gregholmberg

[0] netbsd.org/ports

I have no experience with balloons at all, but even a plowed field feels like a sidewalk to me when I run into it at the same speed under canopy. I leave a sizable mark, and I have seen friends break femurs, pelvises, etc at the end of a similar descent.

http://www.theguardian.com/technology/gallery/2013/jun/16/go...

Will this be the decade where I get to experience being struck by falling routers and switches, far from any data center?

(EDIT: Simply remarking on novelty of problem and upward growth of network. No criticism of any firm implied or intended.)

If not for the patient and curious scribes of medieval Islam, many significant works from Ancient Greece and Rome would have been lost forever.

    Mo: un méga-octet, un million d'octets
  Mo/s: méga-octet par seconde

There also seems to be variation between Wikipedia articles in different languages that I would not ascribe to incompleteness. Does generally-accepted cultural "truth" about a topic or incident really differ so much?

A truly delightful name for such a dreadful place.

One argument in favor of capping work hours: unrestricted overtime discriminates against workers unable or unwilling to compete. A race to the bottom is not in the long-term best interests of any participant or the society.

An imbalance in society might show up as (under|un)employment in groups outside the core demographic -- in this case, anyone who is non-20-something, non-male, or non-white, who is not "crushing it" at work after nine on weeknights.

"Caretos are masked young men dressed in suits made of yellow, red, black, blue and green fringe wool quilts, wearing brass, leather or wooden masks and rattles in their belts. ... They appear in groups from every corner of the village running and shouting excitedly, frightening the people and “robbing” all the wineries." [0]

The trojan referred to in this story is known by another name. [1]

[0] https://en.wikipedia.org/wiki/Careto

[1] https://en.wikipedia.org/wiki/The_Mask_(malware)

The concept of allowing evidence from secret sources should offend anyone who believes in a just society.

Nine results were about the space-partitioning data structure. Also, there is a tree-trimming service on Long Island called K&D Tree Masters.

Surprisingly, pleasantly -- all ten items are useful to me.

Naturally, it is doomed.

  [pdf] http://www.sanders.senate.gov/imo/media/doc/Restore%20Our%20Privacy%20Act.pdf

Here is the entire text of the Patriot Act.

  [pdf] http://www.gpo.gov/fdsys/pkg/PLAW-107publ56/pdf/PLAW-107publ56.pdf
        http://thomas.loc.gov/cgi-bin/query/z?c107:H.R.3162.ENR:

        http://www.aclu.org/free-speech-national-security-technology-and-liberty/reform-patriot-act-section-215

        https://en.wikipedia.org/wiki/Foreign_Intelligence_Surveillance_Act#Provisions

In radioastronomy, there are arrays of dishes that use this technique. There are good pictures in the Wikipedia article. [0]

One military implementation was a connected three-satellite constellation, built by the US Navy for scanning the surface of the world's oceans. [1]

But those tools work in (relatively) long wavelengths.

I have never heard of free-flight multi-aperature interferometry that works in optical or near-optical wavelengths.

.

[0] "Very-long-baseline interferometry" https://en.wikipedia.org/wiki/Very-long-baseline_interferome...

[0] "How VLBI Works" https://en.wikipedia.org/wiki/Very-long-baseline_interferome...

[1] "White Cloud | PARCAE | NOSS" https://en.wikipedia.org/wiki/Naval_Ocean_Surveillance_Syste...

I used to love to go to the Roxie Theater in the Mission, where they showed movies you might never encounter otherwise. During a retrospective for some anniversary of a Friedkin film, the great director himself just came walking up the sidewalk right through his speechless fans.

On the other hand, I have been stuck in traffic on Mission near 17th while the SFPD put down rows of little yellow numbered plastic cones across three lanes of traffic -- one for each shell casing.

Is there any public transit facility less welcoming than the urine-drenched BART station slash homeless zoo at 16th? Yes, actually. Just pop down to 24th, if you dare.

Heroin dealers used to turn little alleys off of 26th into open-air markets in the wee hours. Maybe they still do.

I just can't bring myself to visit the Mission anymore.

[0] "Your Drive to Compete May Come Down to Dopamine" http://healthland.time.com/2012/05/02/your-drive-to-compete-...

[1] "Amphetamine Spurs Slackers to Work and Workers to Slack — at Least For Rats" http://healthland.time.com/2012/03/29/amphetamine-spurs-slac...

.

... some research by Michael T. Treadway and nine others.

[2] "Dopaminergic Mechanisms of Individual Differences in Human Effort-Based Decision-Making" http://www.jneurosci.org/content/32/18/6170.full

> " ... We found that individual differences in [dopamine] function in the left striatum and ventromedial prefrontal cortex were correlated with a willingness to expend greater effort for larger rewards, particularly when probability of reward receipt was low. ... "

You could argue that this mechanism is a neurochemical amplifier of entrepreneurial behavior.

I admire James. Wading into the interpreter takes balls.

I have been running Node.js and Lua at fairly large scale on EC2 recently.

I like the V8/Lua/Luajit platforms just fine. Each has some nice strengths.

With Lua, I see significantly less memory use, and moderately increased performance.

As long as RAM is the expensive bottleneck for hosted servers, it looks like I will be able to get better performance (in operations per dollar) by using a smaller runtime.

Why not client-side rendering?

