More technical discussion of Austin's Open Government Online Initiative

My last post put together a theoretical system that should be able to handle a near "real-time" posting of information in an online format if required. It was meant merely as a basis for discussion; to set the technical parameters of a system so we can break down specifically and in more detail, the functionality of the system in specific situations. In this post, I'll try to highlight and explain exactly how this system *should* work in situations that have been mentioned publicly. But first, a standard disclaimer:

Although I currently work for the City of Austin, I am in no way representing or advocating *any* official position of the City in this matter. I am writing this as a concerned citizen of Austin who would like to see our government as transparent and efficient as possible and I am using information that is publicly available elsewhere to make my points.

With that in mind, this is a theoretical plan so assumptions have been made which could be incorrect. At this time, I am neither for nor against the proposed Charter amendment. My focus with this post is to try and bring a technical solution to the mandated goals of the Charter amendment in which to serve as a base for discussion with the Austin technical community. This opinion is put forward in the hopes it will facilitate discussion and will bring other concerned citizens into the mix and to showcase the alternative technical ways to satisfy the Charter amendment if passed. So please, do not read anything into this post that is not explicitly stated.

I'll first discuss the situations mentioned in the ballot language proposed by the City Council. Here is the language that will be going on the ballot this May (italics are direct quotes; my responses are normal text. Further, I've changed the format of the language to make the points more clear [I don't think anyone can read big, run-on sentences very well]):

City Language

Shall the city charter be amended:

1. to require that all private citizens emails to any public official be placed on the city website in real time, including emails or electronic communications between private citizens and public officials in all City departments, including the library department, police department, city health clinics and city departments handling utility bills and code enforcement, and limit the ability of citizens to keep private the details of these communications;
2. to require that the heads of all city departments, including the police department, parks department, library department, all city manager s staff and all city council members and their staff post online in real time information about all meetings and phone calls with private citizens;
3. prohibit the city from exercising state law protection for information that could expose the city and taxpayers to greater financial and legal liability and risk;
4. to require the city to create at taxpayer expense an online electronic data system for most city communications and documents, which for the most part are already available to the public; and
5. to install and permanently operate such a system at an estimated cost of approximately $36 million initially and $12 million annually thereafter if fully implemented, which could require a tax increase equivalent to three cents per $100 valuation or a reduction in city services?

First, let me say that there are those that feel the ballot language is not an accurate portrayal of the amendment's goals. I won't comment publicly my thoughts on the language at this time but I will address the technical concerns of the ballot language in the proposed system.

a) to require that all private citizens emails to any public official be placed on the city website in real time, including emails or electronic communications between private citizens and public officials in all City departments, including the library department, police department, city health clinics and city departments handling utility bills and code enforcement, and limit the ability of citizens to keep private the details of these communications;

Email is already subject to Open Records requests and state document retention policies so the only difference in this clause from a restatement of current policy is the "real time" bit (which I will address). Further, the amendment does not require breaking any existing federal or state laws in its effort for transparency (i.e- City health clinics, police departments and utilities are bound by federal laws like HIPAA and the USA Patriot act which supercede the City Charter's authority for online posting of information.) Privacy laws (whether federal or state) still supercede any authority the city charter may have (an important concept that seems lost within the context of the current ballot language).

The relevant section(s) from the amendment is:

SECTION 2: Privacy Protected. Nothing within this amendment should be interpreted in a manner that would violate an individual’s existing constitutional or common law rights to privacy.

and

(C)OPEN ACCESS TO CITY ELECTRONIC COMMUNICATIONS

(1)In order to better preserve written electronic communication for public disclosure, the City must establish a system that automatically archives all incoming and outgoing electronic communication that deals with City business to and from the following people in their official capacity:

1. City Councilmembers and their staff;
2. City Manager and his or her staff;
3. Assistant City Managers and their staff; and
4. all department heads.

The amendment merely calls for an archiving system for the above personnel and does not require "real time" capabilities. The city currently follows all state document retention laws which require email archiving for a period of no less than 2 years.

As it is correct to legally interpret the language of the amendment as broadly as possible, the current ballot language does not recognize that the ultimate authority of "real time" resides with the city council and not with this amendment. Interpreting the amendment as broadly as possible, it is correct to assume that "real time" posting of email would be a goal if that action was practical (that is the keyword from the amendment). The city council is the determining body of the "practicalness" of any issue that falls outside the explicit conditions set by the amendment. Hence, in this limited example of "real time" email, because it falls outside the explicit condition set by the amendment, it should be referred to council for determination of how practical the idea is (the key clause from the amendment is "expeditiously as possible and to the greatest extent practical,"). If the council found it was practical, it would then be up to the council to determine the action to best fit the situation.

