. With server push, would the server typically send all linked assets on every request? Or can it somehow keep track of which clients have cached which assets, and avoid wasting bandwidth?


Yes that would be one possible implementation. It's up to the server and can vary quite a bit depending on the scenario, which is why Nginx hasn't tackled the problem yet.

I believe that it's going to take a while to find good heuristics that work in most of the cases. Until then only big companies like Google will have the financial incentive to develop their own custom version that's tailored to their specific use-case.


The browser can cancel the stream after the first "stream open" frame (assuming latency isn't too high, obviously)


I thought the above comment was funny and harmless. Why so harsh?


Don't worry about it. I rolled a 16 on my accept criticism check. According to the table, I shed a single tear, that slowly rolled down my cheek, then I got over it forever.~


Those are metaphors and metonyms. MS Word is not a word. MS Office is not an office. Facebook Messenger is literally a thing that carries messages. Apple Watch is literally a watch.


Although SQL Server is an SQL server :)


To be fair, MS Word is two letters, a space and a word.


Apple Watch is to watches as iPhone is to phones. In both cases they are supersets of what their name suggests, where the consumed set is a tiny fraction of their purpose.


Here's a good comment from a reader, which I think addresses a misunderstanding behind this article about what programming actually is:

> The thing that takes time is humans understanding what they want. Programmers are like therapists who help them crystalize their thoughts to the point where they are at least vaguely consistent. The program is just a representation of that idea.


Yet the widely recognized most effective way to crystallize the idea, is not via therapy but through rounds of prototyping. Now only if prototyping is as easy as snap a finger.

Coding automation too will help to make better software.

What programmers have done is to translate human ideas into machine understandable symbols, or graphic layout. Not he latter is probably far more difficult than the former which itself can be as hard as NLP and essentially the goal of the almost entire computer language evolution, to improve the effectiveness of translations human ideas to computer understandable symbols.


This is disappointing. This screen is 7" at 800x480, so its sharpness is about 133 PPI (pixels per inch).

For comparison, my original Android G1 (several years ago) was 180 PPI, and it looked shit.

This is $60 plus taxes and shipping. I just found a 7-inch tablet for £28 ($43.10) on Amazon (plus a camera and RAM and stuff). Including taxes and shipping. Why is this so expensive?


It's expensive because the economies of building electronics are weird. As they said in their announcement, a big challenge is getting a display which is guaranteed to be available for a long period of time. That means not buying the bleeding edge, but something designed for lower volume markets. If you want something that's available for a long time, the likelihood is that it's already been available for a while. If you want the bleeding edge, it means they'll stop making it and start making the new bleeding edge in 18 months (with subtly different connectors, sizes, drivers etc)

The vast majority of the cost of this display is driven by factors unrelated to the technology level. Labour to build it, logistical costs, EMC qualification etc. That's the case for almost all cheap consumer electronics.


>I just found a 7-inch tablet for £28 ($43.10) on Amazon (plus a camera and RAM and stuff). Including taxes and shipping. Why is this so expensive?

Sales volumes determine economies of scale, it is likely the £28 tablet was made in quantities much larger than the pi screen (I'd guess at least 2 orders of magnitude).

Also, not all screens are equal. It's not just resolution: there's colour reproduction, viewing angles, brightness, contrast and response time. On the non-technical side mentioned on the blog, they mentioned they wanted a manufacture who would make the panel for a long time. I would bet a dollar that Pi screen beats the £28-tablet display on all the above parameters


> there's colour reproduction, viewing angles, brightness, contrast and response time.

Which seem to be all crap for this panel anyway – 70° viewing angles implies it's the cheapest TN panel they could find.

For a non-profit(!) like the RPi foundation that can neither guarantee sales nor buy them in advance in massive bulks, availability is the only factor that really matters.


Indeed. They mention it's an "industrial" panel, which in my experience seems to mean "low contrast and brightness, narrow viewing angle, but wider temperature range[1]". The Innolux panel I mentioned in another comment here (https://news.ycombinator.com/item?id=10185433 ) is an example of this type.

[1] http://www.pacificdisplay.com/lcd_temp_range.htm


But the £28 one comes with RAM, CPU, wifi sensor, camera, and it's two thirds of the price. Even if the other screen specs (brightness, viewing angle etc) are really terrible on the £28-tablet and incredible on the raspberry pi display, this still seems very expensive, and the difference seems like more than just an economy of scale thing.


I mentioned 3 things that might explain the price difference in my comment, not just economies of scale. I should have been clearer

1. Economies of scale

2. Difference in quality of the display

3. Booking factory capacity in advance (there's likely a huge opportunity cost to the manufacturer)


I bought one of those £28 Android tablets from Amazon to experiment with for work a while back. The thing was utterly unusable - it turns out the hardware you can get for £28 is just barely enough to boot to the Android launcher. Don't even think about doing anything crazy like opening a web browser, or reading an eBook on it.


You had poor luck; I've used one of those for a couple of months and all my heavier apps (Firefox, Aldiko and Komik Reader) worked fine, albeit the memory pressure was felt (background apps were immediately dropped when opening a new one). You can easily get a tablet with 1GB of RAM for less than £28, which is equivalent to my Nexus 7 (2012), which I still use every day.


> I just found a 7-inch tablet for £28 ($43.10) on Amazon (

Good luck taking the screen apart and making it work on Raspberry Pi.


This seems so common with prototype parts. I'm not sure what the difficulty is building a sheild which you can just drop on a commodity component.

Mass producing a PCB is 'fairly' simple these days assuming the quantity is sufficient.

Anyone have any insights into the cost?


It's a Pi accessory, and one where the RPi Foundation have a lock on the market - because all display output configuration goes through the binary GPU blob, they control what displays you can use with your Pi.


There's no way to take the surface of a sphere and flatten it out into a rectangle without distorting it in some way. It's called 'projecting' from 3D to 2D.

The projection used for the underlying map here is the Mercator, which stretches everything near the poles. It makes Greenland look almost as big as Africa, which it is not. It's the most common projection, but it's awful for accurately visualising the whole world at once.

But the Mercator is actually a really good projection for stuff like Google Maps, because it keeps shape and direction (not scale) accurate all over the world. Say you zoom into Iceland and then Ecuador... OK you'll need a different zoom level to get to the same real-world scale (e.g. to get to 10 miles per inch you need to zoom in closer to Ecuador than Iceland) but who cares. What matters is that, in either location, north is pretty much straight up, east is right, south is down and west is left. And also, a 10 mile road going east–west will look about the same length as a 10 mile road going north–south.

Other projections don't have these qualities, but are better for keeping overall scale the same all over the map, making relative sizes of countries look more like what you get on a 3-D globe. But most people just use Mercator for everything.


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

Search: