Birds supposedly have UV vision, and some of their markings are for each other, and are more clearly distinguished by other birds that can see UV.

So you might consider observing birds and making sketches of their markings, much like the naturalists of 100 years ago, so that the rest of us can see what the birds can see of each other.

This is good. I take it to really mean that we'll see the end of PayPal by 2015.

I've run my own mail and web servers from my home for years, so I was worried about this when I was making the switch from Time-Warner cable (who didn't care, BTW) to Verizon Fios. One person I know reported that Verizon did indeed block port 25 (and port 80) inbound, while another told me his setup worked fine, once he replaced the free router they gave him with his own router.

And indeed, after making the switch, I still am able to get mail inbound on port 25. The modem/router they gave me does have controls to adjust security settings, which look a lot like Windows trusted -vs- untrusted controls. But I can't really tell if they do much. The key thing I found was to set up port forwarding to send port 25 to the machine in my internal network with the mail server. Same for port 80 to the web server.

Outbound I route through the ISP. Postfix makes that easy. I found a few years ago that certain domains I sent to, though not all of them, started bouncing mail even though I had an SPF record, just because my IP address was in a range listed as "dynamic" (i.e. "residential"). Clearly that's for spam control. Not everybody does this, and I could send direct to those who do not, but it was simpler to just send everything out via port 587 to the ISP.

How about a compromise? Keep the ethX names, but use the proposed scheme to have an unambiguous order in which the names are assigned.

If you take a hash (eg SHA1 is better than MD5) of the UDID you get a unique string that is not the UDID. Of course if other apps do the same then these could be compared to identify users -- not necessarily by name, but connecting a user on one server with a user on another.

So concatenate the App Id, which is unique to the app, with the UDID, which is unique to the device, and then take the hash, which is then unique to both and not invertible. Do this once, on the device (not on your server, or the UDID has to be transmitted), and use that as a unique identifier of the user/device.

RIM needs to make a blackberry device that can use the Iridium satellite network.

I've noticed lots of slow responses or no responses from Google servers lately. (Gmail, Google Docs, Google Search, Youtube, Google Maps, etc) It doesn't happen all the time, but many many times throughout the day. Has anyone else noticed this? If so, is anything known about it? DOS, etc?

Yes, I use Google for DNS and I've notice in the past day or two that it fails for long intervals (many minutes). I also wondered about DOS.

I found the audio to be distracting, whereas the video display gives me positional awareness, and I can look at it when I choose to, not when the box decides to say something. I found I was much more relaxed when I found how to turn off the audio.

So I guess having both at the same time is the real problem.

Please don't associate BOINC with you little project. It will confuse the casual reader into thinking it is something bad rather than something good.

This will fail. A dedicated terrorist has no fear, while a law-abiding citizen can now be in fear of being accused of what is essentially thought-crime.

You'd think that somebody that's light-years ahead when it comes to parallel processing would rule the roost in the Top 500 supercomputer list. I'm sure there's a good explanation, though....just waiting to hear it. :)

To get on the Top 500 list your machine is measured against the LINPACK benchmark. It's not the best benchmark for parallel performance, so in many cases nobody has bothered.

I'm working on a science education project called I2U2, which is looking for teachers like you.

The main idea of the project is to give students (and their teachers) direct access to data from major physics experiments, along with access to grid computing resources so that they can do interesting investigations with those data. We have access to data from the CMS test-beam, as well as Monte Carlo data simulating CMS itself. We will have real data from CMS once the LHC turns on. We have access to environmental data from LIGO, the gravitational wave detection experiment (not the gravitational wave data itself, I'm afraid; but there's still cool things you can do with the seismometers and other sensors). And we have an array of several hundred cosmic ray detectors in place in schools across the US (and a few abroad) from Fermilab's QuarkNet project (http://quarknet.fnal.gov)

All this data can be used for inquiry-driven projects which the students design themselves, with guidance from their teachers and materials we are preparing for the teachers. These 'e-Labs' are not scripted labs (though we do provide a general structure for developing those investigations), they are an opportunity to do real inquiry with real data. And yes, this will include tools to let you track your students' contributions and progress.

We will be doing some teacher workshops this summer, and we need some teachers to be beta testers. If you are interested in that, or in the project in general, check us out at http://www15.i2u2.org/ We are not set up for production yet, so please excuse that it's not very polished, but it should be possible to learn a bit more about the project from that site.