When there's only one

...there's only one choice
Everything here is my opinion. I do not speak for your employer.
February 2013
April 2013

2013-03-02 »

Hi, I'm in Paris this week!  Annoyed at the state of wifi sharing networks.  Not so much that they don't work - reputedly they work great, and France has at least two big ones with millions of hotspots, free.fr and sfr.fr - but that it's way too hard to just buy a 2-week subscription to one.

The way these things work is your ISP (in this case, free.fr and sfr.fr) include a second and/or third wifi SSID in all their home wifi routers, and anybody with a home Internet connection (or in some cases a cell phone plan) can access the connection of any other member, on a firewalled-off network.

This is great - if you live here and have a home Internet connection.  If you're visiting like me, it falls apart pretty fast.  As far as I can see, the only way to get a non-SIM-locked subscription (the SIM-locked ones auto log into WPA2-encrypted networks using EAP-SIM, which is neat, but requires a cell phone with an appropriate SIM... but my phone is CDMA, argh, and my laptop obviously has no SIM) is to get a home Internet plan, which is obviously not going to work for me.  To just go buy a SIM isn't so hard for sfr.fr, since they have retail stores (will experiment with this tomorrow, hoping they're open on Sundays), although you have to sign up for a subscription since their pay-as-you-go plans don't have wifi access. That means you have to remember to cancel the subscription within 30 days, sigh, and thus probably deal with the company's retention department, double sigh, in French, triple sigh.  free.fr is even worse, since they seem not to have any retail stores at all, so they have to mail you a SIM.  Again, not so bad if you live here, but sucks if you're just visiting.

The SIM locking, by the way, is probably all because otherwise it's too easy for one subscriber to give access passwords to all his friends, who then never need to buy their own connection.  That's a problem, all right, but surely there's some other way.  (The claim is that EAP-SIM is a way to auto-connect securely to their WPA2 networks without any need for a UI.  Which is true, but I'm pretty sure you don't need a hardware dongle for that.  It's all about the copy protection.)

Upon further investigation I found that sfr.fr is also a member of the FON network (fon.com), so basically any sfr.fr subscriber can connect to any FON-enabled wifi router, of which there are zillions in several European countries because several huge telcos (including BT) auto-opt-in all their customers (they do provide an opt-out but most people don't bother).  Since sfr.fr is a FON member, theoretically any global FON user can also then reciprocally use any of the sfr.fr hotspots - except, oops, while most FON-participating telcos opt their customers in automatically, sfr.fr requires you to manually opt in, and again, most people don't bother to change the default.  So the net effect is that sfr.fr subscribers who want to travel probably opt in at the last minute and use FON whenever they want, but most other people don't end up enabling it on their own account, so most sfr.fr hotspots are not FON enabled.

Which is too bad, because FON, at least, unlike sfr.fr and free.fr, at least claims to have (I haven't tried it yet) a useful web portal where non-members can just pay some money to get access for some time period.  Which is all I want.  "Oh, you're not a subscriber? Enter credit card number here."  Or as Fry would say, shut up and take my money.

I think I'll try chasing down some FON hotspots tomorrow.  There are actually still quite a lot in Paris, in the grand scheme of things, at least from how it looks on their map (http://corp.fon.com/maps).  I just don't happen to be sitting near one as I do these experiments.  In contrast, every time I look, I am always near a free.fr or sfr.fr shared hotspot.  This place is seriously saturated.

FON has virtually, but not completely, zero penetration in North America.  It's pretty clear that none of the North American telcos have signed on, although individuals can buy a FON router and try to participate, and it looks like some have.  It's so sparse that it's nearly useless though.

I don't think FON is a good enough reason for the average citizen to to out and buy a special router.  On the other hand, an extra special router with extra special features (whatever they may be) could be a good platform to bundle FON into.

2013-03-03 »

Wandering around the 3rd Arondissement in Paris (I haven't ventured further yet), there is a "Free(tm) Wifi" ("Free" is a company, it's not actually free) literally everywhere, usually with full signal bars.  I don't really trust signal bars after having read that people can configure routers to blast full Tx power but that doesn't help reception, so I actually tried joining the networks from my phone - they connect fine.  I don't have an account so I can't test the actual Internet (see previous post), but at least I can see the portal page.

There's also an "SFR Public Wifi" in almost, but not quite, all locations, but the signal strength is almost always less.

There has only been one place where I noticed FON availability, but its signal strength was so weak I couldn't connect to the network.

Although I haven't actually gotten to use the Internet at any of these nodes yet, the experiment is still interesting - as is wandering around with Erin, who doesn't care at all.  She wants a SIM card so it "just works." (Her phone is a proper GSM unlocked iPhone, unlike my useless-in-Europe CDMA Nexus S.)  Of course she does: why would you want something that doesn't just work?  (We'll try to get a SIM card later, but things like that don't happen in France on a Sunday.)

For me this is an experiment in whether public shared wifi could ever be a viable option on its own.  Free.fr is about as close as you can possibly get, and it really does saturate this place so entirely that I think it might be usable as an alternative to a cell phone plan, assuming you dealt with mobile IP roaming issues, bandwidth limits, etc, which ought to all be manageable, albeit difficult.  But it doesn't work if you leave the city, and it seems to require a city with the density of Paris to actually pull it off.  I don't know how popular SFR.fr is, but they're not available quite everywhere in town, so it would be endlessly frustrating to try to use it (eg. for maps) instead of a cell plan.  The wifi is purely a second-choice network that lets you offload the cell network a bit and sometimes avoid metering.  Regardless of whether you can subscribe to it without buying a cell phone plan (you can't), it wouldn't be a positive user experience.

