DG Lewis's Techdirt Profile

DG Lewis

About DG Lewis

DG Lewis's Comments comment rss

  • Feb 10, 2011 @ 06:00am

    Specious argument

    "It wasn't that long ago that no one had mobile phones and emergency responding still seemed to work."

    As others have pointed out, when no one had mobile phones there were other ways to get in touch with emergency responders - every store, restaurant, gas station, and bar had a pay phone and they were on half the street corners in any decent-sized town, and calls to 911 (or the operator, before 911) were free.

    How many operational payphones have you seen in the past week? You could probably count them on one hand - if you've seen any.

  • Jan 20, 2010 @ 01:45pm

    "When it comes to dealing with a "hard looking fbi agent", you still have to follow procedure."

    They did follow procedure. The FBI agent had an "exigent circumstances letter," which was treated as a legal authorization equivalent to a subpoena. Whether they were called that or not, the equivalent mechanism has been in place for years. A telco security manager (pre-9/11) described its use to me as, "When a kid's been kidnapped and the FBI needs a wiretap to find him, they'll have an emergency authorization, and we always comply with those."

    Sure, the telco could "push back when handed a bogus demand to hand over records that did not match the official process and violated the law." But that wasn't what they got - they got an official demand that matched the official process and was in compliance with the law. The FBI may have been misusing the process, may have ignored the follow-on part of the process, and the law may have been unconstitutional -- but telco security compliance managers aren't constitutional law professors.

  • Jun 23, 2009 @ 06:37am

    Coke Machine

    AC@#16: Actually, it was CMU (er, sorry, make that "Carnegie Mellon") that had the first internet-connected Coke machine. See here.

    (E '87)

  • Apr 13, 2009 @ 12:17pm

    WH writes..."If your neighborhood was a good business model, your phone company would install a remote CO to run DSL. Ia m suspecting that it isn't a good business model for them. have you considered forming a small company, getting service from an internet connectivity supplier and reselling it? Apparently there is a need in your end of the woods.

    Oh yeah, no 3G? no Wimax? No Dish?"

    My neighborhood is about 170 single-family homes ranging in price from $575k to $950k. I would guess that high-speed internet penetration is north of 95%. The houses that don't have cable TV have DBS; many of those have two dishes. It is a lovely place for an ISP to do business. "Business model" has nothing to do with it. It's all about politics -- NJ got kicked to the bottom of Verizon's priority list when they were fighting with the state BPU over a statewide video franchise, and our neighborhood hasn't made its way back up the list yet.

    FiOS will be there, eventually - but Verizon isn't going to invest in deploying RDSLAMs that will become irrelevant in two or three years when they get around to deploying FiOS.

    3G as a viable alternative? Don't make me laugh - I can get about 200 kb/s on my AT&T 3G wireless card. No WiMAX providers. Dish, as I said, is present for video, but satellite internet can not hold a candle to cable, FTTX, or DSL as far as capacity or latency is concerned.

    And please put away the strawman about whether or not it's "Cablevision's fault" - I never said it was; I was saying that your claim that "there are at least two alternative providers in every market" is, when examined at anything lower than the SMSA level, utter bullshit.

  • Apr 10, 2009 @ 10:53am

    "there are at least 2 alternative providers in every market."

    Define "Market". If by "Market" you mean "SMSA", well, hell, sure. If by "Market" you mean "at any given residential address", absolutely not.

    I live in a three-year-old upscale housing development in suburban NJ. My choice for internet is Cablevision or dialup. We're too far from the CO for DSL; Verizon isn't installing remote DSLAMs, and FiOS hasn't been built to our development yet. (When the utilities were being installed in the development, Verizon was in a pissing match with the state of NJ over statewide video franchises, so decided to stop building out FiOS in the state. I therefore have a lovely 4-pair cable coming to my house.)

  • Apr 09, 2009 @ 02:29pm

    Last mile != last 50 feet

    The "last mile" that's the problem isn't the home to the curb; the problem is what telcos call the "feeder plant": most generically, from a "hub" point (which may be a building or may be a pedestal) to the customer's property. The "Homes with Tails" article you cite calls for a shared-ownership model for the feeder plant. Presumably, you could hire Verizon's subcontractors to build it for a cost approaching the $750 per household passed that Verizon is at. (Plus someone's got to perform ongoing maintenance - it's not a lot, but you've at least got to have someone with the appropriate expertise on call to handle any problems that arise, like backhoe fade.)

