If this were a live FonBar, you would log in to the hotspot through the form above.

NYT Article Flubs FON Statistics

May 28, 2008

(An much-expanded version of my comment to WiFiNetNews blog by Glenn Fleishman:)

The statistics that Fon throws around are nothing if not misleading, and Fon will make no effort to correct journalists who get confused! Martin has used similar figures as from the NYT article, but attributed them to *different* elements, and I compare them below.

Glenn Fleishman writes that “Non-Foneros” pay for access to the Fon Network’s hotspots, but this is not correct. The Aliens, who do pay for Fon wifi, are registered, and Martin counts them in his figure of 830k Foneros. They’ve registered their email addresses, they’re Foneros. It’s almost as though there was a table of figures “Registered Foneros – Registered Hotspots – Active Hotspots” which got bumped aside one notch in the NYT article! Well-researched indeed!***

NYT uses the figure of 332k as the number of active hotspots, but Martin blogs that figure as being total hotspots ever registered. I also suspect that the figure also includes Linksys and La Fonera routers, which were shipped by Fon but still never registered to join the Fon Network.

Martin’s blog posits the number of “active” hotspots is really only 212k, but by downloading the database of hotspots “POI files” (for use offline to find Fon hotspots), the total number of hotspots only comes to 198,366!* This “number” also disagrees with the figure provided by FrancoFON’s surveys (see below), but neatly comes almost perfectly in alignment if it includes non-reciprocative (they can hop on Fon’s Network, but no arrangement is made for Foneros to hop on theirs) “partner” hotspots hosted by Neuf, Livedoor, BT and others.

Examining these POI files shows that they are filled with hundreds, probably thousands of “junk” entries with fake addresses and impossible coordinates (like the North Pole). Though the blame here lies with Foneros who lied about their real personal data, it is Fon’s responsibility to weed them out. Additionally, Foneros have been complaining that the POI files still contain hotspots which no longer appear on the Fon Map, due to extreme length of inactivity.

Going further, a cursory examination of Fon’s Map shows that a signifigant percentage of the hotspots featured there are offline (ringed with grey). Foneros have also reported that their “offline” icons remain on the map for months after they have turned off their Fon hotspots. Clearly, Fon cannot be trusted to provide accurate numbers of online/offline hotspots, but FrancoFON polls Fon’s database directly and has determined that there are really only 96,170** online hotspots worldwide! That makes 37% of all Fon hotspots offline (Fon POI) or 45% offline (FrancoFON).

Last, but not least, just because a hotspot is “online” does not mean that it is accessible to the public. Most Fon hotspots are in people’s homes where the range may barely reach the street, or not. Even then, the street is an awkward place to sit and browse the web. Many Foneros reported that they could not connect to any of their neighbor’s hotspots, and could rarely even detect them.

Sadly, Fon does not motivate Foneros to place their hotspots in useful locations. The kickbacks for Linus/Bill Foneros (free worldwide roaming, 1/3 of daypass sales) just don’t happen often enough. Foneros who pull their hotspots down do not feel that they have lost something. In fact, they may feel relieved, and more secure from evildoers. In the English Fon Forum, we Foneros are currently discussing ways to display a public success-rating for Fon hotspots, but we doubt Fon will add such a feature when it will surely reveal bad news.

Many remaining Foneros are frustrated that Martin and Fon continue to pursue “partnerships” with major telcos. We don’t share Fon’s enthusiasm because these deals do not include anything for us. Martin grants free access to our hotspots to these “partners” as incentive, and this whittles away what remains of our profit opportunities.

Fon is charging ahead into new territory, (some of which isn’t even wifi related – but still sustained by Fon funds) and left unfinished business with the Fon Community. Fon has abandoned the repair of long-existing shortcomings and flaws in our firmware, security, hotspot management interfaces, customization options and profit-sharing terms. This makes us feel that we are taken for granted, and that Fon is more interested in selling the company than in keeping their promise to us to lead the “biggest” and best wifi-sharing community in the world.