Of course, if both providers would just auto-connect all their customers to FON, then the combined coverage would be astonishing.  That would reduce Free.fr's advantage over SFR.fr, though, so why would they do that?

All this makes the concept of public wifi look rather grim.  The business model isn't there (you make much more money selling cell phone plans) and the user satisfaction isn't there (nearly anyone would prefer a slow, always-on cell data plan over fast, flakey, often nonexistent wifi data).

Or maybe I'm being shortsighted and need to think bigger.  Maybe it's possible to super-ultra-saturate an area, for a reasonable price, and make a business of it.  I think a key component of such a plan would be the resolve to not become also a cell provider.  Cell providers have a perverse incentive to make sure wifi coverage is imperfect.

2013-03-07 »

There is one constant in computing: No matter how fast the CPU gets, the apps can never keep up with my typing.

I'm talking to you, New Compose Window.  And all of Android.

2013-03-15 »

I'm getting confused by all these timezones.  I vote we rotate Europe by 90 degrees.

2013-03-18 »

Dear Robot Vodka Salesperson,

That would have been a lot funnier if you had said 2038.

Albeit maybe only to me.

– Avery

2013-03-19 »

Disabling the CPU fan, then running a system stress test, to see if it increases the frequency of the kinds of crashes we already see sometimes: bad idea, or worst idea?

2013-03-20 »

"I sort of wish we told new engineers 'don't work on something that you don't think is a good idea'." - Phil 

2013-03-22 »

Got up from my desk to go to the washroom.  Turns out the area of the floor where I sit has been blocked off with yellow "Caution - Do Not Enter" tape.  A lesser employee might have worried at that point, but I realized: no problem!  I'm not even trying to enter!

2013-03-23 »

Imagine for a moment that Hypothetical Web-based Product X was canceled because even though it is "successful," the moderate success doesn't outweigh the opportunity cost imposed by devoting energy to its ongoing maintenance.  (It's a hypothetical product, so we can imagine that it's true.)

If Product X were a real-world product, or even a shrinkware software product, the effects of its cancelation would be contained.  Existing users could keep using it until they get tired  of not having any new bugfixes/features, and slowly leave at their own pace.  Like, say, Windows XP. They might be sad that it is no longer maintained, but they aren't violently angry, because there are no immediate consequences to them.

Web-based products aren't so lucky. But I got to thinking: why is that? What's the big difference?  Well, the big difference is that unlike shrinkwrap software, web software keeps on costing the service provider money to keep on hosting the service. That seems pretty much inevitable, barring near-impossible-sounding options like open sourcing it and the entire set of hypothetical infrastructure the Product X depends on.

On the flip side, however, computing power is getting vastly cheaper every year. Naively, the cost of running Product X should decrease every year, assuming you don't have annoying things like a growth in user population.  In fact, given Moore's Law, one might hope the cost would decrease exponentially year over year.  Interestingly, if that were true, you could actually account for the total costs of a web service (again, assuming users don't increase) up front, using the net present value of all future costs.

But that's not how it works, of course.  Why not?  My theory is: because the infrastructure APIs keep changing.  If you leave Product X alone, it will soon kill itself irreparably because the underlying infrastructure changes incompatibly.

There are different ways to deal with the stability of an underlying platform.  One way is the Microsoft approach of putting tons of effort into backward compatibility (such that many DOS programs still work in present-day Windows).  But that method is expensive and annoying.  On the other hand, I don't need Windows in order to run, say, Lotus 1-2-3 for DOS: I just need DosBox on my Mac.  If I need Windows apps, I can also run Windows XP just fine in a virtual machine that will keep working as long as I have an x86 processor - and by the time I don't, the emulators will probably be fast enough.

I wonder if, one day, we can create a world where web services infrastructure stays stable long enough for it to be shrunken down and virtualized, like DOS, so that it can keep working forever, and there is no economic reason to shut down a popular-but-not-popular-enough service.

2013-03-27 »

If you instead filed bugs about devices that didn't crash, you'd probably save time.

2013-03-28 »

Whacko problem of the day:

It appears that MoCA (network-over-TV-coax) hardware is so good that it can actually communicate wirelessly over short distances, accidentally, via crosstalk between unconnected cables.  In apartment buildings, it's possible to actually connect to your neighbour's network by accident, or bridge your two networks together.

100 MBits/sec, apparently.

Come on, physics.  Cut us some slack here!

2013-03-29 »

What, you mean "it's much less crazy now" isn't a compliment anymore?

2013-03-30 »

Friday's project:

The y axis has one pixel per device, sorted by a few interesting attributes.  The X axis has one line segment for each one-hour period over the last week. The darkness of each resulting dot/rectangle corresponds to how much activity was done by that device during that time.

Pretty self-explanatory, really.

2013-03-31 »

This has been a test of the emergency alerting system.  This is not a real emergency.  If it had been a real emergency, you would still have had no idea what to do about it.

February 2013
April 2013

I'm CEO at Tailscale, where we make network problems disappear.

Why would you follow me on twitter? Use RSS.

apenwarr on gmail.com