    The problem is, those subs are going to expect to get paid $750 per household passed regardless of how many people actually buy into the ownership model. Figure 30% of households don't even buy high-speed internet; figure half the remainder are perfectly happy with their current provider. So to cover the cost, the remaining households are paying about $2300 to buy in. If your "over the top ISPs" charge $10/month instead of the telco/cableco $30/month, a homeowner could save $20/month. But I suspect few homeowners will want to pay $2300 up front to save $20 a month on broadband. Thus driving down your take rate, and driving up the cost of the buy in, further driving down the take rate...

    What this leads to is homeowner associations with mandatory buy-ins, as part of a maintenance or community association fee. So you have to pay $15/month to the HA, even if you want FiOS or cable HSI. Which doesn't seem to me like a terrific solution to the problem.

  • Feb 26, 2009 @ 06:09am


    Why would we want to give Hulu some spiral-horned antelopes of the African bush?

  • Jan 23, 2009 @ 08:20am

    Two words:

    Melancholy Elephants.

  • Nov 11, 2008 @ 07:02am

    Railroads and Telecom

    Ultra @#4: "i've read about this railroad analogy before, i think in the context of a book or article review. if there's more to read on this, could you link me to it?"

    Andy Odlyzko at the University of Minnesota (formerly at AT&T Labs) has written about this fairly extensively. Check out his publication list; in particular, "The evolution of price discrimination in transportation and its implications for the Internet" goes into this for a number of transportation systems (canals, turnpikes, railroads), and the "more extensive manuscript" even more so.

    Odlyzko rulez.

  • Jul 25, 2008 @ 09:24am

    Innovation since 1996

    "The only Innovation since 1996 was the Integrated T1. Big deal. All other innovation is the result of the Carterphone decision..."

    Really. Try hooking up your FTTH ONT to a POTS loop and see how it works. Or your VDSL modem. I guess the 80 wavelengths of 40 Gb/s capacity carrying the internet traffic across the country are CPE as well. Plus the gigabit routers that get the bits from Techdirt's hosting site to your computer - all because of Carterphone.

    Oh, and Local Loop Unbundling only "stunted" DSL because the telcos decided that if the rules weren't slanted their way, they weren't going to play. They could have rolled out DSL service at any time -- I remember reading requirements for ADSL in the late 1980s -- but chose not to because they didn't like the rules, so DSL service stayed on the sidelines until they managed to get the UNE-L rules sufficiently gutted by the courts.

  • Dec 04, 2007 @ 07:43am

    Re: Re:

    FiOS can be installed in multi-family dwellings. In townhouses and the like, VZ will pull fiber to the same point at which it currently terminates the copper (an NT on the outside of the unit) and use existing inside wiring (coax and twisted pair) for video, data (using MOCA) and voice. In high-rise apartments and the like, where the copper typically terminates in a basement wiring closet and the inside wiring isn't suitable for this topology, VZ can either cut a deal with the building owner to pull fiber up the riser to a wiring closet on each floor and install the ONT there, or can drop a multi-unit ONT in the basement wiring closet and use VDSL2 and coax to get to the individual units. Just because it's not being widely offered today doesn't mean it "cannot" be offered.

    In many new housing developments, VZ is pulling fiber before the houses go up. In others, it's going back and trenching. Most new housing developments have underground utilities, and once the municipality allows the utility a ROW, it can't prevent them from going back and pulling new plant.

    Private subdivisions (i.e., "gated communities") are a different story; they can limit utility access, sign exclusive deals with a particular provider, etc. And those kinds of deals typically lead to homeowner dissatisfaction (and a decline in housing values) which either leads to changes in the terms of the deal or breaks the deal. Plus, VZ will compete for those exclusive deals as much as will the cablecos and any other provider.

    Most hotels, actually, have satellite service, not cable.

    In the long run, fixed-line communications (telephone, internet, video) will basically be a duopoly, with telcos and cablecos splitting a market roughly down the middle (40%-60% each). Satellite will keep 10%-20% of video, and there will be 10%-20% "cherrypickers" who will pick and choose individual services from the various providers (including second-tier providers like Vonage); but in a given market, FiOS triple-play market share will approach 40%-50% in the long run (10 years).

    On the other hand, VZ is only planning to pass about 47% of their households with FiOS, and VZ only has about 40M households in their service territory (out of about 110M nationwide) -- so even at 50% share of FiOS HHP in-region, that's less than 10% of the national video market.

  • Oct 22, 2007 @ 07:22am

    "AT&T tried, and failed, to compete with Vonage in the marketplace."

    AT&T's lost maybe 800k subscribers to Vonage (compared to about 2.8M lost to cable), and still has 59.4M lines in service. Vonage's growth is slowing down and will probably plateau at around 3M, about 5% of AT&T's total line count.

