From Eric Blake's bug-bash post

bash -c "export f=1 g='() {'; f() { echo 2;}; export -f f; bash -c 'echo \$f \$g; f; env | grep ^f='"

If you see anything like the following:

bash: g: line 1: syntax error: unexpected end of file
bash: error importing function definition for `g'
1
2
f=1
f=() { echo 2

you're still vulnerable. There may be other issues the above does not cover.

Wireless gizmos are becoming very common since they mean you don't need to dig holes in your walls to run the cables.

I have 2 wireless thermostats - the wireless isn't used to set them remotely, it is used for them to communicate with the boiler. On the whole they work pretty well (and yes, I'm sure the protocol is so trivial that someone could probably sit outside my house and turn the boiler on/off if they cared enough). That said, if I could point my browser at the thermostat instead of having to fiddle with a UI that has a limited display and only a few buttons, that'd be pretty useful.

I have a wireless doorbell too. It has to be said that this doesn't work so well because the range isn't great - it certainly won't reach my office. Again, probably really insecure and someone who cared enough could probably make my doorbell ring remotely.

As we get more and more wireless gizmos like this, having them all use common infrastructure, such as the wifi network, rather than communicating using their own point-to-point links is probably a pretty sensible idea - it cuts interference between devices as well as extending the range (by virtue of the wifi network usually covering the entire house anyway, so being able to relay the traffic, possibly via multiple access points). The problem here is twofold:
1. Moving from proprietary protocols to a standard protocol like wifi suddenly means off-the-shelf hardware and software can be used to attack the devices. The old proprietary devices were really insecure too, but no one cared enough to engineer hardware to attack them - now your phone or laptop comes with the hardware you need.
2. These wifi-enabled devices are more powerful and can therefore do nefarious things that the older devices couldn't do - i.e. attacking an old wireless thermostat allowed you to turn the boiler on and off, attacking a new one lets you send spam, etc.

Unfortunately warranty legislation never seems to apply to software - how often do you hear people getting their money back from Microsoft because Windows is buggy (that would be a design or manufacturing flaw, which is certainly covered for physical goods).

I'm not sure that's true - this stuff hasn't really hit the mainstream yet, but the same can be said about a lot of technology early on (how long ago was the internet "only for nerds"?)

I can certainly see a lot of uses for this stuff - my home thermostat lets me set different programs for every day, etc. but the UI isn't great and its time consuming to set. The UI deficiencies are mostly down to the fact that it has a limited display and a limited number of buttons - if I could control it from my web browser it'd be much easier to use.

I'm not entirely sure what you'd expect from an internet connected fridge - it could be useful for stuff like dynamic power use to reduce the load on the electricity grid. But a more consumer-focussed idea would be tracking what's actually in the fridge (would require RFID labelled products or similar) - I can't count the number of times I've found myself in the supermarket and thought "I wonder if we've got any milk left?", or "Is there space in the freezer for this?" - being able to easilly check that kind of thing remotely would certainly be useful. At the moment this is all in the "nerds only" stage, but how long until it integrates with your shopping list, automatically tells you what you've run out of and is used by a large chunk of the population?

I guess something that will hold back adoption of these technologies is that they are in devices that don't frequently get replaced - I've had my fridge for 14 years and I'm not planning on replacing it until it dies. But then the same could be said for TVs and a lot of people have recently replaced perfectly good CRTs with LCD smart TVs so at some point the jump in technology gets good enough for people to bite the bullet and upgrade.

Yep, managed switches seem to be outrageously power hungry. In my cabinet I've got:
  - Satellite patch panel (wired to the dish)
  - 24 port 8p8c patch panel (wired to sockets in the rest of the house/office)
  - 24 port managed gigabit switch
  - Test machine which is completely underpowered and never turned on (at some point I'll get around to removing it from the cabinet)
  - Sheevaplug
  - USB hard drive for Sheevaplug
  - VDSL modem
  - VoIP/POTS gateway
  - USB DVB-S2 receiver
  - RIPE Atlas probe
  - PoE injectors for 2 wireless APs that are dotted around the house
  - Far too many PSUs for all of the above! (Although I have consolidated all the 12v supplies into a single PSU with multiple connectors. I've still got stuff that needs 9v and 5v supplies though)

The total draw is about 90 watts, probably about 50% of it going to the managed switch! About 25% goes to the hard drive I guess.

If anyone has any recommendations for 24 port managed switches that don't draw silly amounts of power when idle, I'd be interested (bonus points if they have some PoE ports).

The other problem I have is that no devices seem to be able to roam between APs sensibly - if I move from the house to the office my phone and laptop try to hold onto the incredibly weak signal from the house AP even though there's an AP in the office for them to use. I have to toggle the wifi off and on again to get them to reassociate. (And vice-versa when I move back into the house).

What amazes me is that the lander has no RCS - it's launched at the comet, and if it bounces off or something there is no second go. I kinda expected the lander to have some RCS so it could automatically correct for unexpected troubles.

ISTR that the big datacentres, such as Google, run drives until they fail - the systems are redundent enough to cope with a failure with no problems and they have so many drives that it's more cost effective to have a resilliant system and just run the drives into the ground than it is to preemptively retire them (and still have to cope with unexpected premature failures).

Yeah, thought as much - I was surprised to see someone saying that Virgin were doing FTTP. They certainly do FTTP for commercial installs, but I've never heard of it for residential.

Or maybe its because they just aren't interested and therefore don't want to pay for a faster connection...

HD on the internet is definitely not the same as HD broadcast TV. When it was first launched, the BBC HD DVB-S channel was doing H.264 at a little over 20Mbps. I think they've reduced that a bit on the HD channels now but certainly nowhere close to 5Mbps. A quick look at a 35 minute programme recorded on my MythTV system from BBC One HD shows 2.6GB, which is a little over 10Mbps - the BBC transponders use statistical multiplexing though, so if you're watching something with more fast action then you can probably expect a higher bit rate than that though. I think BSkyB do around 8Mbps for their HD transponders (and people complain about the quality of BSkyB's HD channels).

The fact that Netflix skimp on the bandwidth a bit shouldn't really be news anyway...

That said, 4Mbps *is* enough for a lot of people - a very high proportion of people use their internet connection for a bit of web surfing and email. The proportion of people who regularly watch hour long HD streaming video channels is probably pretty low. Remember that Slashdot users aren't exactly the "typical" home internet user. (I say this having moved from a 6Mbps ADSL connection to a 40Mbps VDSL connection - for the vast majority of uses the 6Mbps connection was absolutely fine and the only real reason I upgraded was because switching ISP actually worked out cheaper than sticking with the old 6Mbps connection)

Umm, when did Virgin start doing residential FTTP?

The link I posted does mention some idiot woman who decided to chuck a bucket of water over herself while sitting on a horse. The expected thing happened - the horse bolted and the woman was killed.

http://www.telegraph.co.uk/new...

Although I was under the impression that the Leidenfrost effect only worked well on bare skin, so I'm surprised he didn't get frost burns to his scalp and clothed parts.