In the April 98 issue of Chips we looked at some current issues and trends associated with electronic records in the Federal government. In this article, we will broaden our scope a bit and look at the processes and mechanisms that generate electronic records and why computer-mediated enterprise information management (EIM) will become increasingly important to us.
There are four basic elements to EIM: correspondence management, workflow management, document management and records management. While these terms are often associated with various computer software applications, in principle they are all technology independent. We will do them with whatever technology is at hand because they are essential to running our organizations.
Each of the four elements contributes a unique piece to the EIM process and should be treated as modular components of a complete system. A common analogy is how we buy and upgrade stereo equipment. You can buy a separate tuner, control amplifier and power amplifier, or you can buy an integrated receiver that contains all three in one box. The separates may cost more than the receiver, but since each component specializes in doing one thing very well, your system's overall performance will probably be better.
Modular components are also usually less costly to upgrade. Let's say we get a new radio frequency spectrum allocated for public broadcasting in addition to AM and FM. With a receiver, if you want to tune in, you have to replace the entire receiver. With separates, all you need to replace is the tuner. You'll pay less money for the upgrade and probably get a better product.
We might upgrade stereo components every five to eight years. Software and hardware, however, are another matter. Given the frequency of upgrades in the world of technology, maintenance and upgrade costs frequently end up being more significant over the life of a system than the initial fixed capital investment. Buying modular components that we can mix, match and replace piecemeal instead of one- size-fits-all single applications for our computer-mediated EIMs will probably be our best long-term option.
Let's start from the back end of the process. All of the EIM components are important, but the driver for the entire system must be our legal requirements to maintain and preserve access to Federal record material. If the rest of the system works perfectly but we can't retrieve critical information or produce documents in court that meet the Rules of Evidence, we have failed.
To put it another way: the record is the end product of the system in much the same way that a car is the end product of an automaker. The company may be perfectly happy with its manufacturing process, but if the cars are lousy they'll ultimately go out of business. Congress and the public may not care a lot about the process we use to produce records. However, if we can't retrieve them upon demand, they may get somewhat annoyed.
The two basic questions records management attempts to answer are: How long must we keep something, and what happens to it when the time limit expires? The concepts associated with those questions are retention and disposition. Federal records have a retention period, normally specified by law, during which we must preserve both the record and access to that record. At the end of the retention period we either destroy the record or ship it to another location for archiving.
The National Archives and Records Administration (NARA) sets policy for all records management activity in the Federal government. As a general rule, long-term records have a retention of nine years or more. Some examples in the Air Force are listed in the table below:
| Record Set | Consisting of... | Disposition |
| Permanent Orders | Original copies of G-series special orders and MO, movement orders | Transfer to the National Archives in 10-year blocks when latest document is 25 years old |
| Decorations to Individuals (Military and Civilian) Approved Peacetime | Case files of recommendations, decisions, awards announcements, board meeting minutes and related documents | Destroy after 25 years |
| Record Set of Each Publication at HQ USAF/MAJCOM's/ Major Subordinate Commands | Directives, instructions, manuals, supplements, staff digests, pamphlets, which have AF-wide or MAJCOM applicability | Retire as permanent |
All the military services have similar requirements for these types of documents, and each of the record sets listed is eventually shipped to an archive of some type from everyone in DoD.
Much of the impetus behind the push for a paperless office is to reduce the high overhead associated with preparing, coordinating and archiving these documents. However, there is a major area of concern associated with taking these records digital: accessibility.
Will a file containing a plan or publication prepared with today's word processor du jour technology still be readable twenty years from now? Will CD-ROMs written by today's red laser technology still be readable twenty years from now by blue laser (or whatever may come after that) CD-ROM systems?
At present, NARA has placed restrictions on the types of digital formats it will accept for long-term archiving. They include ASCII text, Standard Generalized Markup Language (SGML) tagged text and some database formats. If we intend to have our EIM systems produce electronic records, they must produce formats acceptable for long-term archiving and retrieval.
Records with a retention of eight years or less are usually maintained locally or regionally in a staging area. Since NARA isn't normally involved with them, we can conceivably use whatever digital formats we want. However, given the fact that we're upgrading office automation software, on average, every two years, even short-term records may generate potentially serious retrieval problems if we don't preserve access to them.
The 88th Communications Group at Wright-Patterson AFB, Ohio, recently built an electronic records management prototype system. It is a trailblazing project and a harbinger of what the rest of us will eventually have to deal with in this arena. One facet of their program was to build what they refer to as their museum, a room with a shrink-wrapped copy of every piece of software they've used to create electronic record material and a contemporary computer (still in its original box) on which to run the software.
