Adidas KST-30 (19504) sport sunglasses nose bridge

Note: This post is not meant to be read by humans, unless said humans happen to have a very particular malfunction in their Adidas sport sunglasses.

So my Adidas a124 Gazelle sunglasses broke. Specifically, the rubber nose bridge part came unglued from one of its support clips. I tried to re-glue it, but gluing rubber to plastic is … tricky, at best. I did, however manage to dig up the part designator, KST-30 (KST-30-1 for the extended nose bridge, which is basically just the same thing but scaled up in every dimension). There’s also a 5-digit product code, 19504 (or 19304 or 19301 or 19501 depending on color, size, etc.)

The Adidas phone sales, Adidas-branded stores/outlets, and are all run by a single enterprise … and this enterprise has never heard of a replacement nose bridge. However, I was sure that it was available for purchase somehow, because places like this were selling them … in the UK, and also Germany.

Rather than trying to get it shipped internationally I used the storefinder (not to locate the nearest authorized Adidas optical reseller. I showed up at Optical Shop of Westport, and they pulled the bridge off of one of their display models, then called up Adidas to order a new one and find out how much to charge me. They said it was $30 … almost as much as getting it shipped from the UK, but a lot less than a new pair (~$130).

I hope they were running at a profit.

P.S. Today was the first real classic bright icy Boston winter day … a perfect day for biking in close-fitting sunglasses.

QR Codes on the N950, MeeScan, and libdmtx0

The QR Code reader for Meego Harmattan is Nokia’s MeeScan. This app installs fine out of the box on a Nokia N9, but on an N950 it hits one of the small differences between the two systems. The N950’s repositories do not contain an entry for Nokia’s patched libdmtx, which is packaged as libdmtx0.

You can install libdmtx0 on an N950 by temporarily adding the N9 repositories, in something like a 5-step procedure (since you should probably remove them as soon as you’re done, to avoid conflicts).

Lucky for you, N950 owner, I’ve sped up the process by pulling out the .deb for libdmtx0 and putting it right here. Just “dpkg -i” it as root, and then MeeScan will happily install and run correctly from the app store.

Three cheers for the redistributability of free software.


Yesterday I swung by Staples High School, where I graduated nigh on 10 years ago. A few of my teachers are still around, so I checked in for the latest on the school.

When I went to Staples there were two full-time physics teachers, and one section of multivariable calculus. Today there are seven physics teachers, (one a former postdoc at MIT, others from industry) and three sections of “multivar”.

The school has increased in size, but the expansion of math and science seem far out of proportion to the growth in student population. I can’t help but feel a twinge of school pride.

The end of e-mail

So this is how e-mail dies.

I’ve been using university-provided e-mail since 2002. Partly this was just because it was free and convenient, but there was also something more than that. I’ve consciously avoided forwarding my e-mail out of my academic account and into a commercial provider, like Yahoo or Hotmail or Gmail. I felt I could trust my university not to engage in undesirable practices like handing over the contents of my e-mail to government agencies (or marketing firms), blocking messages that are inconvenient for them, or otherwise compromising what is, fundamentally a private and highly personal part of my daily life. I trusted them in part because I am a client of the university, to the tune of several thousand dollars a year. When you use a free commercial e-mail account, the opposite is true: you are the product.

Universities have been increasingly unwilling to invest in their mail operations though, especially as budgets are squeezed. The sticking point has been handling of spam, which has become more and more difficult as the cost of sending a billion ads for 409 scams has fallen to just about zero.

The best spam handling systems work by observing e-mail patterns of millions of users, enough to catch, identify, and block each new spam wave as soon as it starts. One company, Postini, provides this filtering as a commercial service, and Harvard, rather than run (and constantly update) its own spam filters, has contracted with Postini to provide filtering. Last night they switched my account over to the new system, a Microsoft Exchange server with Postini for spam.