    And AT&T's stodgy old-fashioned POTS is generating about 40% more revenue per line than Vonage.

    So how, exactly, has AT&T "failed to compete with Vonage in the marketplace?"

  • Oct 03, 2007 @ 08:28am

    2.5 GHz spectrum for education

    Back when the 2.5 GHz spectrum licenses were given to educational institutions, the expected application was "wireless cable" - i.e., enabling the institution to transmit educational video programming between locations, such as from the high school to the middle schools, or from the main campus of the community college to outlying campuses. The names given to the service - EBS, or Educational Broadcast Service, formerly known as ITFS, Instructional Television Fixed Service - telegraph the expected use of the service.

    Unfortunately, many educational institutions found that the cost of building a network was excessive, even with free spectrum licenses. Thus, a lot of the spectrum sat vacant.

    Realizing that vacant spectrum did no one any good, the FCC allowed license holders to lease a significant amount of their spectrum for commercial use.

    I doubt that it was "a complete mystery" to many institutions that they held spectrum licenses - they had to file for the licenses in the first place; they weren't just "handed out" by the FCC. I suppose in the ten years or so since the licenses were granted an institution could have lost track of them, just like they could lose track of any other assets - but the notion that the FCC gave out licenses that institutions didn't even know about isn't correct.

    Values of assets change all the time, and organizations will behave differently based on the current perceived value of the asset - not what it was worth five years ago, or what they think it might be worth five years from now. Let's say that a farmer bequeaths 100 acres of land to the local school district, with the condition that it be used to build a new high school within 10 years; if no school is built, the land will be given to the county for open space. But the old high school has plenty of room, it's perfectly adequate for the needs of the community, and the school district can't justify the need for a new high school sufficiently to get a bond issue passed. The land is basically worthless to the school district - all it can be used for is a high school or open space, and the school district doesn't need a high school - so it's willing to let the land revert to the county.

    Suddenly, Sprint builds a tremendous customer support center in the town, bringing a huge influx of families, so the high school isn't big enough any more - and Sprint offers to contribute half the cost of building a new high school. All of a sudden, that land that the school district valued at absolutely nothing is now worth something, and the school district won't want to let the county have it any more. It's the same land, with the same restrictions, but the value to the school district is completely different.

    Like others, I'm really not sure what your point is. Is it that these evil greedy educational institutions are throwing up roadblocks to Sprint's noble effort to build a new wireless network? Is it that the foolish FCC knows so little about what technology will be like 10 years in the future that it doesn't allocate spectrum appropriately?

    You could have made a couple of very interesting points. One is a point of regulatory philosophy - that application-specific spectrum licensing, while perhaps valid in the early days of the FCC when there was a tight coupling between the use of spectrum (e.g., radio/TV broadcast, point-to-point fixed communications, land mobile communications) and the frequency band assignments, is increasingly becoming an obstacle to innovative use of the spectrum, and that a more application- and technology-neutral licensing philosophy, allowing any application to use the spectrum within prescribed technical parameters, better serves the current environment of fairly rapid change in radio technology and applications.

    A second is an economic point related to the law of unintended consequences - that the original purpose of the ITFS spectrum licensing was to enable educational institutions to broaden their reach through video learning and to tie together geographically separated "campuses", but changes in technology, new use patterns and applications (i.e. mobile internet connectivity), and a failure of the business case for investment in the originally planned application has led ITFS spectrum licensing to largely become a land grant available for lease and a means for educational institutions to earn income from commercial mobile operators.

    I expected better from you guys.

  • Feb 22, 2007 @ 10:44am


    DOCSIS 2.0 provides 38Mb/s downstream capacity in a single 6 MHz channel. DOCSIS 3.0 enables bonding of multiple 6 MHz channels. If someone build a DOCSIS 3.0+ CMTS and cable modems that enabled bonding of all 135 available channels, the theoretical downstream throughput would be just north of 5 Gb/s.

    With 500 households on a node, 60% subscribing, that would be an allocation of 17 Mb/s per household.

    Which doesn't sound like a lot, until you do the math for what's there currently - 38 Mb/s shared among 50% of those 300 households that are cable customers gives an allocation of about 250 kb/s per household, but that's enough for cable companies to offer (and usually deliver) 6Mb/s service.

    Ah, the wonders of overbooking.

  • Feb 22, 2007 @ 10:33am

    Re: Math doesn't add up

    First, even on existing systems, you have space for about 400 channels - typically, 83 analog channels (52-550 MHz) and 306 digital channels (550-860 MHz, 6 digital subchannels per 6 MHz channel, less one 6 MHz channel for DOCSIS downstream data).

