grouperlab

Get, share, and use information well

Category: Delay

Inertial Damping

I have been thinking a lot about damping lately.  To be specific, inertial damping.  You know, that aspect of your hybrid electric car’s regenerative braking system that recharges your batteries while you stop?  Or the gyroscopic properties of a bicycle wheel that keeps you moving forward instead of falling over when you turn?  Yeah, that stuff.  People think about that all the time, don’t they?  No?

This is what happens when a geek starts talking about their internal thought processes, and especially since I recently talked about postcards from the future, maybe I do need to explain my terms.  Apparently, one of the first things I need to do is to explain that I am not talking about “inertial dampening,” which, as far as I can tell, is a science fiction plot device highly likely to get yourself into a fight with physicists for dissing their man Newton.    That’s not my focus today.  I’m trying to take a real physics and engineering term, and see how application of that term in a complex human setting helps me design, analyze, or improve sociotechnical systems in a more effective way… because that’s something that engineers do.

Actually, I started thinking about damping a lot based on a question that someone asked me at the end of the Jefferson Science Fellowship (JSF) lecture I gave on January 24.  (For reference, the point of this lecture is to summarize the general area of work that each Fellow does, both for the policy audience of the State Department and Agency for International Development, and for the scientific audience of the National Academies.)  I spoke about information flow and distributed expertise (because that’s something I do), including the challenges of appropriate coordination during event response for either physical (civil unrest, natural disaster) or cyber-physical (network or security operations) events.  I got quite a few questions, as well as invitations for additional discussions with various groups across the State Dept.  This was a very good feeling, in that it gave me the sense that some people could finally hear some of what I have been trying to study and communicate for years.

However, that does come with a price: when one of those people asks a question, can I give an answer that they understand and know what to do with it?  In essence, that was the challenge when someone asked me a damping question.  (They didn’t really ask it as a damping question, but since I am likely to see lots of things as connected feedback control systems, it’s not surprising that I heard it as one.)

If you have a large bureaucratic organization which lives on sending lots of messages to lots of people for their opinion and approval (aka “clearance”), don’t you run the risk of taking too long to respond to emerging, critical timeline events?

That’s a very reasonable question.  And it takes me immediately to thoughts about damping.  Imagine your new event as some sort of input function.  However, the event isn’t always purely evident immediately, and it doesn’t just go from off to on instantaneously.  There might be multiple events that may or may not be related to each other.  You want your response (output function) to match the demands of the input function.  The engineering version of this problem is one of “critical damping”.  If your damping ratio is too high (over 1), your response to the new event is very slow.  Although you may never over-respond to the event, it takes you a long time to actually respond to the event, and in fact, you may fail to do what needs to be done within the deadlines (people need fresh water and shelter and warmth within a matter of hours to days, or they die).  We tend to assume that faster is always better.  However, there is a limit / problem with that, which we now understand from the world of social media.  Someone can respond *too quickly* with *too little* information, and be unable to tell the difference between the actual event that needs to be responded, and some distractor or misinformation.  (Remember, I’m not trying to be political here, but since the lecture was just a few days after the Inauguration, I may have made a reference to a social media event or two.)  This would be an example of having a damping ratio that is too low (close to 0), which is a different problem.  (You might ask what is the inertial property here.  Well, I have talked in the past about knowledge as “little inertial balls of expertise,” in the sense that expertise allows you to devote energy to efficient processing of the world and move to where you need to go in the future.)  People going off on their first impression without checking sources or others’ understanding would be an “underdamped” response (damping ratio too low), which can be just as bad (but in a different way ) than a bureaucratic, “overdamped” response (damping ratio too high) that takes too long and doesn’t want to risk or challenge anything for fear of being wrong.