* as of May 19

** as of May 21

*** ADDITION: In Martin’s latest blog entry today, he explicitly states that “People who are not Foneros (we call them Aliens) will be able to access Comstar-FON’s network using pre-paid cards, SMS or the usual options available on our captive portal.” Yet, in Martin’s recent blog about Fon’s statistics, he CLEARLY counts all email-registered Aliens as Foneros to reach the sum of 830,000. This includes all of those unverified crap email addresses which the leeches made up to get free 15 minute trial connections each day. Doesn’t it occur to him that, if the only way to become a Fonero is to register a Fon hotspot, then there would really be fewer than 332,000 Foneros (his figure for all Fon hotspots ever registered)? Some Foneros have more than one router, so the number of Foneros would logically always be less than the number of all registered Fon hotspots. 😉


Fon Math 2008 (repost with more info)

May 21, 2008

It’s been over two years since Fon’s Official Launch on Feb 6th, 2006. The two-year anniversary passed without any news or fanfare, besides my own blog post. Today Martin Varsavsky has released some general figures about Fon’s current size, income and expenses. I’m assuming these are worldwide aggregate figures, with no omissions:

  • Total number of registered Foneros: 830,000
  • Total number of registered Fon hotspots: 332,000
  • Number of recently active Fon hotspots: 212,000
  • Fon monthly revenues: €100,000
  • Fon monthly cash burn (forecast): €350,000
  • Fon employees, worldwide, 2 years from launch: 61

Martin hopes to count 300,000 active hotspots by the end of 2008, though He does not say if this includes “partner” network hotspots. He states that cash burn was still €450,000 last month, and expects profitability by the end of 4th quarter, 2009.

It is interesting how Martin describes 61 employees as a good thing – his spin is that “so few” people have managed “so much” progress. However, this just tends to shed some light on Fon’s notably poor customer service. Fon actually had nearly 100 employees last year, and I’ve just learned that Fon is closing their Swedish and Korean offices, reducing staff in France, and firing one of the two USA employees, for a loss of 14 more employees.

The two USA employees are Joanna Reeves, and English Forum Moderator/Official Fon Blogger Steve Ross. This last item is a great concern to us who support the Fonero community through the forums, for Steve has worked hard and been very successful in restoring civility there after the damage caused by that Moderfon person. Gutting the USA staff like this is a surprising and devastating decision, especially since two years ago, the USA became, or almost became, the home of the largest group of Foneros.

Some more interesting facts from Cincodias:

  • Investor financing in 2006: €18M
  • Investor financing in 2007: €10M
  • Investor financing in 2008: €6M

It’s surprising how Martin Varsavsky promises that Fon will continue to grow and succeed, when the last two years, especially the last several months, have seen so many Fon leaders quit by their own decision.

Reprising my blog adventure on Jan 21, 2007, I’ve downloaded the POI files (.CSV format) from maps.fon.com, to compare with today’s new statistics and see how up-to-date they are. The 90 POI files have grown to 172, representing single regions as large as the USA and the Russian Federation, and as small as Ascension Island and Liechtensten! Some places i’ve never heard of, like “Iles” and “Burkina Faso”. “Serbia and Montenegro” is in the list twice; one of those entries actually seems to download the POI of Yugoslavia.

Most of the POI files are 1k or 2k, and opening them up shows they’re mainly just the results of punching in random words and letters when registering routers. This bogus data really should have been edited out years ago, and the POI dropdown list could use some quality control!

It took me a while, but I appended all of the files together as one, and opened it for a record count:

  • Total number of POI indexed hotspots: 198,366 !

Keep in mind that the POI lists still contain many hotspots which no longer appear as icons on the map, due to long periods of inactivity. The map itself contains numerous “inactive spot” icons, which includes hotspots that have been dead for months or years… So even with hundreds or thousands of junk records in the POI, why does it still fall so far short of Martin Varsavsky’s statement of 212,000 active hotspots?!?!?! With no publicly-accessible data to back up Martin’s claims, these numbers are hard to prove credible.