So why is this the end of e-mail? Because this means Postini reads all of my mail, every message sent to my primary account … and Postini is owned by Google. Now Google is dramatically better than most, thanks to their transparency report, in which they tell us that they received 5,950 requests by US government agencies for private user data in the past year, of which they complied with over 5,500. The fact that they actually review requests and occasionally deny them is remarkable. Still, it’s hard for us to know for sure how they strike their balance between their users’ privacy and their own legal safety.

There are other threats too. Google has recently seen a series of hacks by apparent agents of the Chinese Communist Party. They have a reputation for never deleting anything. They’ve also caught their own employees snooping on users.

Traditional e-mail mostly doesn’t have this kind of privacy problem. On the modern internet, e-mail is point-to-point, so there aren’t too many people who have a chance to inspect or record your messages … or at least it was. Now if you send an e-mail to me, Google will see it first. They’ll even get a chance to block it from ever hitting my inbox, if it gets classed as spam.

This isn’t really e-mail anymore. It’s GMail, or close enough. When everyone has GMail, GMail is more like the old-fashioned bulletin-board message systems, run by a single host where all the users have accounts. Like a Facebook message. (An oligarchy is hardly better, if none of the providers are accountable to their users.)

If Harvard is going to route all my mail through Google, then there’s hardly any reason not to forward my account to GMail entirely … and indeed precious little reason for Harvard to operate its own mail system at all.

If you’re sending me a private e-mail, I hope you’ll consider encrypting it to my public key. Otherwise, I can guarantee that Google is reading every word … and someone may be watching over their shoulder.

CoinStar Amazon $5 minimum

Short: If you don’t have $5 in coins, don’t use CoinStar for Amazon.

When buying stuff on, I usually take the opportunity to run my coin pile through a CoinStar machine first. CoinStar machines are simple coin-counters, and normally they take a cut, recently increased to 9.8%. However, if you convert your coins into an Amazon gift certificate, they charge no fee (or rather, Amazon pays the fee for you).

Unfortunately, as I discovered today, there’s a sleazy slimy catch. If your coins add up to less than $5, you will be prevented from making an Amazon gift certificate … and now the machine has your coins, and there’s no way to get them back. Instead you will be shunted back to the cash voucher pathway, with its 9.8% fee. There is no alternative.

You will lose up to $0.49, as I nearly did today.


I recently discovered that one of my friends, on a whim one weekend, threw together a website called Ex-rated. It’s a very simple system that allows you to post public ratings and reviews of former boyfriends and girlfriends (or indeed spouses), and also see if anyone has rated you, or someone you’re curious about.

It’s all tongue-in-cheek, of course. The whole thing is a deliberately terrible idea. Nobody wants to associate with someone who’s been known to post public critiques of people with whom they were formerly close.

The strangest thing, though, is that despite its humorous intent, the service seems to be getting some real use … mostly in Brazil. Well, maybe that’s the second-strangest thing. The strangest thing is that when they admitted that they were not going to quit their day jobs to work on this joke, somebody else decided to take it seriously and cloned the site.

The internet is a funny place.

Twisted Pair

I just had a cell phone call drop so hard that the audio cut out … and was replaced by a busy signal. The problem wasn’t on my end; I have perfect reception.

If you think about it, this is an absurd phenomenon. A phone call never becomes busy mid-call; it only makes sense before the call starts. Also, busy signals serve to indicate something (the far end can’t tell that you’ve called) which is never true today. And while we’re at it, the reason to use a busy signal is to transmit that information in-band, because there’s no metadata side-channel … and that’s also false on every cell phone.

The truth is, we’re carrying something like a century’s worth of accumulated hacks in our pockets, often for the sake of compatibility with systems that no longer exist. At best, it reduces call quality and forces us to repeat ourselves; sometimes it just falls over entirely.

Everyone can see it’s time to move to a phone system for the 21st century … except, of course, for the phone companies. A truly modern phone system, where voice is just a kind of low-latency data stream, would force the wireless carriers to adapt their business model. Better to capture the regulators, and stave off net neutrality for as long as possible to maintain the oligopoly.

