2centsworth.info

I read an interesting column in the February 2007 issue of Dr. Dobb’s (a software development journal). The column is by Ed Nisley and he is discussing a report on the problems with electronic voting machines in Cuyahoga County, Ohio during the recent Federal election. The column is entitled Root the Vote: Wetware… It doesn’t seem to be on the Dr. Dobbs website (www.ddj.com) yet but it may show up next month.

I don’t think I have ever talked to someone who is professionally involved with computers or software who isn’t anything but appalled by the idea of purely electronic voting.

As Ed Nisley comments in the column, the current issue is how difficult it is to implement a new voting technology with a temporary once-a-year or once-every-couple-of-years organization largely staffed by volunteers. This years problems were due to poorly designed machines, poor organization, and a lack of training. The issue in the future will be deliberate, subtle, difficult-or-impossible to detect, vote tampering.

In my electoral district we use the “fill in the oval” paper ballots which are then scanned electronically. In my opinion, that is the way to go. Most people are able to figure out how to fill in the ovals, there is a permanent ballot that can be recounted. The thousands of paper ballots would be difficult to alter without its being noticed.

Given the tendency towards dirty tricks in US politics and the generally corrupt relationship between government officials and government contractors, I think purely electronic voting is a recipe for disaster for our democracy…. Just imagine the closed room conversations… “Not only can we guarantee a smoothly run election… For only a few million more we can make sure your guys win!”

If we end up with a significant fraction of the votes being counted in a purely electronic manner then we will also get deliberate biasing in the outcomes. And it will be, essentially, undetectable. You can at least argue about a “hanging chad.”

In fact, the first we will know that there is anything wrong is when some techie gets sufficiently disgusted (or decides he wasn’t paid enough) and we wake up to find we’ve just elected Mickey Mouse president.

Ed Nisley called attention to a recent New York Times article (U.S. Investigates Voting Machines’ Venezuela Ties – October 29, 2006) about the Venezuelan government buying a US company that makes electronic voting machines… That were purchased for use in Venezuela’s elections. Apparently this is being investigated by the Bush administration… One is being asked to believe that a Republican administration is suddenly interested in free and fair elections in South America?

If you want to find out more, check the following links:

Cuyahoga Election Review Panel – Final Report

Election Science Institute – DRE Analysis for May 2006 Primary

Avi Rubin’s blog

VerifiedVoting.org

This is an apolitical public service announcement for everybody that is sick and tire of listening to these automated customer service phone systems (I suspect the irritation transcends all political, religious, ethnic, and racial bounderies!).

Most of us have been confronted by these automated phone systems asking us to press a particular number on the phone keypad depending on which category of disgruntled customer we are. And one frequently finds that your particular beef does not correspond to one of their automated bins… And you really, really want to talk to an actual human being even if he is in some third world country and barely speaks English…. But you have to wade through interminable layers of “Press 5″, “Press 2″, etc. to actually get to a live person.

Somebody has set up a website GetHuman.com which tells you the secret key combinations that allow you to get to speak to an actual person at a lot of large organizations. Actually, that is a little too optimistic… First you will probably hear “We are sorry but all of our associates are assisting other customers at this time but your call is important to us so you just sit there while we play you some elevator music…”

But at least GetHuman.com will help you skip the thicket of keypad options.

Since this is the first award we are giving out, a little explanation about the Anal Leakage Award is in order.

This award has been created for technologies that are sufficiently frustrating, obnoxious, or evil that we would all really prefer that the inventors had “puckered up” and kept it to themselves. It is named in honor of a certain undigestible fat used in many diet foods.

This particular award goes to Royal Philips Electronics. A recent New York Times article (May 7, 2006) entitled Someone Has to Pay for TV. But Who? And How? by Randall Stross describes a recent patent filing by RPE for a TV remote control which will lock the channel during commercials. And there is apparently a companion design concept for a video recorder which will not allow the user to fast forward through commercials.

This is a really brilliant use of technology and I am sure the engineers responsible for designing these devices are truly proud of their work. But, really, why stop there? Why not patent a device which locks the refrigerator during the commercials?… Or heck, how about the bathroom?

Google still seems to be the darling of Wallstreet but I wonder…

If I was invested in Google stock my recent experiences with Google would be rather sobering.

Google Adwords

I have been having a lot of problems with Google Adwords. The basic idea is that you bid on a list of keyword phrases… If a Google end user searchs for one of the phrases you have bid on… He gets the regular Google search results on the left hand side of his browser page and your sponsored link will appear on the right hand side. How prominently your ad will be placed depends on whether your bid (say $0.25 per click) for the search phrase is higher or lower than your competition.