Further, the OGO amendment only mentions "real time" in two explicit circumstances; calendars of the city council, city managers, division heads and their staff and written communications between the City and businesses and individuals seeking ecomonic development benefits. Anything outside the scope of the following sections would be referred to council for a practical definition. The amendment *does not* require "real time," online posting of email (even though the press continually miss this point).

SECTION 3: Open Government Online

(A)OPEN ACCESS TO CITY BUSINESS

(4)All public information concerning the matter subject to Section 3(A)(2) must be posted to the website. All written communications between the City and the applicant relating to the matter must be posted online in real time in a manner searchable by the public.

(B)OPEN ACCESS TO CITY CALENDARS

(3)Calendars and logs must be posted online in real time and be accessible to the public.

(I like to think of this situation as the city council's "Mars Mission." Even though a manned mission to Mars is possible, it still is not practical and the council could rule for various reasons [the cost is prohibitively expensive; no one is trained for the mission; the chemical rockets are not powerful enough to make it to Mars, the citizens don't want the extra $.03 per $100 valuation tax increase, etc] that the "Mars Mission" would not take place.)

The proposed system could meet the goal of "real time" posting of email if required by council at a later date. On email reception, the email is "tagged" with an OGO compliant label (the label is contained within the metadata for that message). Retrieval and display of those messages would be a simple ad-hoc query from the Zope application server to the Zimbra MySQL-based metadata store for the appropriate tag (pseudo-SQL: SELECT msg WHERE msg.tag MATCHES "OGO" AND msg.user MATCHES "Council Member"). Dependent on the tags, messages could also be grouped so a query could pull up not just one member's email messages, but a whole group (psuedo-SQL: SELECT msgs WHERE msg.tag MATCHES "OGO" AND msg.group MATCHES "city council" WHERE msg.date MATCHES "last week"). The nice thing about this approach is that by tagging the messages with metadata as they come in, we can do efficient ad-hoc querying from the Zope server to the message store while utilizing a single data repository. Furthering enhancing system performance, Zope will cache any queries on the Zope server so multiple queries for the same data will not need multiple "trips" to the message store.

I am having trouble finding where the council feels that the amendment is " ... limit(ing) the ability of citizens to keep private the details of these communications;". The closest section I can find is the waiver of rights in section 4(B). That section does not deal with the public-at-large though; it only deals with those individuals and businesses that are seeking economic benefits from the city and does not deal with private citizens email. I'm at a loss for where the council found that clause applicable under the "real time" section.

The next section from the ballot language:

b) to require that the heads of all city departments, including the police department, parks department, library department, all city manager s staff and all city council members and their staff post online in real time information about all meetings and phone calls with private citizens;

And the relevant section(s) from the amendment:

(B)OPEN ACCESS TO CITY CALENDARS

1. For all matters involving City business, the following people must maintain calendars of all meetings and maintain logs of all telephone calls: (a) City Councilmembers and their staff; (b) City Manager and his or her staff; (c) Assistant City Managers and their staff; and (d) all department heads.
2. These calendars and logs must contain the time, date, subject matter, and persons involved in all meetings and telephone calls involving City business. These calendars must be used to schedule and record all past and future meetings that occur after the implementation date of this section.
3. Calendars and logs must be posted online in real time and be accessible to the public.
4. “Meetings” includes all informal and formal meetings including but not limited to telephone conferences, videoconferences, happy hours, and luncheons.
5. This provision must be implemented within six months of approval of this amendment.

In the proposed model, this requirement is already addressed in the current product line. Zimbra has partnered with another open source product, Asterisk to provide VoIP and PBX functionality within the collaboration server. Asterisk is a full-featured, enterprise-class PBX so the logging of calls in or out should be easy and display of those calls should be the same as the calendar view (more information is here and here.)

The next section from the ballot language:

c) prohibit the city from exercising state law protection for information that could expose the city and taxpayers to greater financial and legal liability and risk;

There are no technical issues required within this clause but, for completeness, here are the relevant amendment sections:

SECTION 4: Public Information.

