I spoke at the Gilbane conference on content management technologies in government on Wednesday. It was a good conference experience, and I hope they do it again next year. There were about 50 attendees, so it had the personal intimate feel of a retreat. The attendee list seemed to be balanced 50/50 between feds and consultants. There were three plenaries and eight pairs of sessions in two tracks. Almost all the sessions were case studies, which was amazing. The conference was in the Reagan Building downtown, which is just fabulous. The food was good, although the space for lunch was a bit odd. There were about fifteen vendor tables from a range of companies. Paying the bills and so forth. Tony Byrne of CMSWatch was the conference chair, and he definitely did a good job selecting speakers.
I spoke about Content Management and the Federal Enterprise Architecture. I got good feedback and a number of interesting questions. My favorite was whether the Business Reference Model (a government-wide taxonomy of business functions) is appropriate for the public as a means of organizing government content. My answer was ‘probably not’ since the BRM is meant to be a back office artifact used for planning. It uses language that is inappropriate for the general public. Maybe the general organization scheme might be useful even if the specific labels aren’t. But that asks the question – is there a taxonomy of the whole federal government’s operations that is appropriate for the general public? I believe OMB or GAO maintains a taxonomy intended for librarians (I need to do more research obviously). But librarians are not the general public, and taxonomies for libraries may or may not be appropriate for the web. Firstgov has gone down that road a bit, but they have struggled – the problem is very difficult. (Full disclosure – the original question was from someone who works at Firstgov).
I took copious notes in seven sessions, which I’ve included verbatim after the break…
Gilbane Notes
- James Melzer – FEA and ECM (these are my notes for the presentation, created in Freemind – the actual presentation is here)
- what is FEA?
- IT planning tool
- business-driven – place IT under control of business decisions and executive leadership
- IT performance measurement – improve effectiveness and efficiency of gov’t
- budget planning
- identify redundancies
- gap analysis
- opportunities for collaboration between agencies
- made of 5 reference models
- what’s a reference model?
- Performance Reference Model (PRM)
- Business Reference Model (BRM)
- Service Component Reference Model (SRM)
- Technical Reference Model (TRM)
- Data Reference Model (DRM)
- how does ECM fit into the FEA?
- no content reference model
- (unstructured) content supported in DRM, at the bottom of the stack
- content services might be supported in SRM
- how does the FEA help with ECM?
- BRM functions taxonomy
- intro to the BRM
- foundation of the whole FEA
- high-level taxonomy of all govt functions
- functional NOT organizaitonal
- designed to compare govt functions across agencies
- and to provide a business-based anchor for all the other models
- not appropriate granularity for workflow
- not (necessarily) appropriate lingo for agency public affairs
- Lines of Business
- mission
-
from a citizen’s perspective, what does the agency do? (e.g. EPA manages environment, FCC regulates spectrum, State conducts foreign affairs)
- resource mgmt
- common res. mgmt. functions for all orgs (e.g. HR or Financial Mgmt)
- mission support
- agency-specific support functions (like inspector general)
- mode of delivery
- how is the mission delivered (e.g. grants to states or direct federal services?)
- Examples
- Services to Citizens
- Environmental Management
- Environmental Monitoring and Forecasting
- Environmental Remediation
- Site and Area Evaluation and Cleanup
- Clean Up Contaminated Land
- Site Assessment
- Sampling and Analysis
- Remedial Investigation/Feasibility Study
- Community Involvement
- Remedy Selection
- Remedial Action
- Operations and Maintenance
- Perform Removals
- Federal Facilities Restoration
- Pollution Prevention and Control
- Disaster Management
- Disaster Preparedness and Planning
- Emergency Response
- Natural Disaster Response
- Anthropogenic Spills and Incidents Response
- International Affairs and Commerce
- Law Enforcement
- Mode of Delivery
- Regulatory Compliance and Enforcement
- Federal Financial Assistance
- Support Delivery of Services
- Controls and Oversight
- Internal Risk Management and Mitigation
- Legislative Relations
- Planning and Resource Allocation
- Public Affairs
- Regulatory Development
- Applied Research and Science Support
- Legal Services
- Management of Government Resources
- Administrative Management
- Financial Management
- Human Resource Management
- Information and Technology Management
- Supply Chain Management
- DRM data exchange model
-
common framework to describe, categorize, and share data
Data Description: Provides a means to uniformly describe data, thereby supporting its discovery and sharing.
