What’s With Diebold?
I have noticed people [Eccentricity and Sideshow ] have been wondering how a company like Diebold, that produces thousands of ATMs, can make a product as flaky as the electronic voting machine they sell.
The first thing you need to know is that Diebold didn’t develop the voting machine in-house, they bought a small company that created them. After the purchase they kept the sales and marketing people and fired the technical people. The assumption was that Diebold’s own technical people could deal with the new product.
The technical people working on Diebold’s ATM equipment have a stable, proven platform that was created to exacting standards and reviewed by technical people from the banking industry. There are no cutting edge technologies in a Diebold ATM, everything is bullet-proof and over tested.
The voting machines were created by a start-up and given the infrequency of elections, a couple of times a year at most, and the lack of technical staff by the customers, the multiple standards that do not address programming issues, it would be amazing if you could produce a stable product. The people who created what Diebold purchased are long longer with the company and the probably of that they produced really good documentation of what they created, or actually did extensive testing is almost nil.
Microsoft actually has an extensive testing program for new software, but, as we all know, the bug reports start appearing within hours of any release. It is obvious that these machines haven’t passed the gamma testing, much less the beta, and the are being released to people without any desire to be technical.
In the typical voting precinct, the people in charge will have a few hours of training and then be handed several of these machines with a couple of typed sheets telling them how to set it up. If there are any problems, there is no local support to help them. Diebold’s attitude about the software means there are no local people trained to deal with problems.
I’ve been messing with computer equipment since 1970. I’m a “bit diddler”; I have worked in machine code; I go back to the days when booting the machine meant hand loading the first part of the boot program with toggle switches; I have written better than a million lines of code; but if I have to look at something I haven’t touched in a year; I have to refresh my memory on what I wrote.
The Diebold system, and other computer voting systems, expect people to set up a piece of equipment that they see a couple of times a year, at most. To work, that equipment has to be more trouble free than an ATM, and it needs a printout for diagnosis, if nothing else.
Diebold isn’t doing its reputation or elections any good with their current system. If they think this is a serious market, they should have invested some serious resources in producing a serious product.
4 comments
Bryan,
Thanks for the background on Diebold & voting machines. It reminds me of my father’s adage, “Never assume malice where incompetence will suffice.”
I’ve certainly dealt with my share of systems designed by engineers, for engineers, with every cool bell and whistle, minimal documentation, and less testing. (OK, I’ve designed my share of such systems as well. ;-)) It hadn’t occurred to me that Diebold had pulled the same stunt. I had thought they were better, at least technically.
Diebold has the tech people, but I doubt they want to go near this mess. As near as I can tell they are using a version of of Windows and MS Access. a pathetic combination, which uses VBA to tie things together. A nightmare to secure.
Question: How come the open source folks don’t/can’t/won’t write new software for these machines? Seems it could be a win-win situation.
That’s what they did in Australia, and they submitted the code to a hacker contest to squeeze all of the bugs out of it. The clowns currently in charge in this country want to “privatize” everything, so you end up with crap like this.
All of the government contracts I’ve ever worked on required the source code as part of the contract and the language was specified. This current mess reflects the fact that the government has outsourced their IT, and they don’t have anyone left who can write a decent request for bids, or a contract. If you look at all of the major failures for government IT contracts, the underlying problem is that the contractors fulfilled the contract as written, rather than doing the job that was needed.
The last contract that I was asked to work on I turned down because I knew what they specified in the contract would not work. It would not provide the information that was needed and the people involved didn’t understand what I was telling them.
They went ahead with it anyway, and the guys who had the primary contract told me they had finally gotten their money, but the project ended up being scrapped because it didn’t work. That was a few hundred thousand dollars that the taxpayers will never see again.