Diebold Election Results Released By AZ Judge

Windrip writes "A judge in the case covering the nature of the database used in Diebold Gems software during Pima County, Arizona elections has ruled the DB is not a computer program (pdf). The result is that the Arizona Democratic party will have the chance to review previous elections for transparency and accuracy. ''The Pima County Democratic Party sued the county this year for the electronic databases from past elections. The party requested the databases and passwords be released according to Arizona public-records law. Pima County denied that part of the request, while turning over other records the party asked for. In closing arguments of the four-day trial that began Dec. 4, Pima County argued the databases meet the definition of a computer program, which is protected by state law, said Deputy County Attorney Thomas Denker."

Re:Good.

[Steauengeglase]

A little old, but as I was saying: http://www.scoop.co.nz/stories/HL0309/S00106.htm/ [scoop.co.nz]

Access confirmed in the court ruling

[Dr_Barnowl]

The text of the PDF requires them to release "every file .. that ends with the extension 'gbf' or 'mdb', and the password for 'gbf' files." It also mentions that the data has been scrutineered with Access.

The arguments about an Access database being a "program" are probably related to the ability of MDB to contain queries (aka stored procedures).

GBF files are encrypted / compressed MDB files. The dockit claims that "a gbf file can only be created and opened by the GEMS program", but I suspect it unpacks them to a temporary file somewhere before it opens them up with the normal library.

Other little GEMS (sorry, couldn't resist the pun)...

  * "Microsoft has warned against using the mdb format for some critical applications, such as election management software."
  * Each expert witness endorsed a statement that the GEMS software has significant security flaws.

Re:Good.

[morgan_greywolf]

OMFG. You are serious. The Jet database has long been considered deprecated by Microsoft [microsoft.com].

Re:DIebold Defeats Democracy

[sm62704]

Let me fix those typos for you:

Diebold is the corporation's choice for subverting democracy.

Imagine a world where people vote, but the votes don't matter because the corporations have bribed both wings of the single party in this plutocracy. They just sit in a machine controlled by puppets of the Corporation. We are living this dream.

Re:Not really

[AeternitasXIII]

Except that Diebold's CEO is a member of the Republican party, and one of George Bush's Rangers, a class of high donation supporters for his election campaigns. Money doesn't buy loyalty when your target is already paying off someone else he supports.

Re:Programs, Data, fuzzy distinctions

[tlhIngan]

Most modern architectures blur the distinction by allowing data and code to reside in the same storage, and even allowing you to treat a section of memory as data at one moment and code at the next (which in theory allows for some neat self-modifying code (but that hasn't proven useful in the consumer market at least) but in practice is the root cause of every email virus ever).

Actually, you're referring to Von Neumann architecture. The other architecture is Harvard. Harvard has separate code and data memory (mostly - you still get the convenience of immediate mode addressing in Harvard). But code can only work on data memory - it cannot work on code memory. However, it's only really useful for speciailized computers running the same code on different data (e.g., signal processing - the data is transformed the same way all the time, so the code can reside in ROM, while the data comes in from whatever source is providing it).

The Von Neumann architecture (code and data are intermingled, and one and the same) is your standard computer architecture. However, the behavior is used very often. Think every time you call exec() or CreateProcess() - the OS has to allocate memory, copy the code to memory (i.e., to the OS, your executable program is data), then tell the processor to run the code (now the data is code). Or even consider the bootstrap program - it has to find the OS loader program, which it copies off some storage to memory (data), then runs it (code). It's this architecture that makes modern computing possible...

Re:Then it would be defective by design

[cHiphead]

You are correct, its an .mdb Access db.

Re:Access confirmed in the court ruling

[dcavanaugh]

From ancient times, I remember there was such a thing as an Access "developer edition". It included the ability to take an .mdb file and create a "compiled" executable that was essentially the original .mdb file bundled with a crippled version of Access -- just enough to distribute a database and embedded VBA application to a computer that had nothing beyond ordinary Windows installed. It was a fragile solution -- many ways to screw it up. Along the same lines, the dev kit also included a freely distributable program that could synchronize databases across the internet. It was even MORE error-prone. Typical Microslop.

The original concept of Access was very good -- a personal database with snazzy query, forms, VBA, etc. Problem is, whenever the data has more than one interested party, Access goes downhill pretty fast. Choosing Access for voting machines tells me a great deal about Diebold's IT capabilities. Based on nothing more than circumstantial evidence, I think they chose MS as a vendor, right off the bat. Then they considered MS products that would be useful. Then they tried to limit the cost while meeting someone's hyper-aggressive Gantt chart (prepared in MS Project, of course). Put them all together and you end up with Access. If you release any of those constraints (MS, cost, time) the solution can be made more reliable, more secure, and cheaper. It would be hard to choose ANY other alternative without picking up some kind of benefit.