- data discovery
- data reuse
- data sharing
- data harmonization
- semantic interoperability
Data Context: Facilitates discovery of data through an approach to the categorization of data according to taxonomies.
- Communities of Interest agree on taxonomies for categorization
- Internal and external COI
- Explicit connections to the BRM
Data Sharing: Supports the access and exchange of data where access consists of ad- hoc requests (such as a query of a data asset), and exchange consists of fixed, re- occurring transactions between parties. Enabled by capabilities provided by both the Data Context and Data Description standardization areas
- SRM services framework
- a taxonomy of IT service components
- organized horzontally across busines functions
- promotes reuse and collaboration of IT service components across governement
- PRM IT systems performance
- a taxonomy of performance measures
- Common language to describe outputs and measures
- comparative resource allocation – which programs are most effective and efficient?
- line of sight visibility between inputs and outputs
- Paul Fontain, FAA – iECM
- requirements at FAA
- trying to integrate all their disparate ECM functions
- work with E-Gov initiatives
- records management
- FOIA
- E-Government Act
- Presidential directives
- federal enterprise architecture
- Federal Ent. Arch. Framework vs. DoD EAF
- president’s management agenda
- citizen-centric, results-oriented, market-based
- stovepipes at FAA
- each stovepipe had its own tightly integrated stack of applications
- but no cross-stovepipe communication/integration
- OMB rotated the stacks – now they are integrated federal-wide, but there is no vertical integration
- 3 approaches
- standardization
- pick one product, more everybody to it
- advantages
- cheaper
- simpler contracting
- simpler support
- simpler infrastructure
- disadvantages
- everyone compromises
- no single product meets all requirements
- migration of existing stuff
- vendor dependence
- change management
- external interfaces
- not an option for cross-enterprise solutions
- integration
- build point-to-point integration between apps
- advantages
- disadvantages
- high development cost
- proprietary nature of enterprise application integration apps
- Nx integrations
- interoperability
- any app can speak to any other app or its info anywhere on the network
- advantages
- vendor independence
- interface with any system
- optimize external ISP options
- minimal integration cost/complexity
- minimize program lifecycle costs
- minimize implementation costs
- new functionality possibilities
- disadvantages
- requires coordination of multiple stakeholders
- not real yet
- iECM – interoperable Enterprise Content Management
- worked with AIIM, since they’ve got all the vendors as members
- goals
- goal: global access (not just federal gov’t)
- goal: increasing effectiveness of IT investments
- goal: better service to citizen/customer
- framework – requirements in, solution out – a blackbox sort of thing
- activities
- will develop a Technical Reference Model
- identify other related standards
- identify/develop best practices
- facilitate engagement of stakeholders
- pilot implementations
- content management services
- component descriptions
- content management services
- domain and data model services
- model of ‘semantic search continuum’ from TopQuadrant
- technical deliverables
- reference model
- best practices
- web services instantiation
- how does iECM apply to you
- Federal Web Governance Best Practices
- Lisa Welchman
-
web governance: how decisions and policies are made and implemented (content, data, applications)
web governance: policy-making and editorial+technical standards to meet the strategics objectives of the organization
3 manifestations of web governance
- ad hoc
- no implemented standards
- silo’d approach web development
- poor user experience (graphical diversiy and difficult navigation)
- diverse un-integrated technical infrastructure
- informal
- well understood web policies and standards
- some policies undocumented and unevenly implemented
- user exerience okay
- little ability to support whole-scale changes
- a few site-wide infrastructure apps and tools (e.g. CMS) but inconsistent success
- formal
- fully documented and implemented web policies and standards
- web site managed holistically as a product
- measured effectiveness
- high quality user experience
- well architected technical infrastructure with well-implemented enterprise tools
- appropriate meta-structures (content model, taxonomy, etc.)