In essence, an effective inertial damper takes energy that comes at you, with bounces and noise and possible confusion that you don’t want to respond too much to, and turn it into energy that works for you in a time frame that makes for the tasks you need to do.  That sounds great, and it’s a very interesting problem to work on.  Perhaps the additional challenge is, How do I apply this to my own life?  As much as I enjoy a string of fist-pumping, high-fiving successes in a non-athletic context,   there is the challenge of appropriately damped responses when shifting from State Dept. to Purdue stuff.  Reminder to Barrett: it’s not good to try to do two full-time jobs simultaneously for long periods, and I am feeling now the stress of trying to complete a large number of Purdue (or Indiana Space Grant) activities after spending all day working on Japan Desk activities.  In fact, that stress might be better described as hysteresis, rather than damping.  (Discuss.)  More accurately, damping is the ability to take the frustration of emails and news feed updates and channel that energy into productive work, such as a book chapter, or journal manuscript, or even a blog entry.   Like this one.

Back to School

The Labor Day holiday weekend is drawing to a close, and I have finished up my second week of the Fellowship.  Even though the start dates of the semester and my tenure here in DC were the same, I have gotten to notice how much the routines differ between the two environments.  Unlike my academic routine that can adapt and adjust based on the day of the week and the differences between class and no-class, committee and research schedules, things feel distinct here.  There is a bus I catch, most days, between 8:14 and 8:40.  On Wednesdays, there will usually be lunch with the other Fellows.  There are Monday and Thursday morning “huddle” meetings.

 

However, that is not what I notice the most from the past two weeks.  I admit that I have developed a particular appreciation for my manager.  Each day, there is a specific new thing I have to learn.  How do I send a particular type of email?  What is the formatting for this kind of documentation?  Who do I contact for this activity?  Of course, he’s seen this all before, but it’s my first time.  And it’s not like I have had 3-4 weeks of easing into the situation.  I’ve already worked on international memoranda, and meetings between embassy staff and local representatives, and sat in on planning discussions with the offices of some folks whose name might appear on someone’s bumper sticker.  (But notably, the importance of the office is communicated by an acronym, or even a single letter; the people whose names are used are names I don’t recognize, and even those names go with acronyms.)  The most appropriate phrase for this experience is one that I learned during my first few weeks as an undergrad at MIT: “Drinking from the firehose.”

 

In that environment, where I’m supposed to come up to speed quickly, it seems like a luxury to have someone check in with me as many as 3-5 times per day to help me with one task or another.  In truth, some of the help sessions seem a bit remedial, teaching me things I do already know.  But he doesn’t know that.  And more importantly, I don’t always know when something I think I know how to do isn’t exactly how this organization does it.  So, I find myself learning to be more patient when being taught, and listening all the way through the lesson.  I even have a guiding document for goals to achieve over the next month or so—distinct from a to-do list of tasks, and an in-process list of assignments.

 

One of the things that surprises me most about this firehose experience is a new-found empathy and appreciation for the situations that confront new students in the lab.  We’ve been working on SoS and PoSE conceptualizations of ICT use in the SHARK and DOLPHIN and PERCH* streams for years—why are you nodding blankly at me?  Of course.  I’ve been doing it for years.  You just got here.  I just used a bunch of acronyms—shorthand for me, incomprehensible jargon for you.  Even when we get to time for a thesis outline, or a prelim draft, or a set of PhD defense slides, it does take some reminders to recognize that two dozen years of practice and 75 or more iterations don’t get transmitted easily to someone who is experiencing it all new and in an intense, nervous state.

 

I would like to hope that this lesson comes back to Purdue with me next Fall.  For a new student, or new faculty member, each new item can be part of an overwhelming onslaught of novelty and complexity.  Maybe it won’t stay that way for long, but it feels like that now.  In the senior capstone design course I teach, I remind the students to take the time to capture those initial moments of novelty and first attempts at processing and decision making, because it will be really hard to recall those feelings (and assumptions, and senses of confusion) again later.  I can tell them that, but it was a long time since I have felt that at the level I feel it now.  It’s good to be reminded of what the first few, chaotic weeks of new experience feel like.

 

IMG_3685

Photo of Little Kern Golden Trout by Middleton and Liitschwager (1988), hanging in the C Street entrance lobby of the National Academies.

