Politics and Informatics don't mix well

[miekec]

The Help America Vote Act of 2002 (HAVA) orders each state to create a "single, uniform, official, centralized, interactive computerized statewide voter registration" database that will be linked to other records such as ones stored by motor vehicle agencies. HAVA does say the database must be protected with "adequate technological security" but offers no details and fails to require encryption, for instance.

full article at http://news.com.com/2100-7348_3-6040781.html?part=rss&tag=6040781&subj=news

I'm sure this isn't the biggest threat going on in our lives right now. But the mere stupidity of it just makes me cringe. Leave it to each state to make their own. No guidelines (or rather, no useful ones). Probably no input from people who are good (or even competent) with security or privacy issues.

All states permit voter registration data to be sold for political purposes such as campaigning and direct mail. But 20 states and the District of Columbia also allow unrestricted access for commercial purposes such as marketing

Sure. Register to vote, and lose any privacy you had left under a barrage of unwanted crap.

And no doubt version 2.0.1 will have the added feature of keeping track of who you voted for.

Comments

[johngnassi.livejournal.com]

Duh. Process and security here are much more important than the intrinsic CS database issues. Human and administrative issues will guarantee that this will be done wrong. The quality of the information obtainable from this database in a few years will be so low that you can rest assured that, even though your voting history may be recorded, there will be no possible scheme or method to assure non-repudiation. Just deny everything and they won't be able to say you're wrong. I predict that Donald Duck, Fred Flintstone, and Cthulhu will all be duly registered with validly cast votes. At least according to the DB.

Hospitals can't do this right - they have duplicate "unique" identifiers for people (one of my residents actually had *7*!), ~15% error rate on inpatient charges, at least 1 med list error on > 50% of all patients past the 3rd day of admission - and they have the almighty $ incentive to get this right! Why so many errors? Too many people without vested interest entering data they don't personally value, with too little oversight, too little training, and powerful counterincentives of the underappreciated bureaucratic slacker culture. Add to that lack of personal responsibility (do you think there will actually be a DBA who has knowledge, experience, & commensurate responsibility and decision making authority?), lack of integration and coordination (which leads to the spontaneous generation of the Adams' Not My Problem Field, effectively cloaking the downstream consequences of poor data management practices), and an arbitrary date deadline (instead of a performance and surrogate goal outcomes deadlines).

Given how poorly people and systems perform when patients' lives are on the line (in a hospital), what do you think state administrators are going to do when getting to Happy Hour early on Friday afternoon is on the line (at the soon-to-be-announced Massachusetts Registration of Eligible Voters office)? I'm just horrified.