I’ve noticed that blue LEDs have become quite popular with electronics manufacturers, and especially in computer components these days. 

During my last cleaning I was advised by the hygienist to start using an electric toothbrush.  The braces give me fits trying to get everything out with regular brushing.  She gave advice on a couple of brands, one of which was Oral-B (Braun).  So I went out and acquired an 8850 DLX, mounted it on the wall in my bathroom, and plugged it in. 

The first thing I noticed was that it’s inductively charged.  I thought this was an ingenious solution to dealing with the problem of getting power to a unit that needs to be watertight, since it didn’t require any external electrical contacts.

Anyhow, the thing has a blue LED on the front that flashes while charging and stays on once it’s charged.  The problem is that the thing is so bright that the flashes illuminate my bedroom if I leave the door to the bathroom open.  I suppose I could just close the door, but I’ve gotten so used to having it open that doing otherwise doesn’t feel right (you get used to a certain “feel” to the room at night).

I suppose my point in all this is that blue LEDs are much brighter than the old red and green ones, to the point of being capable of lighting up a room.  I wish the manufacturers would take this into account and dim them a bit in the future.

In my spare time (such as it is), I’ve been attempting to build and install a J2EE application server on a Linux system.  My intention is to use it for testing and some other development activities when the new project gets rolling (we’re in that interim phase between submission of the cost/timeline and being given the go-ahead to get started on development). 

I built an AMD Athlon 64 system and put SuSE 9.2 professional on it.  Now my office is already full of hardware, and it’s on the verge of being uncoolable in the summer.  So I decided I needed to put this new system somewhere else.  With all the fans it’s pretty noisy, so I decided to put it in the guest room.  It’s not used often, and if it’s being used then I can shut the system down.

Having decided on the placement, that left the issue of getting network access.  I didn’t want to hard-wire an Ethernet port in there, and I definitely didn’t want wires strewn across the floor between there and the office, which left me with the wireless option.  Wireless on Linux is something of a difficult area, though, since most chipset makers won’t release either a driver or the specs to create a driver.  However, Linux geeks don’t give up easily, so there are various reverse-engineered drivers available.  And for those cards that won’t easily give up their secrets for reverse-engineering, someone created a handy hack called ‘ndiswrapper’, which lets you load the Windows driver in Linux.  I did some research on the various sites and decided on a Linksys USB 802.11g adapter, since it showed support through ndiswrapper.

After installing SuSE, I noticed that the adapter wasn’t seen automatically, which was to be expected.  So I tried ndiswrapper with the Win32 driver.  It reported that it couldn’t see the hardware, which was sort of an inauspicious start.  Some research showed that a newer version of ndiswrapper corrected this problem, so I downloaded the package, installed the kernel sources, and built a new ndiswrapper.  That went well, and ndiswrapper reported that the driver had been installed and the hardware was present.

Then I ran into the 64-bit wall.  The driver wouldn’t load, since I was running a 64-bit system and the driver was 32-bit.  Some checking showed that there are very few 64-bit Windows wireless drivers, and to use ndiswrapper on my system I’d need one of those.  I briefly considered trying to run the system with the 32-bit kernel, but the SuSE installer was too smart for me.  There is no option (at least that I could find) to override the automatic architecture detection in the installer.  So I was stuck with 64-bit (even though the AMD64 architecture allows full backwards compatibility with 32-bit code).

A little more research into the area of 64-bit Linux with wireless showed that the solution was to find an adapter that used one of the chipsets that had a “native” Linux driver.  Some more research on drivers and chipsets allowed me to make a list of potential PCI cards to use and a quick trip to Staples yielded a NetGear WG311TNA PCI card.  I popped the card in, booted up, and SuSE automatically found the new card and let me configure it.  After all that fiddling, it was almost a letdown for it to be that simple. 

Despite having been out for over a year, I suppose 64-bit processors in the PC world are still pretty new.  What’s bad is that I didn’t even think about 64-bit being an issue.  Oh well, they don’t call it the “bleeding edge” for nothing.  If nothing else, I can serve as an example to others of what not to do.

I generally have low expectations for my dealings with large companies.  In fact, I usually expect the worst.  This allows me to be pleasantly surprised if it doesn’t happen.