 

 

*Acronym decluttering:

SoS: Systems-of-Systems. or a description of complex systems engineering settings where individual components of an overarching system represent complex systems in their own right (such as individual aircraft, with pilots and co-pilots, in the airspace over Washington, DC while Marine One is traveling across town).

PoSE: Perspectives on Systems Engineering.  This is a course that I developed to teach about four distinct traditions of systems engineering, ranging across systems thinking, cybernetics, component-whole relations, and project management.  Only in its second iteration as a hybrid distance / on-campus course, it is one of the most subscribed courses in Engineering Professional Education (and I’m not even teaching it this semester).

ICT: Information and Communications Technology.  When I first started as a faculty member, most computers had line-by-line display screens in single colors of amber or green; email and word processors and bulletin board chat groups were the most sophisticated information exchange tools available.  Even with all of the changes in capability, it’s still important to recognize that the point of these technologies were, and are, for humans to communicate.

SHARK, DOLPHIN, PERCH:  These are designations of project areas within the research lab, referring to knowledge sharing architectures, information flow delays, and applications to healthcare delivery improvement, respectively.  Check them out at https://engineering.purdue.edu/GrouperLab/streams/.

Filling in the Blanks

How can it be that it’s been nearly 18 months since my last blog entry?  Well, I could wax philosophic, and point out that the path to such an outcome is like the path to other, more positive outcomes.  It’s an accumulation of daily habits, and a series of perhaps small, but sometimes very distinct, decisions.  So, a bit of a review of what’s been happening, and what lessons can be drawn from both the period of silence and what has filled that silence.

 

I’m a big fan of waiting for a big, dramatic highlight to emphasize in an announcement.  Back in November, 2014, I was applying for a campus-level directorship position; I was pretty excited about the opportunity, and the ways that I could use my skills to connect research, and STEM engagement, and educational improvements at K-12 and university levels.  I thought I was going to get the position.  I didn’t.  In retrospect, it’s not necessarily that I was a bad candidate for the job, but a bad match for the view (by others) of what the job needed.  This is actually an important distinction, and I am convinced that I had never actually seen the idea of not being selected for a position in that light before.   Well, a few weeks of anticipation were followed by days of anger and frustration, which in turned into a more circumspect view of job searches and candidate interviews no longer just being about showing that someone is “good enough” to be considered.  Imagine that all of the finalists may be “good enough,” in some generic sense, but every complex job is a combination of factors on a very large vector of possible criteria (utility), where different people involved in the selection (stakeholders) have different ideas of the importance (weights) of the criteria, and decide what “best” looks like (stakeholders maximizing their objective function according to their multi-attribute utility weighting).  I was a really good candidate for one version of the job.  I wasn’t the best candidate for another version of the job.  That doesn’t make me a good or bad candidate overall, and certainly not a bad person.  An important lesson to learn, but not one I was ready to write about in Spring 2015.

 

The lab was going through a significant shift in 2014-15, both conceptually and physically.  We spent the first half of calendar 2015 in Wang Hall, learning how to conduct a different type of meeting with a different configuration of students (three new, first year grad students with only four or five continuing students).  We’re back in Grissom Hall as of August 2015, but the only thing about the building that’s stayed the same on the inside is the walls and bricks and windows framing the building’s outer boundary.  And we’ve had to learn an even more interesting set of dynamics: we are now at a point where much of the lab’s activity officially qualifies as a distributed enterprise.  Dissertation-writing students are working in industry, and other doctoral students are doing co-ops, internships, and other work in multiple time zones.  Lab meetings and 1:1 individual interactions are more likely to occur in Google Hangouts than Grissom 335 (my new office) or the GROUPER dedicated lab space (which doesn’t exist).  So, we have had to learn new lessons about information alignment and distributed knowledge sharing.  That’s a topic for another entry, coming soon.

