What about things like the identities of spies, plans for future military operations, and details of ongoing criminal investigations?

It'd be nice to have fewer secrets, but it's completely unrealistic to argue that we should have none at all.

> In reality, the shift from hard drives (HDDs) to SSDs has thus far been confined to the upper end of the PC market.

Not entirely. I have the cheapest netbook I could find, and I replaced its hard drive with a cheap low-capacity SSD. I don't keep much big stuff on it so the capacity isn't a problem. In terms of performance and power usage and not having to worry about my data when I drop my computer, it's been entirely worth it.

Sorry for the typo in the headline. Should be "Australia rules..."

> dead links wouldn't be a huge problem because the only links you would need to maintain are with the major search engines anyway

Seriously? The great innovation of hypertext was that websites can link to each other, and it's used all the time. Slashdot itself is a pretty good example. Almost every site, not just search engines and Facebook, has links to other domains on it.

If you remove the ability for websites to link to one another reliably, you kill the web. I am not exaggerating.

You know you can trust mybank.com because it's a memorable name and you've used it before. And once you know mybank.com, you know that subdomains like accounts.mybank.com or invest.mybank.com or whatever also belong to your bank.

It's not so easy to remember your bank's IP address, and it's impossible to tell whether an IP address you haven't seen before belongs to them or not.

This idea has many problems:

You can't change your ISP, or renumber your network, or move your website to a different server on your network, or switch to IPv6, without making all existing links to your site invalid.

A link can only point to a specific IP, not to a website that has multiple redundant servers with different public IPs, or a website with both IPv4 and IPv6 support.

Anyone can create a site like (for example) http://203.0.113.135//westpac.com , and no user can distinguish it from the 'real' westpac.com without consulting some authority. This makes phishing easy.

Also, I think you underestimate the importance of being able to remember a site's address and tell people what it is. Even ICANN is better than having to look up a directory of IP addresses manually.

We all know the new top-level domains (and some of the existing top-level domains) are basically a money grab and a way to force people to pay as many times as possible for their name.

And the registrar system, which supposedly enables competition, is also just a money grab. For each top-level domain we have one registry, which is a simple database run by one organisation, but then we have a whole lot of commercial infrastructure and multiple companies around it which serve no purpose except to skim profits off the top.

Now the problems with the new TLD registration process are starting to make ICANN and the domain industry look incompetent as well as greedy, for those of us who hadn't decided that was the case already.

So, what can we do? I know it's been suggested and unsuccessfully tried before, but is it time someone replaced ICANN?

People keep suggesting decentralised DNS, but I'm not convinced it's a workable solution. If there's no central authority controlling the DNS, there's nobody who can give your domain back when someone breaks into your system and steals it, or when you accidentally lose your crypto keys.

> First, the very fact the DNS is the way it is now, subdomains.domain.tlds, is their fault.

This was decided at a meeting in 1982. The minutes of this meeting are available as RFC 805. As this decision was made before ICANN and the domain name industry existed, it would be wrong to blame them.

You're right that the ordering of the DNS names is inconsistent with many other naming systems. It seems to me that the rationale (which made a lot of sense at the time) was that you can email someone at a local host the same way you always could, user@host, and you can email someone at another domain by just affixing the domain to their email address, user@host.domain. Makes more sense than sticking something in the middle and having user@domain.host.

> [ccTLD] shorts to your local country-code if it is omitted.

I think this is a horrible idea which would encourage the fragmentation of the global Internet, create many name conflicts and create a huge opportunity for phishing attacks. A URL referring to a public website should point to the same website no matter where it's accessed from.

We all know the new top-level domains (and some of the existing top-level domains) are basically a money grab and a way to force people to pay as many times as possible for their name.

And the registrar system, which supposedly enables competition, is also just a money grab. For each top-level domain we have one registry, which is a simple database run by one organisation, but then we have a whole lot of commercial infrastructure and multiple companies around it which serve no purpose except to skim profits off the top.

Now the problems with the new TLD registration process are starting to make ICANN and the domain industry look incompetent as well as greedy, for those of us who hadn't decided that was the case already.

So, what can we do? I know it's been suggested and unsuccessfully tried before, but is it time someone replaced ICANN?

Where's this "virtual economy" money is being pumped into? Doesn't the profit from HFT go to the owners of the HFT company, who then invest or spend it? It's not just disappearing down a black hole, it's part of the same economy as the rest of us.

Basically every simulation is "only partially correct", but by comparing simulations to experiments we can get a good idea of how accurate the simulation is and at least use it as a guide for future experiments.

Obviously nobody's suggesting that we go straight from simulating a drug to selling it without testing it first.

Because this isn't just about a lack of coverage or network capacity. There are actual 4G LTE networks in Australia, and Apple was selling a device that wouldn't connect to them but advertising it as 4G.

I was surprised to read that they still use a device with a needle to play these. I would have thought that they'd be scanned with lasers, to avoid wear entirely and possibly to reconstruct the groove more precisely.