- Jonda Byrd, EPA
- intro to EPA.GOV
- 3 million visitors
- 74.8 million page requests
- 730k pages
- 386 staff and contractors authorized to post to the web site
- 2 years ago
- no formal structure or representation
- CIO’s office ran the site and infrastructure
- grass roots web work group
- mgmt roles not defined
- OPA takes interest in the web site -> tension with CIO’s ofc
- web governance task force established
- today
- formal governance structure
- web council
- nat’l web content manager
- nat’l web infrastructure manager
- content and infr. coordinators from every program
- OPA – content
- OEI – infrastructure
- 46 people on this council
- how it is going
-
formal structure instituted by EPA Administrator
many informal policies collected into a huge omnibus formal policy, approved by web council and agency policy-making board
web standards, approved by council
- what didn’t work
- web council meeting failed
- contractors sent out to figure out what went wrong
- interviews with council members
- council members didn’t feel empowered
- council didn’t go back to programs and evangelize for decisions
- what worked
- required decisions, not just rubber-stamp
- make members work together
- meeting is in conjunction with larger web workgroup to make the two distinct
- Renee Truillo Lockhart, SSA
-
about SSA.GOV
- since 1994
- 50k pages
- 109 web masters
- 2.2 million transactions
why did they need governance
- succession planning
- multiple OMB mandates
- no formal structure or governance
- CMS – can’t automate business processes that aren’t documented (or don’t exist)
areas of governance
- web content
- web applications
- web strategy
working group created a web steeting committee and gave them a mandate (and created a temporary governance sub-cmte)
- document procedures
- escalation procedures
- roles and responsibilities
- formalize and publish procedures
- implement standards and infrastructure
- recommend web enforcement policies
- disseminate standards
- train on procedures
identify all web work groups
- identify
- roles and responsibilities
- overlap/redundancy identification
- escalation procedures
- establish how related to web working group
- establish communications
lessons learned
- executive buy-in across the enterprise is critical to success
- create a sense of urgency; state the business imperative
- promote the concept that Content Management is more than a system – it’s governance
- document policies, standards and controls; assign roles and responsibility for accountability
- CMS should leverage governance, not create it
- Sheila question
- staffing changes
- staff performance measures
- PDs for content-specific roles
- lisa – not a lot more content-specific job postings
- right skills
- Open-Source CMS in the Federal Sector
- Seth Gottleib, Optaros
- is it still a classic build vs. buy decision?
- in either case, you buy things AND build things
- x2 – x4 license cost in customization, minimum
- plus databases and other infrastructure tools
- when to use a open source CMS
- when you have a business process that is common and well solved by an open source product
- lower cost of basic purchase to focus funding and effort on customization
- want to standardize on a platform (for many coordinated non-competing groups)
- CM use cases
- informational website
- examples
- corporate identity
- departmental intranet site
- online information resource
- requirements
- page centric design
- compliance
- ease of use/WYSIWYG
- products
- Magnolia
- tree of pages
- in-content editing
- rigid page architecture – no reuse
- no workflow (?)
