The sanctions imposed on South Africa in the 1980s absolutely crippled the country and are widely seen as a successful contribution in the efforts of bringing down Apartheid.
> The sanctions imposed on South Africa in the 1980s absolutely crippled the country and are widely seen as a successful contribution in the efforts of bringing down Apartheid.
They played a role, but Western institutions are biased to exaggerate it and gloss over the fact that the sanctions regimes were only adopted after the reinvigoration of the armed struggle fueled by stepped up material and training support from the USSR brought into serious doubt the survival of the South African regime even with continued active support.
The West didn't want South Africa to lose, and wouldn't concede to it losing until the alternative was it losing anyway and armed struggle and support of the Communist bloc being the entire narrative for why, which they didn't like for either international or domestic political optics.
> In this case Google provided a reason - a ToS violation.
When the ToS are 15 pages long this is about as useful as hearing "You're being arrested for breaking the law" when you're in the back of a cop car. Doesn't really narrow it down and provides you no way of actually defending yourself.
I agree that being too specific can help bots but the current way of handling these things is obviously flawed.
Can't check other countries since Google automatically adjusts the country version to your location but you can check yours here: https://policies.google.com/terms
//edit: but you're correct considering this doesn't contain any service-specific ToS.
Not a tautology but yes, that is essentially what I am saying.
I just found it weird that people seem to be focusing on all sorts of factors like family structure, median income, winter holidays, Fasching/Karneval celebrations etc (which of course also play a role) while completely ignoring that higher population density facilitates the spread of a novel virus.
Assuming what is and isn't valid for user inputs is a dangerous game because there are always exceptions.
I ran into a similar issue with many online retailers when I was living in the inner city of Mannheim, Germany because a lot of online systems make assumptions on how a valid address looks.
Addresses in Mannheim's inner city follow the format "Char Number, Number". "A1,1" is a valid address if you want to send a letter to the district court. A1 being the city block the court is located at and 1 being the house number within that block.
I didn't get to do a lot of online shopping for years when I lived there.
Simple workaround would be putting in a human-readable long string. Anything like, "Daniel's residence, A1,1". Post people know how to read.
Upd: reaction to this suggestion shows that some people don't understand how post office operates. They go to great lengths to understand where to deliver the mail/parcel. In most cases, addresses like "big yellow house with a red door overlooking the cliff near the lighthouse" would work. So the only challenge here is to get past the whatever dumb rule the service developer imposed on the address format. Likely it is just filter by string length.
Post people know how to read, but I think now most nail sorting/routing is done with computers and OCR. I sometimes get mail addressed to people who used to live at my address but have long since moved. I tried writing “return to sender, addressee not at this address” or similar things but the mail kept coming back to me. I finally went into the post office and they said that the machines would just rescan the address and send it right back to my address for delivery. So I think relying on postal employees to see/interpret things on address labels is no longer a viable approach in many places.
No, not really. Sorting is always manual when automation fails.
In your case, automation actually didn't fail, it just didn't recognize your additional instruction. Probably, you could have just patched the address with an easily removable piece of tape and that would definitely trigger a human attention, and delivery would go where it should
-My parents (living in rural Norway) once had a postcard delivered where the address given was simply their first names - no last name, no street, no town, no nothing.
Having a database in which every citizen's domicile is registered does have its occasional advantages.
Similarly, there was a story going round a few years back about mail being delivered in Iceland where instead of an address there was a map to the house to be delivered: http://i.imgur.com/1GVjLKF.jpg
I live in Japan. I once had a package delivered from overseas where their printers couldn’t print CJK fonts and thus the whole address resulted in just small empty boxes.
The post office inferred my address from the post code + my name and delivered it correctly. There wasn’t even a (noticeable) delay.
> Probably, you could have just patched the address with an easily removable piece of tape and that would definitely trigger a human attention, and delivery would go where it should
Fair point. And that was the advice given to me by the postal service.
Cross through the wrong address. Every year or two somebody from the management agency tidies the noticeboard for the building I live in and removes my hand written sign explaining how this works. Then, next September/October when lots of people move in (some fraction of the occupants are students) the noticeboard gets envelopes pinned to it with undeliverable mail. I write a fresh sign.
The sign is a flowchart, it says first, is this mail for a different address? If so, either redeliver it (duh) or write "Misdelivered" in bold leters and put it into any postbox.
If not, but you don't recognise the recipient, strike through the whole address in black pen and write clearly "Not at this address" then put it into the postbox.
This won't stop you getting more mail by the way, I still get letters labelled "Urgent" with the name of the previous owner years after I bought this place. But it does stop literally the same mail coming back since the OCR will reject the crossed out address -- it's just that the sender may not have any effective process for what to do when they get the mail back undeliverable.
Define 'accuracy' though. That's what this whole thread is about. When what you're comparing against is itself wrong, what does accuracy mean? Also, "big yellow house with a red door" etc. is in fact accurate for that place.
That's the whole issue: a random coder decides "and this is my idea what's acceptable: Google Maps/whatever finds it from the input string; worksforme, done!" without second thought or even authority to make such decision; this, an operational decision, gradually becomes doctrine, even dogma.
A friend of mine living there works around it by using „Quadrat A1 1“ (translated: square A1 1 - since most blocks are roughly a square in Mannheim) and it seems to work okay.
But the naming in Mannheim causes a lot of issues, I remember early navigation systems having a hard time with the format. A IIRC TomTom even crashed when trying to announce the street.
Why not have a click-through at install or on version upgrades that says something like "we want to install an easy way to add products to your search results for which we get payment: Allow, Deny" ... rather than default install something that is going to cause such negative press.
IFF then there is a wide-scale uptake then there is warrant for auto-install of those features.
The iPhone App is the best example of this waste of space.
That banner image takes up about 1/3 of the screen, leaving 2/3 for UI elements, actual information like follower/following and tweet count and one single tweet.
It's not only a waste of space if you are already following someone but also if you want to discover new people to start following.
What users are talking about should be the primary focus in every case - whether you are already following the user or not.
The feature list of the paid version does not mention removing advertisements though. Only virus protection, the ability to play more file types, transcoding and remote access.
"Only virus protection, the ability to play more file types, transcoding and remote access."
Which I would rather pay to NOT have included in my version. I had been using microTorrent for what its name implied, for being micro.
IIRC, the first torrent client was written in python. Then ported to Java. The first implementations where so bulky that people hated those. uTorrent probably was the first port in C and everyone loved it for being light weight. Hence the "micro".
And all that is (or should be) completely irrelevant since we are talking about a British guy hosting a website that not primarily targeted US audience.
I am not a lawyer but I am fairly sure O'Dwyer does not have to abide the Digital Millenium Copyright Act. Google, as an American Website, however has to.
At that level of income, a significant part of his audience could be proven to be from US. It would also matter if he used US servers, US registrars (.net is controlled by Verisign which is an American corporation) and so on, but I doubt prosecutors would have much issues in proving US jurisdiction once he's in USA.
What you say makes sense, assuming that jurisdiction is clearly with the country that is requesting extradition.
What I fail to understand is why the US has jurisdiction here. What brings this crime to the US? If it is just that the US is on the Internet too, then the logical extension would be to pass this guy around every country in the world, which is clearly ridiculous.
He should be answerable to the law. But the UK seems like the appropriate jurisdiction here, not the US. If the problem is that UK law doesn't cover this case, then that is what should be questioned, rather than pretending that the US has jurisdiction in order to work around the rule of law.