We are marketing Boston Christmas cards… So one of my Google Ad phrases is “Boston Christmas Cards” but it does not show up… Originally I had set an option in Google Adwords that said that I wanted my sponsored link to be “regional” i.e. it should only show up if the person searching was connected to the internet in the Boston region. And since I live and work in the Boston area, my sponsored link should show up when I Google for “Boston Christmas Cards”. But it doesn’t. The reason being that Google Adwords thinks I am browsing from Alabama. The Google Adwords diagnostic tool identifies my location (based on my Verizon.net ISP connection IP address) as follows:

Domain: www.google.com
Language: English (US)
User Location: Alabama, United States

But if I go to a publicly available tool at dnsstuff.com it easily finds my correct location:

Your IP is 141.154.254.246.
Welcome visitor from [City: Boston, Massachusetts] United States.

This is just one example of the sort of technical glitsches I am encountering with Google Adwords. Even when presented with this sort of “smoking gun” evidence of a problem, the Google CS folks just say everything is working fine.

Google Print

I have a customer who has published a book. He wanted to have the book made available on Google Print. We submitted all the necessary PDF files well over a month ago and the book still is not available on line. When we check the book’s status we find that it is listed twice as “pending”. This was called to Google Prints support folks attention a number of weeks ago but the book is still double listed and is still listed as “pending.”

Google “Sandbox” phenomenon

Finally, the heart and soul of Google, their basic search capability seems to be slipping. If you do a search on “Google Sandbox” (heck try it in Google) you will find numerous references to a phenomenon whereby new sites are ranked much lower than sites that were created and indexed earlier. The general consensus amoung the people specializing in Search Engine optimization is that the date threshold for being ranked lower is currently sometime in early 2004.

Now Google is perfectly entitled to use any search criteria they want but, if it ends up producing search results that are of less utility to the end user, they do so at their peril.

As you will find in the various articles on the “Google Sandbox Effect”, one way of identifying if a site is being “sandboxed” is to search using a relevent search string in Google, MSN, and Yahoo. If the target site is well ranked in MSN and Yahoo but doesn’t show up in Google then there is a possibility it is being “sandboxed”. The actual criteria are more complicated than that. My point is that, in many cases, the “un-sandboxed” results of MSN and Yahoo are much more useful to the end user than Google’s results are.

Conclusion

It is certainly too early to tell if Google is going to be displaced as the king of the search world but I see some disturbing trends. They seem to be diversifying faster than their technology can be implemented reliably, their CS can be best described as complacent, and they are being targeted by competitors with considerable resources and capabilities. At this point Google is synonymous with searching it may end up being metaphor along the lines of the Titanic.

Outside of the techie world, remarkably few people know the history of the Internet and the World Wide Web… Or even understand the distinction between them. I am certainly not an authority on either but I have “grown up” with both.

Back in the 70’s, the Defense Department’s Advanced Research Projects Agency (ARPA) saw the potential value of networking computers together. They funded the development of a network architecture by a mixture of private companies (e.g. Bolt, Beranek, & Neuman) and various academic institutions. My understanding is that they specifically asked for an architecture that would resist the loss of various nodes due to battle damage. The result was a network (called ARPANET) that tied together computers at federally-funded academic institutions, military facilities, and government research labs. The networking protocol that was used is called TCP/IP. Ironically, no ARPANET or Internet node has every been destroyed in battle but the battle-resistant architecture has proven to be reasonably well suited to dealing with the loss of nodes due to routine equipment failure. [Hmm, maybe the resistance to battle damage requirement is a myth... According to the Resource Center for Internet History, ARPA developed the Internet for purely scientific purposes with no intent for military use.]

For the first couple of decades, ARPANET (later the Internet) was known to, and used by, a few thousand academic and military types. The primary use was for email and file transfers.

In the late 80’s, the government decided to open up the Internet for public and commercial use. It is largely forgotten now but this was quite controversial. The priviledged elite who had used the Internet up to that point objected to masses of unwashed newbies getting access to the network.

Of course the real revolution was not the opening up of the Internet for public use. After all, the average person saw little use for email or electronic file transfers. The real revolution was the development of the World Wide Web by an obscure Englishman called Tim Berners-Lee. He was working for CERN which is a multi-national physics laboratory sponsored by the European Union (roughly corresponding to one of the US Governments national laboratories like Lawrence Livermore or Argonne).

His intent was simply to provide a means for sharing documents (physics papers, primarily) among the staff of CERN which numbered in the thousands and which was located at a number of different sites around Europe. All these sites were networked using the TCP/IP Internet and he built a computer architecture that layered on top of TCP/IP and allowed people to publish information on a server on one part of the Internet that could be viewed by someone else using a client browser at a remote location.

Berners-Lee apparently had no plans to “change the world” but he was a very competent programmer and designed an architecture that was scalable, robust, and very flexible. The same architecture he designed to serve the needs of a few thousand people in Western Europe in the late 1980’s is still being used… But now it is being used by millions of people across the planet. We now use it buy products, make travel arrangements, check the weather, read the news, publish our opinions, settle bets, and a million other day-to-day tasks.

In my opinion, no single individual can have had as great an impact on technology, business, and politics in as short a time as Tim Berners-Lee but hardly anyone knows his name.