- online periodical
- examples
- ezine
- online version of a print magazine or newspaper
- requirements
- new pages all the time
- workflow (roles)
- flexible presentation tier
- high traffic
- products
- bricolage
- MacWorld
- workflow based on ‘desks’
- apache lenya
- midgard
- zope CMF
- boston globe
- san diego union tribune
- collaborative workspace
- examples
- project team space
- document repository
- requirements
- very efficient UI – must compete w email
- security
- versioning (but people often don’t get it)
- desktop integration
- products
- document management
- Alfresco
- documentum people
- adding web content soon
- Knowledge Tree
- wikis
- hybrid
- CPS
- collab tool
- doc mgmt cmty
- Open sTeam
- Plone
- weak versioning model
- strong cmty
- online community
- slashdot
- community of practice
- platform for custom development
- why
- you just want the platform anyway
- investing in a proprietary product is risky
- adherence to open standards is critical
- scalable solution
- framework
- buy the common pieces
- build the unique stuff
- figure out how much you can buy
- don’t make the system do something it can’t/shouldn’t
- government of New Zealand
- based on Plone
- fully compliant with ALL their standards
- make a local copy and customize
- local look and feel in CSS
- multilingual supported natively in Plone
- community of feds run this version
- larger Plone community supports basic technology
- how to be successful
- identify a community – active and aligned with your needs
- contribute back to cmty anything you don’t want to maintain solely (use the cmty)
- control scope
- only customize to solve unique problems
- have people live with the defaults first, before planning customizations (pilot and get feedback)
- select right support options
- focus on non-technology factors that really drive success
- good content
- training
- information architecture
- invest money saved on technology for these other things
- W. Page Glennie, Department of the Navy
- background
- office supports head of acquisitions for Dept of the Navy
- policy writing
- facilitate and improve the acquisition system
- 180k civilians, military, industry use their policy
- $60B in major acquisitions
- problem
- create a new public web site
- reduce operating costs by 60%
- maintain 10 years worth of policy
- decrease time for updates (minutes)
- multiple non-technical personnel to maintain the content (no contractor support for content)
- implementation
- approach
- market research analysis
- contractor
- evaluated open source and COTS
- evaluation criteria
- functional requirments spec
- govt usage
- dev support (community, user groups)
- available documentation
- selection
- eZ Publish
- 95% of manditory requirements
- free
- topical sections with own managers
- easy interface for non-technical content creators
- large community of support
- results
- acquisition.navy.mil
- saving $400k per year
- 2.5 years of implementation
- vendor support down from 4 people to 1.5
- people maintain their own content
- Making RM Invisible
- Tim Sprehe
- taking RM out of user hands
- …as much as possible
- devise ways to make records capture in the background
- business rules for capture (user role, nature of LoB, nature of document)
- workflow analysis – know when decisions made, what record is needed to record the decision
- Carol Brock, GAO
- intro to GAO
- document mgmt early adopter
- 1200 blue books (investigation reports) annually
- what they do
- audits/investigations
- administration
- policy
- project background
- requirements
- easy to use
- almost invisible to end users
- DoD 5015.2
- had a document mgmt solution, with no records functions
- solution overview
- based on document profiles (metadata)
- records decisions in workflow
- invisible to users
- sits on top of doc mgmt solution – enables retention decision
- a single retention descision with three options
- mission
- administration
- policy
- and the system remembers the user’s default profile
- rather risk having too much than too little, so over capture
- 3 retention schedules, pilot with NARA for this kind of solution
- intentionally overly simple
- 5 years for all mission recs
- permanent policy and special collections
- 7 years for all administration
- not automatic destruction – they ask the owners, espcially about holds like litigation
- separate category for ‘knowledge base’ with 10 year ‘reference’ retention
- simple file plans
- not mimicing paper filing
- each major bucket has a different (single) organization scheme
- functions
- operations
- jobs/projects
- user decision to save draft as record
- global read access to metadata, but docs may be secured
- personal profile(s)
- more than one per person
- one per project/person
- project metadata built in, included on every document created
- creation IN CONTEXT
- document locks
- entire project bucket locks all at once
- so, documents not individually locked
- and, all saves are to the record bucket, but it’s open until the whole thing locks
- 30 days to change things
- then completely locked
- but can open and reuse documents (saved as new version)
- hummingbird is vendor
- 26 suscessful pilots
- includes interns and contractors
- $425k per year on backlog reduction
- 104 users
- pilot objectives
- access user acceptance