Top 20 Fonero Countries:

  1. 36,366 FR – France
  2. 33,421 JP – Japan
  3. 26,082 DE – Germany
  4. 15,512 ES – Spain
  5. 14,205 US – United States
  6. 12,306 IT – Italy
  7. 11,060 TW – Taiwan
  8. 9,237 SE – Sweden
  9. 6,296 KR – South Korea
  10. 6,494 NL – Netherlands
  11. 4,011 UK – United Kingdom
  12. 3,369 HK – Hong Kong
  13. 2,811 FI – Finland
  14. 2,438 PT – Portugal
  15. 2,496 CN – China
  16. 2,017 AT – Austria
  17. 1,814 DK – Denmark
  18. 1,531 BE – Belgium
  19. 1,199 HU – Hungary
  20. 980 CA – Canada
  • POI/Countries with 1000 or more records: 19
  • 101-999 records: 14
  • 51-100 records: 9
  • 5-50 records: 42
  • 1-4 records:91

It’s perplexing why Fon would reduce staff in France, when this is the largest group of Foneros, why Fon would close Sweden and Korea when they are both in the top 10, and why Fon Russia would get so much recent news when they have only 23 hotspots!

Steve Ross has written me to let me know that his Fon schedule remains full at this time, and he has received no indication that he may be let go. The La Fonera 2 Beta Test has not been cancelled, though Steve says the hardware apparently did not ship as expected yet. Fon has experienced a number of unspecified changes, which Steve feels has addressed some of the most-complained about areas.


I’ve been using a web-spider program to test the age of Fon’s POI files. Though Fon Maps makes downloading the POI a 3-stage process “select, create, download”, they are not really generated on demand. I had expected to find that they were several months out of date, and thus, my analysis above would be simply irrelevant. Alternately, I thought I might find that the files always test to be a certain age, no matter what time they are checked. This would be dependent on Fon’s web server, and could suggest that an age difference was merely due to a difference between the clock on my PC and theirs. However, at the time of this edit, Fon’s POI files appear to be 2.5 days old and growing! Since this is too much time to be a clock offset, and too little to explain the differences with Martin’s claims, I am reasonably confident that this is the file’s true age.

Another excellent source of Fon statistics is Francofon’s Fon Maps. I am told that they update their figures every night, and use data from Fon’s own POIs (or perhaps direct database access?). However, they come up with considerably different figures than I got directly from Fon Maps! FrancoFon lists 144 countries, instead of 172. They count 206,886 registered Fon hotspots, not 332,000 or 198,366. FrancoFon has determined that 95,324 Fon hotspots are active, (but who knows if they are accessible?) not 212,000. When the statistics include Neuf and Livedoor hotspots, who are “partner” wifi networks, (though not fully reciprocal to the Fon Network’s members) the numbers come closer to Martin’s claims.

This suggests that Martin is including these, and probably British Telcom and other “partner” networks, in his statement. This practice certainly makes the Fon Network look bigger and more active, but is it reasonable and honest?

ADDITION 3: It’s May 24, and Fon’s POI files now appear to be 5.5 days old, so it looks like they were indeed built late Sunday/early Monday. Now we’ll be keeping an eye out to see if they build weekly, monthly or whenever.

ADDITION 4: It’s May 26, and Fon’s POI files now appear to be 7.5 hours old. Therefore, it would appear that they are currently being generated once a week, Mondays at 1:30 AM CST (Sundays at 20:30 GMT). There are still 172 POI files/countries to download, and “Serbia and Montenegro” is still listed twice.

ADDITION 5: May 27. Concatenating all of the new POI files and counting the records/hotspots results in a total of 199,614 for an increase of 1,248 hotspots in one week. Interestingly, the overall filesize is 87,956 bytes smaller than last week. However, as of this addition, FrancoFON finds 96,170 active hotspots, an increase of only 846 over last week… Have 402 new hotspots already been binned?

Why have so many Foneros abandoned Fon?

April 4, 2008

Fon President Martin Varsavsky has posted an adorable leetle survey on his blog today. Either he has no idea why Fon is failing under his leadership, or else he knows, and won’t list those reasons because he has no intention of fixing them.

The reason most Foneros have quit Fon is due to anemic equipment and firmware imposed upon them, and the culture of dishonesty in Fon’s press releases and business practices.

After promising to give us firmware which supported dual-SSIDs, Fon switches the bait and presents us with their proprietary, locked-down 1-port router with this feature. No dual-SSID for us Linksys and Buffalo Foneros. It’s just as well, because it turns out that many wifi adapters can’t cope with the little transmission trick that produces two SSIDs.

People with pre-existing home networks discover that they can’t access their LAN resources, even when using the private WLAN. There is no “bridge to WAN” feature. This device *looks* like an AP, but is instead a NAT router. This is one of the main reasons people abandoned Fon. They didn’t want to *start* a network. They needed to *expand* one (and on a budget).