I’m now using Verizon Fios for my Internet service (15M/2M) and I’ve been very pleased with it (there has only been one outage and that was caused by a technician who accidentally unplugged my connection while hooking up another one; the helpdesk got it corrected within an hour).  In fact, I unplugged the cable modem the day they installed Fios and never plugged it back in.

So it was with a bit of trepidation that I phoned Charter a few months ago to cancel my Internet service with them.  I had already experienced problems with their billing system trying to charge me rent for a modem that I owned.  While they eventually corrected it, I suspected that this would cause a problem during the cancellation of service.  And I was not to be disappointed in that respect, since my next bill arrived with a charge of $45.00 for “unreturned equipment/cable modem.”  I called them the same day and they removed the charge.  All was well for the next bill.  But this last bill came marked “past due” and the $45.00 charge had returned with the addition of an extra $3.25 late fee (although there were no details given, it just appeared as the “outstanding balance”).  As I was sitting here this morning summoning the wherewithal to call them again, they called me.  At first I was a bit ticked, since it was an automated message that informed me to call customer service about “an important issue with your account.”  But the message allowed me to press ‘9’ to connect to a rep, so I went ahead and did it.

I guess I’m not sure how I feel about this whole thing.  The reps have all been very polite and nice about trying to fix the problem.  I’m just a bit frustrated that it won’t stay fixed.  Verizon is in the process of getting a franchise agreement with the city of Keller and if Charter keeps screwing things up it will be pretty tempting to switch.  I’ve already got the Verizon equipment installed and it would just require them hooking up to the coax on the outside of the house.  There might be a few equipment switching pains (e.x. getting the Tivo to control their digital cable box as well as getting their guide data), but overall it seems fairly simple to switch.

The Keller city council was supposed to vote on the agreement last night, although I haven’t seen anything in today’s paper indicating the outcome.  Given how eager they’ve been to get Fios in the city I’d hope that the agreement would happen pretty quickly.  However, this council seems arbitrary and capricious to me at times, so there’s no telling how it turned out.

Whenever I do a Hotsync with my Tungsten T3, I wait for it to finish and then hit the power button to turn it off.  Sometimes, though, it turns itself back on after a few seconds.  Usually, turning it off again takes care of it, but there have been a few times where it’s waited a few seconds and turned itself back on again (sometimes it decides to be even more persistent and I have to repeat this cycle three or four times to get it to stay off).

I think my Palm is haunted.

There is a format war underway for the next generation, high-definition DVD standard.  Interestingly enough, one of the major players in determining which format ultimately wins in the marketplace may be the porn industry.

One of the next big issues in which pornographers could play a deciding role is the future of high-definition DVDs.

The multibillion-dollar industry releases about 11,000 titles on DVD each year, giving it tremendous power to sway the battle between two groups of studios and technology companies competing to set standards for the next generation.

“It’s sort of like the buzz around the campfire,” said Peter Warren, DVD editor at industry bible Adult Video News.

One side of the divide is a standard called Blu-ray backed by consumer electronics heavyweights like Sony, Philips Electronics and Thomson and movie studios Fox and Disney. Blu-ray offers storage up to 50 gigabytes, enough for nine hours of high-definition content.

On the other side of the fight is HD DVD, which has much the same structure as current DVDs and, backers say, is cheaper and easier to manufacture as a result. Supporters of the disc format and its 30GB capacity include companies like NEC, Toshiba and Warner Home Video.

While the porn industry may influence the format war, it got me to thinking about another issue.  I’ve heard that HD really brings out details and any flaws will be magnified tremendously.  I wonder if HD is really such a good thing for this industry….

I think the advent of digital controls for audio devices has been something of a two-edged sword.  While they often offer easier and more precise control in some areas, the one where many of them fall flat is in volume control.

Of particular annoyance to me is the fact that most digital volume controls operate based on a digital step scale of volume versus a continuously variable scale like the older potentiometer-based controls.  This means that even if you have a volume control knob on a digital device, it likely just operates by steps.  If the designers of the device took this into account, then the steps are very small and the device simulates a potentiometer pretty well.  The problem comes in when the designers of the device use larger steps.  The one that’s been bugging me of late is the volume control on the radio in my Avalanche.  It’s a factory system and I just can’t seem to find the right volume level, especially when I have someone in the vehicle with me.  It’s either slightly too low or slightly too high.  Clicking the button on the steering wheel or moving the knob on the face both act the same way: they move the volume by one quantum at a time in either direction.  There is no in-between or fine control.

