LA county’s Voting System for All People (VSAP) had a difficult start. Super Tuesday’s debut caused widespread complaints of faulty machines and long wait times, some in excess of three hours. A pre-certification assessment of the system completed in December found multiple vulnerabilities. But the issue with VSAP touchscreens is more fundamental; even if they worked perfectly, the paper output they produce isn’t reliable. For this reason, security experts recommend hand-marked paper ballots for voters who are able to use them and touchscreens for those needing assistance. (1, 2, 3, 4)
洛杉矶县的全民投票系统(VSAP)开局艰难。 超级星期二的首次亮相引起了人们对机器故障和漫长等待时间的抱怨，有些甚至超过了三个小时 。 该系统的预认证评估于12月完成，发现了多个漏洞。 但是，VSAP触摸屏的问题更为根本。 即使它们工作完美，它们产生的纸张输出也不可靠。 因此，安全专家向能够使用选票的选民推荐带有手写标记的选票，并为需要帮助的选民提供触摸屏。 (1,2,3,4)
The VSAP touchscreen is a type of ballot marking device (BMD) — essentially an electronic pen that generates a printed listing of the choices a voter selected on the screen. Initially developed to assist voters who need help filling out their ballot, in recent years commercial vendors have been heavily marketing the devices for all voters. This is a problem because if BMDs fail to mark ballots accurately, studies show that few voters will notice. (1, 2, 4) In the latest study from University of Michigan, only 7% of voters told poll workers of errors. (3) In elections where both candidates and voters depend on accurate results, this disqualifies machine-marked ballots as a universal voting protocol. Instead, hand-marked paper ballots should be used for most voters, with BMDs reserved for people who need help marking a ballot.
VSAP触摸屏是选票标记设备(BMD)的一种-本质上是一种电子笔，可以生成屏幕上选择的选民的打印清单。 最初是为了帮助需要帮助的选民填写选票而开发的，近年来，商业供应商一直在为所有选民大量销售该设备。 这是一个问题，因为如果BMD无法正确标记选票，研究表明很少有选民会注意到 。 (1,2,4)在密歇根大学的最新研究中，只有7％的选民告诉民意调查人员错误。 (3)在候选人和选民都依赖准确结果的选举中，这使以机器标记的选票无法作为普遍投票协议。 取而代之的是，应该为大多数选民使用带有手工标记的纸制选票，而BMD则保留给需要帮助进行选票的人。
To fully appreciate the risks universal use of BMDs create, it’s useful to consider the idea of election transparency. Strangely, voting in the United States is based on a paradigm of trust rather than evidence. Voters and candidates rely on computers operating from instructions that are hidden from view to capture and count votes. We then accept the reported results as accurate and the “winners” take office. Other important transactions, like banking for instance, don’t operate on trust; and no one expects them to. The fact is, we don’t really know what happens to our votes when we walk out of the voting booth.
为了充分理解普遍使用BMD带来的风险，考虑选举透明度的想法很有用。 奇怪的是，美国的投票是基于信任的范式而不是证据。 选民和候选人依赖于从看不见的指令中操作的计算机来捕获和计数选票。 然后，我们认为报告的结果准确无误，“赢家”上任。 其他重要的交易，例如银行业务，则不是依靠信托进行的； 没有人期望他们这样做。 事实是，当我们走出投票站时，我们真的不知道投票会发生什么。
It is in the context of evidence-based voting that election integrity experts critique BMDs. Like any computer, hardware can malfunction and software can be misprogrammed. As the following diagram shows, prior to every election, BMDs receive programming that identifies what races and candidates to show voters and how to interpret their screen input. The programming typically comes from a centrally located election computer (an election management system, or EMS) and is loaded onto a portable memory device, like a USB drive or a memory card, which is then loaded onto the BMDs. Security flaws in the EMS or BMD may allow anyone with access to the EMS computer or the memory card to install vote flipping malware onto the BMDs without detection. These vulnerabilities are demonstrated each year at the DEF CON hacking conference, where white hat hackers have been able to hack into every BMD they’ve attempted (as well as other voting equipment).
选举诚信专家在基于证据的投票中批评BMD。 像任何计算机一样，硬件可能会发生故障，软件可能会被错误编程。 如下图所示，在每次选举之前，BMD都会接受程序来识别哪些种族和候选人向选民展示以及如何解释其屏幕输入。 该程序通常来自位于中央的选举计算机(选举管理系统或EMS)，并加载到便携式存储设备(如USB驱动器或存储卡)上，然后再加载到BMD上。 EMS或BMD中的安全漏洞可能使任何能够访问EMS计算机或存储卡的人都可以将投票翻转恶意软件安装到BMD上，而不会被发现。 这些漏洞每年都会在DEF CON黑客大会上得到证明，在那里，白帽黑客能够入侵他们尝试的每个BMD (以及其他投票设备)。
Simplified diagram showing how vote flipping malware can be secretly installed on touchscreen voting machine and how the QR code hides the official vote on the printout.