This little overheating brick had WDS meshing built-in at first, but this was undocumented. Hackers learned to use it to aquire an Internet connection without paying or logging in. Fon quickly took WDS out, and has still never admitted it existed. Pres. V pontificates in his blog that the range-extending Fontenna (he sells) is superior to connection-relaying meshing, despite the poor performance of said Fontenna. He should have instead sold us a kit to mount the router outside, with an embedded booster antenna and PoE adapter.

Nearly two years later, La Fonera still doesn’t support MAC cloning, which is such a trivial feature to add. It is necessary for modems/ISPs which lock your service to your WAN MAC. This is another big reason people abandoned Fon- they never got it connected to the Internet. This feature wouldn’t even threaten the sales of additional Fon hardware. :(

Instead of improving the La Fonera firmware (except to rush out patches to keep people from aquiring better access to their device and developing new features), Fon spends R&D on further routers: Want one precious LAN jack? Buy the new router, at twice the price of the old one! This is surely why they don’t give us WAN bridge in the original La Fonera for free.

What would Foneros really prefer that Fon focus their attention on? Bringing the feature set of the router at least up to the point of every other cheap router on the market, nurturing and empowering the creative community that has built up around Fon, and showing some real progress for a change. Instead, here is “La Fonera Orwellian Name”, for $100, which lets you download free bittorrents of Fearless Leader’s video clips. Ugh.

“Buy thees Skype phone and make calls for free at any Fon hotspot in the world!” they said. Well, sure- if you had the encryption key for all of those Fonero’s private networks. The darn thing wasn’t able to log in through Fon’s public hotspot, until many months later, when a firmware patch was provided. Calls were then free if they were Skype-to-Skype, or you were spending the included “free” 20 trial Skypeout minutes. Skype pulled the ads down. Ugh.

“We split the profits 50-50!”. An outright lie. First, Fon takes unspecified “fees and taxes” out, then splits what is left. Fon refuses to itemize this amount, which varies from country to country and depends on the ISP, so there is no way to tell if they are paying you fairly. Fon only pays Bills if their hotspot is the Point Of Sale for day passes, not for bandwidth, length of wifi sessions, or number of customers. If paid-up customers wander over to his hotspot, Bills get nothing for the service he provides.

Fon’s price per day is quite reasonable when compared to other for-pay mobile Internet services. However, wifi is free in virtually every coffee shop in the USA, many restaurants and libraries, and provided by many municipalities throughout the city centers. Fon won’t budge on the price, or add something to make their service more desireable than free wifi, like VPN encryption.

Fon’s system mimics other “instant hotspot in-a-box” offerings, but these competitors are offering more flexible terms in setting prices and managing equipment. The competitors let you have control of your Internet connection, your router, and the appearance of your hotspot to the public. Fon pretends that they do too, but in reality, you have almost no control over what they clearly consider to be *their* router, and *their* hotspot.

It’s been obvious in recent months that Fon is fading away. Varsavsky spends his time supporting side projects, which have nothing to do with wifi (Mexican Wave, Fon URL Sortener, and several ways to abuse Gmail), and writing bizzare articles in his Fon Blog. Varsavsky recently dumped much of his Fon stock.

Fortunately, there is a thriving community of hackers who still develop improvements for La Fonera wifi routers. If a Fonero is willing to void his router’s warranty, he can have his MAC cloning, WAN bridge and much more. While they can do nothing about Fon’s awful profit-sharing, the routers themselves can even be flashed with entirely different firmware, and be used with other wifi networks, or even liberated entirely, including features usually found only in very expensive equipment.

There are so many other points, I could write volumes. Please visit Varsavsky’s blog, and instead of taking his survey, leave him comments which surely will fall outside his carefully selected choices.

UPDATE: Y’all will find this very interesting. Martin approved another round of comments to that post in his blog. While he approved a comment I made under a fake name, he did not approve a more coherent comment I made, as myself, discussing the exact same points, somewhat earlier that day. This is not proper management of his blog, this is censorship of those whom he dislikes. What a skunk!

La Fontenna Review

June 29, 2007

I’ve been waiting for the opportunity to blog about La Fontenna since Martin’s video blog where he demonstrates it in conjuction with a La Fonera and an 802.11n router he has in his office.