The term “public information” means information that is required to be produced under Texas Government Code § 552.021. Public information also includes the following categories that must be produced in response to a public information request:

1. INFORMATION RELATING TO CIVIL LITIGATION. That the City is a party to litigation does not render information relating to that litigation less important; rather it often means the information is a matter of heightened public interest. Therefore, the City must not withhold information relating to civil litigation under Texas Government Code § 552.103, but it may withhold under other Public Information Act exceptions.
2. ECONOMIC DEVELOPMENT INFORMATION. Information relating to economic development assistance or incentives is public information to which the public has a right of access.
   1. The City must require all businesses and individuals seeking to engage in the type of economic development negotiations referenced in Texas Government Code § 552.131 to execute and deliver to the City a waiver of any rights to prevent the public disclosure of all information exchanged with the City. The City is without authority to engage in economic development negotiations with any company that has not first executed a waiver.
   2. The City is without authority to shield economic development offers under Texas Government Code § 552.131(b).
   3. Nothing in section 4(B) prevents a City from withholding documents under Texas Government Code §§ 552.104, 552.105, or 552.108.
3. AGENCY MEMORANDA. Open government in Austin ensures the people have access not only to the final decisions made by government officials but also to the process by which those decisions are made. The City must not use Texas Government Code § 552.111 to withhold information reflecting advice, opinion, and recommendations on policymaking matters, except the City may withhold attorney work product.
4. PERSONNEL FILES. The City of Austin must not maintain an optional personnel file as authorized under Texas Local Government Code § 143.089(g) for employees of the Austin Police Department, nor does the City have authority to enter into any meet and confer or other agreement with any police officer association that requires creation or maintenance of a separate file that is closed to the public.
5. EMAILS RELATED TO CITY BUSINESS. Email or other written electronic communication to or from a public official concerning City business is public information, including communications to or from privately owned email accounts or computers.

I'll address the next two ballot language sections together:

4. to require the city to create at taxpayer expense an online electronic data system for most city communications and documents, which for the most part are already available to the public; and
5. to install and permanently operate such a system at an estimated cost of approximately $36 million initially and $12 million annually thereafter if fully implemented, which could require a tax increase equivalent to three cents per $100 valuation or a reduction in city services?

The proposed model does not require any new systems to be purchased outside of the one's listed in my first post. Jabber serves as a "bridge" to connect the various existing datasources into the display architecture while the "agents" speak the "native" application language to the existing datastores (once again, I'll mention the CAPWin project of Washington, DC.) Some of the cost in the City's official estimate is for software licensing ($6mil alone just for the document management system). As I've stated in my past post, the cost estimate from the city is a good attempt at "herding cats" but until specifics are laid out, noone actually knows how much the initiative will cost.

For example, the model I propose is heavier into programming than software licensing. Most of the products I've used in this model do not have an initial or ongoing software licensing cost (however, service contracts will still need to be procured). So it can be assumed that this model will be less expensive in software licensing than the city's official estimate. I have yet to fully breakdown the initial and ongoing costs for this model (I'm more concerned about the technical issues and not cost; I leave that for people more skilled than I at cost estimation) but a quick rundown of the products listed and their software costs:

Cost estimation
Product	Initial Software Cost	Ongoing Licensing Cost	Service Contract required
Zope	$0	$0	Yes
Plone	$0	$0	Yes
Jabber	$0	$0	Yes
Jabber agents	$Unknown. Based on how many systems need to be tied in.	$0	Yes
Zimbra	$28/user	$0	Yes
OpenOffice.org	$0	$0	Yes

Please keep in mind that the service contracts will cost money as well as hardware would need to be purchased (the city's estimate on server hardware is fair and reflective; the city would not have to replace 1,540 PCs under this model saving an additional $1.2mil). Without delving further into specifics, between just EDIMS licensing and the replacement PC costs, there is a savings of $7.1mil off the projected one-time cost of $24mil and at least $1.1mil off the recurring cost of $11.6mil. (I feel the ballot language is wrong when it states $36mil upfront costs. That figure combines the one-time cost at $24mil and the recurring cost of $11.6mil. The recurring cost would not have an effect on the one-time purchasing cost of the first year. Again, if anyone is interested in fully fleshing the costs of this model out, please email me.)

So, in sum, I hope that the technical model I propose shows the flexibility and expandability sought by the OGO amendment to provide a transparent window into the people's government. I would like to re-emphasize that this model should *not* be taken as a roadmap but merely as a discussion of the salient technical issues the amendment seeks.