Washblog

PS to Washbloggers ... the "Recommend" button is near the top of the righthand column.  Try it!

You're only young once, but you can be immature forever -- Larry Andersen
Blogging at Peace Tree Farm

I'm not ready yet to sign on for open source systems, partly because, however open and scrutinized, they still can be subverted...But that's for a different discussion.

This is just to raise the obvious alternative: machine-free, hand-marked paper ballots hand counted in the precinct...This has been so thoroughly discussed among election integrity people that a fairly good initial view of the problems is already available.  A full-fledged design study could get under way any time.

If some of the same kind of effort Rod is proposing was applied to HCPB, wouldn't it be inteesting to compare the outcomes?

Btw: Canada's elections cost about USD 1.80 per citizen annually [not current exchange rate]; Sarasota County, FL, was spending $8.00 at one stage [DREs].

Currently, there exists some form of a federal and state certification process (that obviously has been found wanting lacking by most accounts) - If every county in the nation ran off and implented its' own vote tallying system that certification process would completely break down.

If there were no federal and state certification process for vote tallying systems, voter confidence in the integrity of our democracy would be further erroded rather than improved.

Ginormous distributed systems that are expected to perform flawlessy in the field when operated by half-trained nit-wits (not ours of course, I mean other county's elections departments) that have not read the manual nor followed the required operational procedures and routinely violate any security protocols they are trained in are very hard to get right on the first try (or second, third, fourth, etc), as evidenced by the poor performance of the various vendor's solutions today...  It's a bit of a stretch to assert with any confidence that the lowest bidder responding to our RFP would do much better, open source or no.

I can, for sure, understand why someone with a background at a Microsoft or similar enterprise might not immediately recognize all the advantages Open Source has when it comes to voting but, listen up and please try to keep an open mind.

First off, just because a system is freely distributed does not mean it is well designed, well written, high quality, designed for security or successfully implemented from the security point of view.  BryceM, you are right on this score, but, then, such logic also applies to software kept proprietary and sold for large amounts of money.  It does mean that there is an opportunity to work in the open, to get feedback at each stage of development from the nation's and the world's leading thinkers on voting and security and, in the end, to allow every county to hire whomever it wants to support that technology rather than being confined to the keepers of proprietary secrets.

If you buy voting systems from the commercial vendors, you can hire RABA, Berkeley professor David Wagner, and/or similar experts.  They find that they or even their students can shatter the security measures of all the voting systems from all the major commercial vendors. Why not work with these same experts and get their feedback before the majority of the investment has been made?

Open Source systems can be built to specifications agreed to in advance by the world's experts in secure systems, not simply shown to be swiss cheese after they are completed and federally and state certified.  How does after the fact expert examination do anyone any good after certification and just as sales people are beating the drums for quick purchase of expensive and unknown (to the public) systems? It is too late to fix the voting system then. The basic design of these systems is totally wrong from the security point of view.  We propose to develop Open Source voting systems that can be reviewed at the design stage, during development, before and after certification, so that all the flaws that anyone can find get addressed as early as possible and as completely as possible.  

The way we see it, one can get rid of the gaping holes in security during Open Source development and not after it is too late to correct shortcomings proprietary voting applications.  The CA Top-to-Bottom review found that any single machine on a network could totally compromise every other machine; this could have been nipped in the bud were an Open Source process in place.  As it is, thousands of these machines have been used in voting all over the United States.  The core of security is going to still be in the procedures designed in and around the Open Source voting application, most notably, in the auditing of the paper ballots produced by the voter and then comparing them to the electronic records.

Open Voting Solutions, Inc. has spent over two years putting together  an organization capable of doing this sort of work. We have, on our current team, 3 engineers working on documentation, specifications, and test plans, 6 developers coding modules for voting solutions, and an Open Source community of thousands of highly qualified developers looking over our shoulders.  We are partners with very large companies such as Hewlett Packard, IBM, and Kodak.  We are highly capable of collaborating to good effect with King County on Open Source voting systems and we are not King County's only option for an Open Source solution.

The report of the consultants to King County does not recommend that King County circumvent the federal and state certifications. Quite the contrary.  King County would be hiring a team (such as mine) to produce a voting system and get it though King County's acceptance testing and then, after that, get it certified by the federal Voting Systems Test Laboratory (VSTL) and by the State of Washington. King County has the budget to do exactly that.  The RFP should not give the contract to the low bidder but to a qualified team with the best proposal - we, for sure, would like to be that team, but even we are not, we would be willing to consult with and assist any Open Source team selected.  We would recommend a process where you do not wait to attack the technology at the end and find the flaws too late, as happens repeatedly with the commercial offerings.  Rather, this process would rather work with outside experts at every stage to review specifications, test plans, and specific modules of the source code as it is being developed.   Reviews can be done in parallel, and they need not delay development once specification have been set.

The main advantages of Open Source, as I see it, are:

The technologies can be developed, as indicated above, with input from and oversight by the worlds leading experts.
The Open Source design and coding can take advantage of the current best practices in software rather than being limited by the weight of past errors and bad designs.
The source will be visible to all, the quality of the code will be on display and this is a forcing function for improved quality.
The Open Source technology can be supported by services from the most qualified, responsive consulting organization rather than by just the vendor only.
Reduced cost: A central count scanner from the company formerly known as Diebold is quite expensive. Once developed, a new scanner system could be fielded for a fraction, often half, of the rate for proprietary system. When counties want improvements, rather than waiting for a new release that maybe satisfies the need, they could band together, hire a consulting firm (we are available although not essential) to make the changes, and get the new technology certified in months and not years.

I am not saying that an Open Source system is, automatically, well designed, secure, of high quality, appropriate for any particular county's needs, and so forth. Instead, I am saying that Open Source provides an opportunity (that proprietary code does not) to realize these goals in collaborative development within a reasonable budget.  Note that the commercial vendors have shown themselves to be either unmotivated and/or incapable of building systems without gaping holes in them. Read the California Top-to-Bottom Review: there is nothing proprietary cited here worth buying. The logic is simple:

    When there is nothing to buy, you must build.

If you are going to build a central tabulator system based on a commercial hardware, using, say, a high-speed scanner from Kodak, why not build software on the collaborative model of Open Source?  If you do distribute the resulting application without charge, then other counties can use it and they can share development and certification expenses in the future.  Most of all, King County will have control of development, of features, and of quality and security designed in from the start.  Development using XML and JAVA in a modular modular architecture can readily be accomplished in such a collaborative project.

A system that seems destined to be used in many counties will naturally get the attention of security experts and others who care about voting technologies.  The puzzle of how to get national proprietary vendors to serve public needs has not been solved; the alternative of Open Source collaboration is an open path for King County to get what it wants.  King County needs to take advantage of this approach rather than to just repeat the proven failures represented by commercial voting-system development.

Richard C. Johnson, CEO
Open Voting Solutions, Inc
rcjohnson@openvotingsolutions.com
www.openvotingsolutions.com