Unfortunately, given the incestuous* way in which the factory radio is wired into my truck, it’d be a real pain (not to mention expensive) to replace it.

* I’ve got the 6-CD changer with XM radio and the Bose system in this truck.  It also has rear-seat audio controls and OnStar.  In addition, in 2003 Avalanches and newer, the radio is used for the door chime system.  If you remove the factory radio you remove the chimes that warn you about leaving the headlights on, that your key is in the ignition, and in my case, the chime that tells me the 5-second delay locks have been activated.  If replaced, the best that can be done is to put in a new head unit with its own XM receiver and an adapter that lets the OnStar use the front speakers.  The rear audio controls would be inoperative as would all the chimes.  But even this has to be done carefully, as the whole damn thing is on the vehicle’s data bus, which can interoperate with the Body Control Module and the Powertrain Control Module (!).  The fuse to the above adapter has to be pulled if they reflash either of these or the modules can be made inoperative.  In this case, losing the PCM would turn the Avalanche into a 5000-lb paperweight.  All in all it’s just too much damn trouble to replace it, so I’ll just keep being annoyed by the volume or maybe get used to it in a few years.  It’s just not the sort of thing you notice on a test drive, though, although I will pay more attention in the future.

I noticed that XM and Delphi are trying to generate buzz for the new Delphi MyFi as a Christmas gift.  The MyFi is a handheld XM radio receiver with a built-in five-hour digital recorder (sort of a cross between audio TiVo and a VCR).  Given my experience with XM over the past two and a half years, I’m not convinced that this type of product is going to go over well. 

To receive XM in real time you need to have an unobstructed view of one of XM’s two satellites, which are in geosynchronous orbit over the continental US.  For heavily built-up areas (Dallas, New York, etc), where the view of the sky is likely to be obstructed XM has installed terrestial repeaters.  So if you happen to live in one of these areas, perhaps the device would work for you.  However, if you’re elsewhere, I would expect that simply going into an inner room in a house would cause you to lose signal. 

XM is great when you’re on the road, but I just don’t see how it would be a satisfying experience as a portable, given the potential reception issues.

A test of one of the potential components of the ballistic missile defense system failed today.

President Bush’s drive to deploy a multibillion-dollar shield against ballistic missiles was set back on Wednesday by what critics called a stunning failure of its first full flight test in two years.

The abortive $85 million exercise raised fresh questions about the reliability of the first elements of the plan, an heir to former president Ronald Reagan vision of a space-based missile defense that critics dubbed “Star Wars.”

Looking at the Google News feed for this story it was hard not to notice the gloating of the media at this turn of events.  It seems to me that the purpose of testing is to uncover these kinds of faults and fix them.  If all you have are successful tests, then it would seem that something is missing from those tests. 

I think the unfortunate thing is that there are a lot of people who for some misguided reason think that defending ourselves against missiles is morally wrong.  There is significant overlap between these idiots and the press, so any failures tend to be trumpeted as a case for the whole system not to work.  Intercepting ballistic missiles is a nontrivial operation, so it will take a while for the technology to reach maturity.  I also doubt that we’ll ever come up with a foolproof system without perfecting our energy weapons capability (or at least it seems easier to me to track and hit a missle with a directed energy weapon like a laser than it would be to intercept it with a rocket, although I’m not a rocket scientist…). 

With the normal stoicism of the Bush administration, I suspect they can easily weather the bleatings of these idiots, although the weaker members of Congress may vacillate when it comes time for further funding.  Given that the government of North Korea is batshit crazy, has nukes and missiles, and is a perfect example of the “success” of socialism we need a missile defense system sooner than later, just in case Kim Jong-il decides to take a few running dog capitalists with him when his country finally descends into chaos.

It’s interesting that Glenn Reynolds would have an article about customer service robots up today, since I had to deal with one this morning.  While his experience was positive, I’ve never liked dealing with them and my experience today didn’t exactly give me any more warm feelings for them.

The power failed here just before 9:00am.  As soon as I confirmed that there wasn’t a problem with my breaker I called TXU’s “customer service” number.  When you select the option to report an outage you’re dumped into a second phone system with an obnoxious overly chipper female voice.  They obviously didn’t think it out carefully, since the first thing it says to you is, “I’m sorry, I wasn’t able to get your account number.”  It then asks you to speak or key in the account number.  After navigating through this and speaking the right phrases at the right times I was assured by the chipper robotic voice that my problem had been recorded.

