Geeking

I was thinking about data archival the other day because David Hagan was telling me about one time when he spoke to a group of librarians and told them to expect a data gap starting in 1950 and extending until we get serious about data preservation. According to David, the time when we stopped being able to save data was in the 50’s because xerography (which is the same as laser-printer technology) came into existence then. All other data archival techniques (magnetic storage on tape or disk, recordable CD’s, etc) are inferior to toner pressed onto paper, which is itself inferior to ink soaked into paper fibers. And, come to think of it, ink on paper is inferior to marks on clay tablets – though ink on paper has proven to be stable enough, since we use knowledge today gleaned from papyrus scrolls from thousands of years ago. ...

One of the headlines that came out of my trip to Freetown was “The Skills Exist, Use Them!” First, a bit more about what I was doing there. I was on a contract with the Health Metrics Network, which is a project running under the auspices of the World Health Organization, but funded by the Gates Foundation. HMN is supposed to go around building a network, not of routers and switches, but people who understand the challenges of constructing and running health information systems. Those people should be researchers, technologists, and public health people. And those people should be a mix of those talking about it (easy to find in Geneva, they are douze pour une centime) and those doing it. The latter are a little harder to find… ...

I am back from Freetown, freezing in Leeds. I’ve spent today trying to get software working including reinstalling Chrome (which started marking all words as misspelled) and Word Press which inexplicably started hanging. I’ll write a note, maybe this weekend, about what I was doing in Freetown. It was fun, and we got a lot accomplished, so it’s a worthwhile story to tell.

There’s a nifty feature hidden in Gnome that’s interesting to play with. Click on the date in the upper right. You get an applet with a calendar. Below the calendar is “Locations”, click on that. Nifty little map showing the line where sunset and sunrise are. Click on edit, and you get to add your favorite locations to the map. But not just your favorite timezones, you get to pick from a huge list of geographic places. My hometown, little Roseburg Oregon is in there. African capitals are in there, though sadly Monrovia got missed. And when you add the clocks, you get weather too! ...

I started a two week trip to Freetown, Sierra Leone yesterday. I am here as an IT consultant for the Health Metrics Network, part of the World Health Organization. We are taking a bunch of random hardware that the government bought with a not-quite-complete architecture and putting in place a foundation for IT services related to the Health Information System. Of course, the HIS relies on lots and lots of low tech forms and paperwork “up country”, but there’s also a need for a datacenter in the capital. ...

From Slashdot: Corporate Data Centers As Ethernet’s Next Frontier Metcalfe’s (apophrycal) Second Law: Never Bet Against Ethernet ‘Nuff said.

The Long Now foundation took a field trip recently. The pictures are cool, as is the irony that they went to a trade show held once every four years. Miners have a time scale closer to the Long Now’s. What does it mean that mining is a slowly changing industry? The economics of extractive industries push them that direction (capital investment is only warranted by a positive outlook in the commodity price). But workplace safety, which is a special case of universal human rights, demands that industries take advantage of the best available technology as fast as possible. This is an interesting story of the tension between capitalism and social welfare. If a mining regulator pushes an industry too much it will give up and move elsewhere. And apparently the mining industry is especially prone to regulatory capture an euphemism for corruption I especially like (like as in hate; in newspeak up is down, enslaved is free and corruption is regulatory capture). ...

Some stuff I discovered playing with Google Chrome: View source works on the various “internal” pages (i.e. the home page, the about:stats page, etc) so you can see how the magic happens Read the comments in the home page to see how absolutely nutso over-the-top the programmer was about load time. He obsesses about the difference between 20ms and 100ms. You go, dude… make it fast! Also in the home page, you can see that internal UI pages have access to a Javascript object called “chrome” that tells the page what to put on itself. (Attention bad guys: this is an attack surface. If you can find a way to get a handle to a chrome object, baddness ensues. So, please, don’t do that. Thanks.) Embedded images, for example the logo on the home page, use the “data:” URL scheme to embed the picture’s data right into the page. Why? Performance? Or simplification. If view source doesn’t work (like in the UI for “error 404”), then the inspector works, and will show you the source. (What’s the point of disabling view source? Setec Astronomy.) In the inspector, you can click on things in the source and change them into input boxes, then you can change the source on the fly, and the results are shown in the page. Cool. In about:stats, you can sort the columns by clicking on them. View source for a really nifty piece of Javascript I might lift and use someplace else myself. About:network is totally cool. Click “Start I/O tracking” then go to another tab and load something. You get a trace of all the transactions. You can also click “Show current I/O status” to see which HTTP connections are open. You have to type the about:foo URLs, if you click on them you don’t get anything. They are checking the referer. Props to Stikiflem and Redneck Programmer who blazed this trail before me. ...

Yahoo published some best practices on OpenID, but I have one they forgot. Imran and I were talking at OpenCoffee Leeds the other day and we independently and together came up with this realization: OpenID providers should be required to be OpenID consumers first. Why? Well, the problem OpenID is trying to solve is “too many usernames”, which really equates to “too many authentication providers”. That means that the last thing we need is more stinkin’ authentication providers. OpenID providers who are not also OpenID consumers, are just making the problem worse. ...

gPXE and dnsmasq do not get along, beware. gPXE wants to find the boot filename in the dedicated slot in the reply packet. dnsmasq does something clever/stupid (I’m not sure which) and puts it someplace else. You have to use the --dhcp-no-override option to dnsmasq to make gPXE see the filename. I don’t have proof, but I suspect this inflexibility in gPXE also makes it incompatible with the DHCP replies sent by qemu’s built-in DHCP server. ...