It may seem like overkill to some people, but they're doing precisely what they need to preserve access to digital records in their original form. I suppose they could batch convert old MS Word 2.0 files to MS Word 98, but it is unlikely the converted files would retain all the characteristics of the original file. Certain metadata and formatting will not translate, which means you no longer have a true copy of the original. Therefore, someone may need to preserve the technology, in addition to the information.
Finding information is a prerequisite to using it. All the information in the world is worthless if you can't find it. That's where document management applications come into play in EIM.
Much of what we consider part of traditional records management in the Federal government falls under the concept of document management. The various systems we use to index, track, find and organize discrete collections of information all fall into this category as we do them, whether or not the material we're managing is actually record material with a legally specified retention and disposition.
I make this distinction because document management software and records management software can be separated based on their ability to perform different access and retention functions. Records management software may have some document management capability, but a dedicated document management application will probably do the job better.
Windows Explorer, a file management application, is an example of a simple, computer-mediated document management system. File managers have rudimentary indexing and search capabilities that include sorting files by name, type, or date, and linear text searching of filenames or file contents. Electronic mail and groupware programs that let you build complex folder sets for storing e-mail, like MS Exchange and Lotus Notes, also fall into this category.
However, those applications are usually, in practice, managed individually, not as organizational resources. Document management in an EIM should provide and control access to the sum total of an organization's information holdings as a centrally managed system. This is not, however, a trivial task.
From a management standpoint, it means identifying the information we use and categorizing it by organization, function, process, use, or some combination of these and other factors. It also includes knowing how much information your organization stores, handles and transmits, and the relative importance of each information stream that flows through the organization.
From the technical side it means providing the technology to support business use of that information, including network services, storage media, user account management, bandwidth and connectivity. This can be difficult enough when information sits in one place. It becomes somewhat more complex when we start moving it around.
Information is not a static resource. It flows through organizations in much the same way that water flows through the global ecosystem. Small drops of information combine to become rivulets which merge into streams and large rivers, droplets of which are extracted to start the cycle over and over again. A single piece of data, like a single atom of water, may circumnavigate the globe many times during its life.
Workflow management includes all the processes we use to channel information through our organizations, including staff summary sheets, suspense tracking systems, chain of custody receipts and all the other things we do to control the transmission and delivery of information in the course of business. As with any large, complex system, we must break the sum of our activities into more manageable chunks if we expect to make any effective changes.
Defining an individual workflow process is akin to driving two stakes into a stream bed one hundred yards apart. We may then measure the size of the channel and the volume and composition of the flow, be it water or information.
If we have some control over the flow of our work, we can change the characteristics of the channel to speed or impede the flow. Events or conditions downstream may help our flow move faster or back up the flow and turn our little stream into a full-sized lake. Changes upstream can reduce or increase the volume we have to deal with and what types of things flow through.
And each piece of the total system may have some effect on every other piece. It's an intimidating conceptual model.
We expend a tremendous amount of manual activity managing paper-based workflow processes, which is why the paperless office has been a bureaucratic Holy Grail for the last ten years. We are, however, actually at a point where our technology can support paperless processes, particularly in the area of computer-mediated workflow.
The main benefit of a computer-mediated workflow is that the system takes care of all the delivery and audit functions now performed manually. Organizations that have successfully embraced workflow have reduced some process times by a factor of ten or more. Actions that used to take thirty days now take three days because all time information used to spend in transit has been wiped out by electronic delivery.
But only part of the effect is technical. There are also social pressures. We all probably know people who are chokepoints because something spends three days in their inboxes before they look at it. (We may even be some of them.) However, when the workflow database will tell their boss or their customer exactly how long they sat on something before taking action, it tends to motivate people.
First-generation computer workflow systems were based primarily on databases. The earliest ones were part of the mainframe computing environment, and versions or copies of those applications were eventually transported to the client/server network and personal computer environment. Workflow processes were programmed in, and a document management application of some type gave people access to their assignments, though some later versions use electronic mail as a delivery system.
A weakness of these early systems is that they had no ability for ad hoc assignment of work. If the workflow path didn't already exist in the system, individual users couldn't just build their own flows.
Then came the second generation of e-mail-based workflow engines. In addition to the predetermined, structured workflow allowed by older systems, newer applications give users the ability to build their own assignment/suspense flows as they need them. This more closely parallels how we actually work in the real world, which means we can effectively eliminate more manual inefficiencies via automated systems.