Technologies often challenge existing business models. Eventually, technology always wins. One day, I think when people say “phone call” they’ll mean Opus and WebRTC, end-to-end. The present powers can certainly forestall the transition though … and so far they seem to be succeeding.


Wednesday-Saturday this week is a takehome final for Physiology. The funny thing about take-home finals is that they truly expand to fill all available time … which isn’t much, in a week with two rehearsals and two concerts, one displaced at the last minute by Harvard’s autoimmune overreaction to Occupy.

I need to get some sleep, or I’ll still be thinking about chemokines and fibrocytes at our concert tomorrow night.

In the Dock

I just attended a public interview by Larry Lessig of Jack Abramoff, under the banner of the Edmond J. Safra Center for Ethics, of which Lessig is the director. The interview was billed as the first of a (potential) series of public interviews of people whose unethical behavior has led to their imprisonment. (In this case the title is something of a misnomer; Abramoff was released almost a year ago.) My friend Chris was there too.

The exchange was distinctly more cordial than one might have expected between a famously ethical professor of law and an infamously corrupt convicted felon. I attribute this above all to Abramoff’s uniformly apologetic attitude; it’s hard to stay angry at someone who says, over and over, “I was a really bad person, and I did a lot of terrible things. Now I’m trying to make up for it.”.

Another reason why the two could hardly joust with any spirit is that both have recently published books that recommend detailed plans for eliminating pernicious corruption in Congress by imposing aggressive restrictions on donations to political campaigns. (Lessig claims to have the better proposal, and I think he’s right, but the difference is mostly in the details of how the limits are structured.)

Instead of a debate or deconstruction, what I heard was a conversation between two great orators. Lessig is famous for his public speaking abilities, but Abramoff was hardly worse. His combination of emotional intelligence and verbal skill is extraordinarily persuasive; it was immediately apparent why he had been so tremendously successful — and dangerous — as a lobbyist.

A few choice (approximate) quotations from the discussion:

Lessig: I’d like to thank Jack Abramoff for coming here tonight to help us better understand a system that practically none of us respect.

Lessig: Is it your sense that there’s a shift toward indirect payment from lobbyists, things that could contribute toward getting re-elected and away from things that could contribute toward, say, getting a boat?

Abramoff: Most people think of a Congressional Hearing as something like a court of law, with lawyers on both sides, some sort of high-minded debate about an issue. It’s not. It’s a kangaroo court, and if it goes badly enough for you, you can be locked up in the basement for Contempt of Congress. Even if it goes well, you’re going to spend at least a million dollars.

Abramoff: The most important influence to establish isn’t influence over the congressmen. It’s influence over the staffers. Congressmen are busy; it’s a staffers who make the real decisions about legislation.

Lessig: This reminds me of Jim Cooper’s quote that Congress has become “a farm league for K street”.

Abramoff: Early in my career I used to hire staffers immediately. I would say “when can you start?” and hire them to start the next day, or the next week. Later I learned to say “when are you thinking about leaving the hill?”. They would say “in two years”, and I would say “great, I’ll hire you in two years”. For the next two years, that staffer is my staffer. In fact he’s better than my staffers, because he’s right there in the congressman’s office.

Lessig: Your campaign finance plan would silence a whole lot of people, despite your supposed libertarian leanings.

Abramoff: You know how they say “a conservative is a liberal who’s been mugged”? Well, a libertarian is a conservative who’s been indicted.

VoiceLab Concert!

VoiceLab’s annual Winter Concert will be next Sunday, December 11th, at 3:30 PM (doors open at 3:15 PM) in the Gutman Conference Center (here’s a map), courtesy of Class Notes of the Harvard Graduate School of Education.

As per tradition, wintry snacks will be served, and admission is free of charge.

If you can’t make it on Sunday, then you’re still in luck. We will also be performing this Friday evening, December 9th, at 8 PM on the Harvard Medical School campus in Tosteson Medical Education Center room 209.