Accessibility? Cross-platform uniformity? (better answer from 'capnrefsmmat: render once, then cache massively)

He was a widely known and well-regarded big wall guy who had freed gigantic routes in Pakistan, Greenland and the Yukon. I read his trip reports with awe. I would have felt out of place carrying his rack from the truck to a picnic table, much less talking climbing with him.

I had not heard that he was tidying up. He was also reputed to be very generous with his time and expertise. I find the idea of him carefully cleaning up a new route (while passing an increasing load through a frayed harness) almost intolerably painful to think about.

http://www.sfgate.com/news/article/Fatal-fall-at-Yosemite-sh...

There are also errors in judgment, but the gear is the center of this story.

The final paragraphs are still hard for me to read, six years on.

edit: adding National Park Service summaries

.

NPS Morning Report for October 26, 2006

  Yosemite National Park (CA)
  Noted Climber Falls To His Death

Todd Skinner, 47, of Lander, Wyoming, was climbing a route on the Leaning Tower when he fell approximately 500 feet to his death. Skinner's climbing partner reported the fall at around 4 p.m.

Skinner was pronounced dead at the scene. He is survived by his wife and three children. The exact cause of death is under investigation.

[Submitted by Adrienne Freeman, Public Affairs]

.

NPS Morning Report for October 30, 2006

  Yosemite National Park (CA)
  Follow-up On Fatal Climbing Fall

Jim Hewitt reported that he and his partner, well-known climber Todd Skinner, had been working on a first free ascent of the "Jesus Built My Hotrod" route on the overhanging west face of the Leaning Tower. Skinner's fall occurred when he was rappelling.

Hewitt told investigators that he had been above Skinner when he fell. As he was rappelling on the low-stretch ropes that they had fixed on the route, Hewitt came to Skinner's Grigri descent device on the rope at the point where he had fallen.

The Grigri had a still-locked carabiner attached which had been connected to Skinner's harness. When Skinner's body was recovered, the belay loop on his harness was missing.

The next day, rangers recovered a broken harness belay loop in vegetation at the base of the wall. It was very worn at the spot where the break had occurred.

Hewitt later told investigators that Skinner was aware that the belay loop on his harness was in a weakened condition prior to the climb, and that they had talked about its poor condition three days earlier. ... .

[Submitted by Keith Lober, Emergency Services Coordinator]

If you don't work in network security, you might find it unsettling to see just how much additional trouble a person can cause by having physical access to the hardware itself -- the cables and ports and LCD front panels and the like.

As an example, here's a dirty secret: in quite a few of the large, institutional settings I have had access to, the hash of the IOS enable password is stored on local flash inside the machine, and set to the same string across many core devices. This means that if you can compromise one switch (perhaps a small one in a basement closet), you could also have privileged access to larger switches deep inside data centers on the same campus.

Compromising the first switch is much easier if you can attach a serial console and reboot it at will. If I were serious about doing something like this, I might even bring an extra switch along to substitute in, so the regular users of the network would see no downtime.

Groups of switches inside a data center (when viewed with eyeballs) have a kind of tedious homogeneity to them. Generic faceplates all in rows, kudzu of brightly-colored generic cables fanning out in every direction, armies of green LEDs flashing with traffic, thick black ropes of power cables in back ready to wiggle loose from a stray nudge. Aloof. Opaque.

The traffic to and from each data center switchport, though, is often highly individual. Many times it is deadly dull for port after port after port. But sometimes, you see that you are watching a machine that appears to be processing payroll. Or saving a series of very expensive and proprietary chip masks to some huge file server. Or, best of all, you might see millions of rows of data describing those things and more, all being stored as tidy SQL.

So yes, I draw the line somewhere short of allowing homeless people into a space where they would be sleeping next to network devices with important roles.

Instead, I saw a small carpeted room containing a half-full rack of telecom gear, featuring several Ethernet switches providing 100 or so switchports for end users, punchdown blocks for terminating phone service to a similar number of incoming lines, and a small number of switchports on what looks like an administratively privileged network via a second smaller switch.

The list price of the larger Cisco with all three power supplies and several 24-port GigE cards was at least $15,000 the last time I had to buy one.

The fiber uplinks to other rooms (provisioned like this one, or better, typically one per wing on each floor of a large building) are likely to carry some very interesting traffic -- not just between end users and their preferred servers, but between the large switches themselves, possibly even routing outbound traffic for the "administrative" switch, as well.

I sometimes use separate "control plane" switched media to access "remote power strips". These allow an admin to remain seated at a desk while rebooting machines all over the campus.

Allowing unrestricted access to a storage closet containing that much gear (uninstalled) is irresponsible. Theft is likely.

Allowing unrestricted access to a wiring closet containing that much gear (provisioned, configured, and running in production mode) is a hilarious wtf. The imagination soars ...

Allowing unrestricted physical access to any administrative switch that carries traffic for power-cycling campus equipment on and off remotely is a fairly serious oversight, and not in the least bit hilarious.

edit: It looks like the photos show two different rooms. The wiring closet itself has a bare concrete floor.

http://en.wikipedia.org/wiki/List_of_cognitive_biases

.

Daniel Kahneman (prominent psychologist, 2002 Nobel in Economics) helped develop "reference class forecasting" to avoid poor project management decisions on novel or unusual projects.

http://en.wikipedia.org/wiki/Reference_class_forecasting