After the excitement left me, I realized that Martin was not actually performing a real test of his antenna, he was merely pointing out the window and making speculations. He fails to mention that any laptop hundreds of feet, to 1.5 miles away, would need to be along the beam of his fontenna, at the same elevation, and need a similar antenna to transmit back. Also, the buildings in his area would probably cause a problem due to the Fresnel Effect.

Here are the range of hotspots which Martin Varsavsky describes in his video blog. Martin also doesn’t make it clear that La Fontenna is highly directional. If one did not know this, they would think he was describing gigantic *circular* hotspots!

I recieved my promo code for a $2 (+ $5 shipping!) La Fontenna in an email from Fon on June 12. I thought about it for a couple of days, and then ordered it the evening of Thursday, June 14. My La Fontenna arrived by FedEx on the afternoon of June 21, exactly a week later! Not bad at all. It appeared to be shipped from the same outsourcer in Libertyville, IL who shipped my La Fonera.

I had been concerned that my Fontenna might be lost, since the invoice emailed to me showed the wrong zipcode. Fon’s Shop displayed my name and address when I ordered, and asked that I verify or update it, and I did. These corrections were NOT carried forward to the order I was actually making! Fon does not include the shipment tracking number with the invoice, so I opened a ticket on Fon’s HELP page for customer support.

Suprisingly (heh), I did not recieve a response within 24 hours, as Fon promises. Five days later, on June 20, a Fon representative responded with my tracking number and advised me that *I* needed to contact FedEx to resolve the problem. The tracking number was appreciated, and I could see that my package was on it’s way. But when I called FedEx, they told me that it is the *sender* who is the only one who they could accept an address correction from. Fon needs to update their procedures, and contact Brightstar US, Inc. directly, and then send the customer a resolution notice!

About the Tests

For each of 6 tests, I kept my laptop in exactly the same positon on my front porch, in exactly the same orientation. I was careful to avoid coming too close to it, or to stand between it and the La Fonera. I chose to move the Fon hotspot instead of my laptop because the object was to see how well my personal experience benefeited. This is where my laptop would usually be. My laptop runs Windows XP SP2, and my wifi card is a Proxim Orinoco Gold ABG PCMCIA card, which has a reputation for providing particuarly accurate signal strength measurements.

Admittedly, this was not a truely scientific test. I didn’t try a multitude of positions or use professional testing instruments. I attempted to make it as “real life” as possible. I do think my tests were very reasonable, given the fact that there were probably fewer obstacles for the wifi signal than average at my house.

When attaching the reflector and La Fontenna, I made certain they were pointed directly toward my laptop. La Fonera’s stock dipole antenna was carefully positioned fully-extended and upright. The stock antenna creates a torus (doughnut) shaped pattern, with the antenna in the center. The reflector focuses half of this torus in one direction, producing a sort of teardrop shape, and the Fontenna is supposed to create a large circular lobe with the antenna at one edge, and a much smaller lobe produced just behind it.

La Fontenna Technical Specifications

From Fon’s online shop:
Frequency Range: 2400 MHz 2500 MHz
Antenna gain-without cable: 7 dBi
Antenna gain-with cable: 6.5 dBi
VSWR 2.0: 1 Max.
Polarization: Linear, vertical
Impedance: 50 O
Temperature: -10 C to +55 C
Connector: R/P SMA PLUG
Cable: ULA168,L = 3M

Indoor Test

For this test, my La Fonera was inside the house, near the opposite corner from the porch where my laptop was placed. It hung 6 feet (just under 2 meters) above the floor. The house is wood-frame, and the only obstacle of any note is the stove in the kitchen, which may cause a slight Fresnel Effect on the wifi signal. The signal passed through my bedroom door, the central wall, and the kitchen window behind my seat on the porch.

In the NetStumbler screen shots below, the “A” section is the signal using La Fonera’s stock 3-4db antenna, “B” is the stock antenna enchanced with a single “Windsurfer” reflector, like I have blogged about before, and “C” is using La Fontenna.

The signal seemed strong at first, then dropped 8db for some reason. This also occurred with my neighbor’s wifi signal on the same channel at the same time. I was unable to convince La Fonera to change channels (my Linksys was unaffected on ch 6), and subsequent tests showed even poorer reception. My guess is that something nearby had cooincidently just begun to generate interference, and continued until I was out of time to perform new tests.