    If all a cable company did was to eliminate analog channels and use every channel (but one) for digital video, that would give them 810 channels.

    Because the cable system is broadcast, every STB gets every subchannel, and picks the one (or two) to decode.

    So your 875-1750 TVs could all be on and watching at the same time - as long as they're not watching more than 810 different programs.

  • Feb 22, 2007 @ 06:50am

    Re: Re: Re: helping other things

    Frequency-division multiplexing. Each channel on the cable plant is its own little 6 MHz slice of heaven, and the headend can put any kind of signal it wants into that 6 MHz. So it puts upstream DOCSIS digital data down below Channel 2, analog video signals on channels 2-78, a downstream DOCSIS digital data channel in one of the channels above 78, and multiplexed MPEG-TS digital video signals on the rest of the channels above 78.

  • Feb 12, 2007 @ 02:18pm

    Re: Wait a minute?

    One more time, for those of you who haven't been paying attention.

    The government is requiring broadcasters to stop transmitting analog TV signals over the air. You know, the way everyone used to get TV 30 years ago, rabbit ears, antenna on the roof, all that.

    Because the government is requiring that broadcast to stop, they appropriated money (from the expected sale at auction of the spectrum licenses that are recovered when the frequencies are reallocated from broadcast TV to other purposes) to subsidize boxes to convert OVER THE AIR digital broadcast to analog. This is a normal practice when the FCC changes the use of a particular area of spectrum.

    The NTIA (the Commerce Department agency responsible for implementing the subsidy program) proposed regulations that only households that do not subscribe to cable, satellite, or other similar service would be eligible for the subsidy.

    Reps. Dingell, Markey, and others have sent a letter to the NTIA saying that the subsidy should not be limited in that way.

    So if you're a cable subscriber, you could get a $40 coupon off the purchase of a box that would convert digital over the air broadcast signals to analog. It's not a digital cable box; it wouldn't allow you to hook up your second TV to your digital cable and get ESPN8 The Ocho or any of those channels above 73. It would allow you to pick up your local broadcast TV signals on your old Sylvania 19" TV with the dial on the front.

    Oh, and it would increase the cost of the program to taxpayers by $40.

  • Feb 12, 2007 @ 02:04pm

    Re: Responsibility

    Sure. If you're not a cable subscriber, you'll get a coupon worth $40 towards the purchase of a box that will convert OTA digital signals to analog, so your TV will continue to work just fine.

    What's that? You have cable? Well, the government isn't forcing the cable companies to convert from analog to digital, just OTA broadcasters. As far as the NTIA, the FCC, or Congress is concerned, cable companies can continue to manage their spectrum however they see fit. Cable companies are converting to digital so they can sell more channels, more VOD, more niche programming. Why should I, as a taxpayer, subsidize Comcast's ability to make more money?

    If you think that the FCC should require cable companies to give away digital cable boxes prior to moving any analog channels to digital, that's a different issue.

  • Feb 12, 2007 @ 01:56pm

    Re: Money grab via stupid Digital Box

    The problem is that a digital set top box isn't a tuner. Or rather, it incorporates a tuner - a cheap, analog circuit that picks out the appropriate 6 MHz channel from the 860 MHz of signals on the coax - but it also incorporates a decoder and demultiplexer to pull out one of 6-8 MPEG2 streams from that 6 MHz channel and convert it into a signal that your TV can recognize. The tuner costs a few cents. The MPEG2 decoder, probably around $20-$25 - for each stream (channel) that you want to decode.

    You want to pull 20 streams (10 TVs, each with PIP) off a box? You need 20 MPEG2 decoders. That's $400-$500. Plus you need to control it at the box, not just tune the TV, because it's the box that picks the channel and stream to decode.

    You want to mass-decode every digital channel and put them all on your home cable? No can do - not enough bandwidth on the cable. The whole reason cable companies are going to digital is because they can get more channels into the existing bandwidth. Cable plant can handle up to 860 MHz, and the cable in your house is the same. So you could get 135 channels or so - now you've got to configure the box in some way to pick which of the 500 digital channels it's going to decode.

    Oh, and it'll cost around $2500-$3000 to build. Add in margin, and you're looking at a product that will sell for upwards of $3500. How many people will buy that instead of a couple extra digital cable boxes?

  • Feb 12, 2007 @ 01:40pm

    Re: Plenty To See

    So you want your tax dollars to subsidize the guy next door who has 6 analog TVs in his house and doesn't want to pay the cable company for 6 digital set top boxes?

More comments from DG Lewis >>