After a couple of hours had passed I started to wonder if they were doing anything about the problem.  The power was still off and I hadn’t gotten a call on my cellphone (which I had given to the chipper robot when it asked for a contact number).  I called back, determined to find some way to break the robot’s programming and speak to an actual human.  I found that when it asks you if your account number is right that it has no provision for keypad input (i.e. some systems allow you to enter ‘1’ for yes, but this one required you to speak “Yes” or “No”).  If you enter a number it says that it didn’t understand you and to either say “Yes” or “No.”  I hit ‘1’ again and it said that it was having difficulty understanding me and was transferring me to an agent.  Ah-ha!  Success at last.  After a short wait I got a person on the line and reported the problem.  The power was back on within 20 minutes of reporting it via the person.  So, from now on, I’m going to be purposefully obtuse in order to get a human on the line.  I don’t trust the damn robot to actually record my problem and dispatch someone to fix it.

Now I’ll be up front and confess that I despise talking to machines.  First, I don’t trust voice recognition to get things right.  Second, it feels silly to talk to a machine.  Third, most of these damn robots are too chipper and familar.  They come across to me as being unserious, which leaves me with the impression that my problem isn’t being taken seriously.  The whole concept of using a robot to keep from having to deal directly with me reeks of contempt for the customer.  It says that these companies will do anything in their power to keep us, their paying customers, at arms length.

However, despite my hatred for the damn things, it would appear that customer “service” robots are here to stay.  While talking to a friend this afternoon, he mentioned that he’d discussed these robotic systems with one of their vendors and was informed that their surveys showed that 85% of people would rather talk to the machine than key in information.  If we assume that the vendor’s survey is correct (and I would look carefully at it, given that it’s in his interest to show that people like the systems he’s selling), that still leaves the other 15% of us who despise the damn things.  I guess it depends on your market whether you can afford to offend and annoy 15% of your customer base.  While TXU has to face competition, there is still a lot of old-style monopoly mindset in the way they treat customers.  Perhaps they don’t care if they piss people off.

Anyhow, if companies are interested in keeping their customers happy, rather than just minimizing their callcenter costs, they might want to consider a few design points when implementing these voice response systems.

1. Every menu must have an “escape” to a human operator.
2. Voice recognition systems must have the ability to accept keypad input for those who do not wish to speak to the machine.
3. If a call has to be transferred from one phone system to another it must not be done in such a way that the call enters the second system in an “error state.”

It seems that printer companies have been embedding the model and serial numbers in the output of their color laser printers and color copiers.

Next time you make a printout from your color laser printer, shine an LED flashlight beam on it and examine it closely with a magnifying glass. You might be able to see the small, scattered yellow dots printed there that could be used to trace the document back to you.

According to experts, several printer companies quietly encode the serial number and the manufacturing code of their color laser printers and color copiers on every document those machines produce. Governments, including the United States, already use the hidden markings to track counterfeiters.

What bothers me about this isn’t so much that it’s there, but that we weren’t told about it before it was done.  The technology is ripe for abuse.  Sure, it can be used to catch the really stupid counterfeiters, but it could also be used to catch people who have legitimate purposes.  Consider a whistleblower who prints incriminating documents and submits them anonymously.  The affected wrongdoers, if they obtained the documents, could potentially decode the serial numbers and work back to the person who printed them.  This doesn’t necessarily require access to the printer manufacturer’s records, either, given that companies could potentially know the serial numbers of printers assigned to employees (or at a minimum, narrow it down to a certain shared printer).

And what about government abuse of the system?  There appear to be no safeguards built into the law, since this is just something that the printer companies decided to do to be “helpful.”  There is no subpoena requirement, no probable cause, no warrant, no review by a judge.  The company just gives up the information to the government upon request. 

If there really is a legal reason for the government to need to know the information, it won’t hurt them to go through the normal due process procedures and get a warrant.  Even though I don’t trust the judiciary these days to do much more than rubberstamp warrant requests, it would would at least provide a bare minimum of oversight to avoid some abuses.

Update: GeekWithA.45 has some more thoughts on the subject.