- identify burden on staff
- identify best practices
- measure processes
- evaluate perceptions
- lessons learned
- big buckets take some getting used to
- you can’t do ERM without EDM
- must be inherent part of business system
- onsite vendor support needed
- lots of IT support (she’s KM)
- next steps
- portals for projects
- workflow
- collaboration
- Denise Bedford, World Bank
- problem statement
-
volume of info increasing
rm requirements growing
no strong approach for RM of electronic records
not an end of lifecycle function – must be full lifecycle
paper-based approach insufficient
dissemination of electronic tools to staff causes serious risk of records lost and unauthorized publications
- vision
-
staff use templates within the context of a business process and task – they can’t just open up Word and start – they must start with the task and the template
metadata hidden in template, filled in by user as part of normal document creation
template knows where it belongs – a ‘best guess’, and knows how it should be managed
reuse of document parts automatically inherits original metadata
filing is synonymous with saving – it happens seamlessly and in the background
can infer/distinguish record copies from convenience copies
- realizing the vision
- institutionally comprehensive retention schedules
- functionally organized
- authority file of organization units
- metadata repository with links back to source system
- rules-based retention and disposition expert system
- management reporting tools
- absolute requirement that all systems meet minimum metadata compatibility rules
- content owner rentention actions ‘fail up’ to records managers as an escalation
- must have agreement on business process functions (like the BRM)
- security classification of individual metadata elements
- she has a slide with an ER diagram for this
- Washington Post – Large Web Publishing Systems
- Chris Contakes – WashingtonPost.com
- intro/team
- 10 java developers
- CMS
- dynamic apps
- washingtonpost.com
- budgettravelonline.com
- newsweek.com
- slate.com
- CMS solution
- washingtonpost.com
- budgettravelonline.com
- same CMS (really?)
- custom built solution
- largely automated
- uses terragram for autoclassification and indexing for search
- everything rendered statically in preview and FTPed to production
- template driven
- all text is plain text
- custom content entry interface – XML on the backend
- 130,000 publishing trips (FTP uploads) per day
- CMS challenges
- every content item has a ‘context’ element, so the system knows what site it applies to
- publish 5000 items per day
- about 10M pag views per day
- for dynamic applications – database apps are replicated out to a public DB
- so things like MyWashingtonPost are actually rendered from a static copy of the content
- Alec Dann – PostNewsweek Tech Media
- iProduction is vendor on CMS
- had 114 templates (many fewer now)
- XML-based now
- challenge of the future
- web design changes
- huge monitors, huge resolution
- various forms of syndication and presentation
- visual media
- migration of content is hard
- coding in the pages
- coding in the templates
- generalization
- page elements
- lists of things
- content elements
- composition tool for making templates, basically a template for templates
- internal use of RSS for exchange, stored search
- Sheila Campbell – Future of Federal Web
- about FirstGov.gov
- office of citizen services within GSA
- firstgov.gov
- citizen info
- web managers advisory council
- webcontent.gov
- best practices
- community
- training
- 2011
- ubiqitous internet – always on
- wireless/internet as a government service
- implications for content and format
- 10 predictions
- dramatically fewer government web sites and government web pages
- design debate will be over – basic layout will be common (for government)
- US Government sites will have a common look and feel
- No government website will be launched or continued without testing with the people who use it
- .gov naming convention will make sense to the public
- public will be able to perform the tasks they want on gov’t web sites
- websites will be used to accomplish agency missions
- leaders will ‘get the web’
- government search will improve dramatically
- government websites do not come up well in results on Google
- government websites are hard to search
- websites will promote greater interaction between people and their government
- e-democracy
- census online
- live help
- content will be aggregated to serve the public
- such as contact us. why find the agency first? have a federal-level contact us.
- community-level information based on your location
- personalized content for each citizen
- government web managers will be judged by whether citizens can find what they need
Great work, James! I’ll be sure to dig deeper into your slides.
As an aside, I find it interesting that the timing of the conference almost coincided with the release of Google’s U.S. Government Search.
Rob,
The new usgov.google.com was definitely a topic of conversation.
This is excellent content, James — congratulations on a great report! Was any scientific research content discussed? We’re quite invested in consolidating research data for scientists. Are there any tools out there specifically targeted on that?