Some of the best workflow engines allow users to draw workflow diagrams by dragging and dropping icons representing various organizations, people and activities which the system then translates into an automated process. While these applications may cost more initially, if they are significantly easier to use we'll most likely see a much faster and greater return on our investment over its life cycle.
You get what you pay for.
The New Webster's Dictionary of the English Language defines correspondence as, "Communications by means of letters." An integral part of any letter is the signature, which serves to identify the author and authenticate the letter. However, we use signatures for many other things besides letters, particularly official forms and documents that may not all be communications between correspondents.
If we're routing a document that will eventually become a Federal electronic record via computer-mediated workflow, authentication will be our most significant issue over the life of that document. While I normally don't quibble with Mr. Webster, for the purposes of this discussion I would like to expand the concept of correspondence to include all properly authenticated transmissions of information between people.
By properly authenticated, I really mean signed. A staff summary or action officer processing form may not look like a letter, but it fulfills the same role with a few extra functions thrown in. The same applies to forms that people use to request or authorize various things. In the digital environment, digital signatures will be essential for EIM systems, as they are the guarantor of integrity.
Signatures are important at all points in the process. During workflow, they authenticate agreement or disagreement with the action being worked. As part of document and records management, they assure readers that the document is authentic.
In the electronic world, digital signatures can give us a higher degree of reliability and integrity than pen and ink signatures on paper. However, they may also complicate some processes if we're not aware of how they work.
Letters are a fairly simple model. They are normally only signed by one person, so authenticating them can be as simple as using the author's private key in the signature, which can be verified by everyone who receives the letter using that person's corresponding public key. This can apply to either e-mail or documents authenticated by a single authority.
Documents with more than one signature are more of a challenge, particularly when the signatures are applied sequentially as part of a coordination process. For example, minute changes to a document, like correcting a spelling error, can invalidate previously applied signatures. Authentication algorithms can't, at present, differentiate between changing content or fixing a typo.
However, given the relative speed with which a document may be coordinated electronically, re-authenticating a corrected document shouldn't take too much time if we design the workflow process around the technical constraints imposed by digital signature technology. It's just a matter of establishing new work habits.
Here are a few cultural issues to consider about EIMs as we move towards employing all this marvelous technology:
1. They tend to enforce honesty. No more fudging about when we got the package, what action we took and where we sent it. Big Brother, in the form of the workflow engine, is always watching. This may make some people quite nervous.
2. Reading from a monitor and reading print on paper have two entirely different comfort levels. As much as I like computers, I'm not likely to curl up with my laptop in bed and read the latest Tom Clancy novel off the screen. It's also much easier for most people to make pencil annotations in a margin than it is for them to use the annotation function in MS Word. I have adapted enough over the last six years that I can work comfortably in either medium, but your mileage may vary.
3. Speeding up a process does not always help. Nitrous oxide will let your car run 150+ miles per hour for a while, but then the engine burns out. Having one office produce staff packages ten times faster won't help if the next office in line can't process them at that rate. It's similar to the application upgrade problem where sixteen people get the new word processing application ahead of everyone else and proceed to send unreadable files to the other 2,084. Changes to the EIM should happen for everyone at once, not piecemeal over three years.
4. Speeding up a bad process just makes it worse. Some level of process analysis and reengineering is absolutely essential for anything you intend to automate.
5. EIM systems are strategic investments, as they will affect the conduct of business throughout every part of an organization. There must be a single, clear vision of the desired end result of implementing an EIM, and it must be understood, internalized and articulated at the highest levels of leadership. Bear in mind that every minute spent coordinating, handling or delivering something is one less minute spent on operations, logistics, security or other mission-related activities. Reducing non-value-added administrative overhead will increase the overall effectiveness of any organization.
If you want to talk with someone who is breaking ground in this area, I highly recommend you contact Guy Sawyer at the 88th Communications Group, Wright- Patterson AFB, OH. His group has documented some priceless lessons learned during their pilot project, which has been very successful.
There are, of course, some people who say this will never really work. For them, I submit the following:
"The concept is interesting and well-formed, but in order to earn better than a "C," the idea must be feasible."
The quote above is attributed to a Yale University management professor in response to Fred Smith's paper proposing reliable overnight delivery service. Smith went on to found Federal Express Corporation. So much for that feasibility model.
About the Author: Maj Dale Long, USAF