The result is that indoors, La Fontenna performs somewhat better than the stock dipole antenna, but only slightly better than the reflector, and seems to suffer more “dropouts”.

Outdoor Test

There is an empty lot across from my house, and since I use an electric mower and trimmer, I chose to run an extension cord across the street and all the way to the rear of the property. This was about 120 feet (36 meters). The La Fonera and La Fontenna were suspended 8 feet (2.5 meters) in the air, in direct line-of-sight of my laptop.

I performed the same three tests. The result is that outdoors, La Fontenna performs noticably better than the stock dipole antenna, but worse than the reflector, and also suffered more dropouts.


I was sincerely disappointed in the results I got. Either I have a defective Fontenna, or there is considerably more than .5db of loss in that 3 meter antenna lead!

Perhaps La Fontenna was worth $7 to me, since I was looking forward to performing these geeky tests, and blogging about it for other’s benefeit. But it performed only marginally better than the Windsurfer reflector I built, which cost me nothing at all. I honestly would not recommend paying $19.95 (plus $5 shipping) for it, but I might pay an extra $2 for it with the purchase of a new La Fonera, as offered in Fon’s Shop, in case it solves some kind of installation problem.

I urge Fon to retire this flaccid device, and instead offer a $19.95 kit comprising a waterproof case to mount the La Fonera into, with an embedded flat-panel antenna on a *short* cable, and a built-in Power over Ethernet (PoE) adapter, so that La Fonera’s 5v DC and network connection can be sent together down a long Ethernet cable. We can buy 100 foot Ethernet cables here for the same price as La Fontenna, which is both cheaper and more flexible than 100 feet of antenna cable. With such a kit, we would not need to worry about running AC power to within 3m of the access point, and where to stash the AC/DC adapter.

If anyone would like to discuss my tests, or even suggest some additional ones, please comment below! ;)

Addition: Check out Marshall’s blog for a La Fontenna unboxing!

Fon Boards Welcomes New Moderator

May 23, 2007

Today we were delighted to learn that Steve Ross, Fon’s English Blogger will now become Moderator of Fon’s English-language discussion board. Ross is an accomplished blogger, Internet entrepreneur, a native speaker of English, and has been a very helpful and courtious presence in the board using the screen name ”Ross“.

We will now find Ross also using the “moderator” screen name, and overseen by US/UK staff; Ms. Dale Taormino (US Business Development and Marketing) and Joanna Reeves (CEO of Fon USA). Dale and Joanna will leave the daily moderation up to Ross, but may step up to the podium from time to time with official guidance.

The moderator known as “Moderfon“ is looking forward to having more time to spend with his family.

Ross will also continue to serve as Fon Blogger, and will be reevaluating the discussion board as a resource to Foneros. We recently observed the posting of written “Board Rules“, and Ross will be surveying the posters for their input on what rules are necessary, which may perhaps result in a Community-backed revision.

Folks on the grapevine may allready be aware of a petition to compose and submit a “Fonero Bill of Rights”, in response to poor wording and vague definitions in those “Board Rules”. This “Bill” would clearly list activities which are not banned, even if the “Rules” were otherwise unclear. Ross is aware of this movement, and has expressed interest in adopting language of that type for additional moderation policy.

Episode XII: Fon Censors Community Protest of Censorship

April 23, 2007

I bet the board meddlerator thought this matter had been sucessfully choked off when he deleted the thread. Well, this is the Thread That Shall Not Die. As you can see from the saved copy, there was no discussion of forbidden topics, just an accelerating protest against censorship, and confrontation of the moderator’s terrible attitude and lack of respect for the community. This is yet another example of how Fon scorns the community of bright and talented wifi-sharers, who will not tow the company line.

Click here to read the deleted thread.

This snapshot was taken just before the final post made, which was just a short compliment to me for having spoken up. Anyone who would like to continue this thread is welcome to use this blog’s comments to do so. :)

ADDITION: Today, megga-blogger Om Malik reported on some kind of overhyped Fon E-partnership that everyone is parroting this week. Awk! At the end of his post he references the long-unfixed bandwidth throttling control bug for the La Fonera APs. His link to one of the board threads on the topic is dead. Fon censored it!

Addition: the board thread which Om Malik linked to in his blog has been restored!