简化的示意图显示了如何将翻转投票恶意软件如何秘密安装在触摸屏投票机上以及QR码如何在打印输出中隐藏官方投票。
So why don’t voters just check their printouts for errors? Many BMDs, including LA’s machines, encode selections in a barcode or QR code that voters can’t read. In the diagram, for example, the vote for SMITH was encoded as JONES in the QR code, but printed as SMITH in the human-readable text for voters to review. But it is the QR code vote that is scanned and tabulated, so it’s impossible for voters to confirm what gets counted. This is not how transparent elections are run.
那么为什么选民不只是检查他们的打印输出是否有错误呢？ 许多BMD，包括洛杉矶的机器，都以选民无法读取的条形码或QR码对选择进行编码。 例如，在该图中，SMITH的投票在QR码中被编码为JONES，但在人类可读的文本中被打印为SMITH，以供选民查看。 但是扫描和制表的是QR码投票，因此选民无法确认计数的内容。 这不是透明的选举方式。
The VSAP documentation describes two ways voters can check their QR codes. But both methods involve using additional technology — a different touchscreen or a smartphone QR reader — and then interpreting the series of numbers and letters they see. Setting aside that a different touchscreen could have the same misprogramming as the first, if voters don’t find errors on their printouts using just their eyeballs reading plain text, what chance is there of them using technology to find errors within coded character combinations?
VSAP文档描述了选民检查其QR码的两种方式。 但是，两种方法都涉及使用其他技术(不同的触摸屏或智能手机QR阅读器)，然后解释他们看到的一系列数字和字母。 撇开其他触摸屏可能会出现与第一个触摸屏相同的编程错误，如果选民仅靠阅读纯文本的眼球就找不到打印输出上的错误，那么利用技术在编码字符组合中查找错误的机会是多少？
But as mentioned earlier, voters don’t adequately check their ballots — even the human-readable text. Many voters don’t even look at their ballots, or if they do they don’t find errors, or if they find errors, they don’t report them. And here’s the key; to have an accurate election using BMDs, every voter must check, find, report, and correct errors.
但是，如前所述，选民没有充分检查其选票-甚至是人类可读的文本。 许多选民甚至都不看选票，或者如果他们没有发现错误，或者如果发现错误，就不会举报。 这是关键； 为了使用BMD进行准确的选举，每个选民都必须检查，查找，报告和纠正错误 。
There are other important arguments against BMDs for universal use, such as high cost and increased wait times, but the principle issue is their inability to facilitate what they’re being pressed to do — assuredly reproduce voter expression. Hand-marking, by definition, expresses voter choice; and errors are easily noticed and corrected.
还有其他一些反对普遍使用BMD的重要论据，例如高昂的成本和等待时间的增加 ，但主要的问题是它们无法促进他们被迫要做的事情-确保复制选民的言论。 根据定义，手工标记表示选民的选择； 错误很容易发现和纠正。
Finally, it’s important to clear up the misconception that BMDs are “safe” because their printouts can be audited to confirm election results. A manual audit, which involves hand counting paper ballots, relies on the accuracy of the human-readable votes. If, for example, malware added a human-readable vote for a race the voter left blank, and the error is not corrected, the auditor would count the incorrect vote as correct.
最后，重要的是要消除对BMD是“安全”的误解，因为可以对BMD的打印输出进行审核以确认选举结果。 手动审核涉及手工计算纸票，它依赖于人类可读投票的准确性。 例如，如果恶意软件为选民留下空白的比赛添加了人类可读的选票，并且错误没有得到纠正，那么审核员会将不正确的选票视为正确。
The universal use of BMDs over hand-marked paper ballots is not just overkill, it doesn’t work if we expect verifiable elections. Voters can’t reliably quality test every machine-marked ballot, and they shouldn’t have to. Yet proponents of these devices are pushing that burden onto them, and compromising accurate election outcomes. For what reason? Someone is benefiting from the universal use of BMDs, but it’s not the voter.
BMD在手写的纸质选票上的普遍使用不仅是过分的杀伤力，而且如果我们期望进行可核查的选举，这是行不通的。 选民无法可靠地对每个带有机器标记的选票进行质量测试，因此他们不必这样做。 但是，这些设备的支持者正在将负担加重到它们上面，并损害了准确的选举结果。 是什么原因？ 有人从BMD的普遍使用中受益，但这不是选民。
What can you do? If your jurisdiction relies on touchscreens for all in-person voting, contact your local election official and insist on hand-marked paper ballots as the standard voting protocol. Explain that touchscreens are unverifiable, error prone, unnecessary, expensive, and create long wait times.
你能做什么？ 如果您所在的辖区依靠触摸屏进行所有亲自投票，请与当地的选举官员联系，并坚持将带有手工标记的纸质选票作为标准投票协议。 说明触摸屏不可验证，容易出错，不必要，价格昂贵，并且会导致等待时间长。
1 There is no Reliable Way to Detect Hacked Ballot-Marking Devices
2 (PDF)要求选民验证选票的因素：选民记忆中选票的定量分析
3 Can Voters Detect Malicious Manipulation of Ballot Marking Devices?
4种投票标记设备(BMD)无法确保选民的意愿安德鲁·阿Perl(Andrew Appel)，理查德·德米洛(Richard DeMillo)，菲利普·斯塔克(Philip B.Stark)
扫一扫
736
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