 

Believe it or not, the lesson learned about being a good candidate vs. a matching candidate for the job had to be taught to me again in 2015.  This time, the position was a campus administrative post, and again, I thought I was a very good match for a visionary leadership role in a broadly influential and interdisciplinary approach to the future of the campus.  Great, right?  Except that this objective function was apparently not aligned with the utility vector of critical stakeholders.    This is neither good nor bad, in itself.  (Remember what you just told them, Barrett.)  I do believe that the transition from anger to acknowledgement happened faster this time, and to be honest, it’s a lesson that does need a very strong reinforcement over multiple administrations for me to actually learn the meaning well.

 

Oh, there’s some outcome productivity in terms of field visits, and journal papers, and GROUPER degree completions.  However, I wouldn’t suggest scheduling MS thesis defenses by multiple students on consecutive days.  We succeeded last summer, and now the number of GROUPER MS thesis grads exceeds 30.  But I’m not likely to try that again soon—it’s a lot of reading, and a MS thesis is often as much a test and oral exam for the advisor as for the student.

 

In the end, I’m better off for it, and I think we in the lab have learned a number of very important and valuable lessons.  It can be dangerous if someone gets too much in the habit of doing without considering, or acting without accepting that both “success” and “failure” can be a benefit or blessing.  One of the challenging, and yet extremely beneficial, outcomes is that the two interviews required me to very explicitly consider the question of how to manage the lab, and in essence, examine what was an appropriate “carrying capacity” of GROUPER at this stage of my career.  (I’m probably more active than ever before, with GROUPER work and GROUPERs in 2015 supported by five federal agencies—AHRQ, FAA, NASA, NSF, VA; it’s not yet the “riding into the sunset” that I had previously considered.)  We’ve been practicing skills that I see in increasing frequency in industry, but not as much in academia—how to become easy and fluent with a team operating across geography, knowledge domain, and a variety of external constraints to be focused and robust to a variety of communication channel capabilities.

 

More coming soon.  I’m expecting a big announcement in a week or so.  No, really.

 

Delays and Synchrony

In our first post, we talked about how our research inspired us to start a GROUPER Blog to provide you with updates more frequently than our journal and conference publications.  We anticipate that the various projects going on in the world are moving faster than the year-long or half-year-long publication cycle permits us to get updates out to the designers.  In the previous post, Dr. Caldwell discussed the idea of people’s perception of waiting.  In this post, we will combine these two ideas with a discussion of waiting during information sharing.  We have talked about aligning the update frequencies of the design and our research output, and we have talked about what people feel when they wait.  Today, we will present our perspectives of a second timing concept, delay.

Delay, lag, latency, and lateness are all words that describe the fact that information does not arrive instantaneously to its intended receiver.  We can go into detail about why this is always true in another post, but to keep things brief we will limit today’s discussion to obvious cases of delay.  Consider cases where e-mails are sent across the Internet, voice mails are left for our family members, satellite news video feeds are transmitted up to orbit and back down on the other side of the world, and memos and documents are shared between team members.  All of these cases experience some large or small amount of delay between when the sender generates information and when the receiver gets that information.

When there is delay, people have to compensate so that they can still do their job well.  The degree to which they have to compensate depends on the medium of communication.  GROUPER classifies communication media two ways:  synchronous media and asynchronous media.  Synchronous media are those in which the sender and receiver wait for each other to send messages and receive messages.  Telephone conversations, chat-room messages, television and radio, and face-to-face communication are all synchronous.  The receiver can sometimes process the message as the sender is generating it, as in the case of a face-to-face message or telephone conversation.  Other times, the receiver must wait for the sender to compose and send the message, as in the chat-room message, but the receiver is still listening and waiting for the message.

In contrast, asynchronous media are those in which the senders do not wait for the receivers to get their message and respond.  This is because the senders can expect the delay in the communication to be longer than they are willing to sit there and wait.  E-mail, voice mail, postal mail, memos and documentation, and blogs are all asynchronous.  When dealing with asynchronous media, the sender can send a message and forget about it, moving on to other tasks.  The receiver can work on her tasks and forget about the sender until she decides to check for, or the medium notifies her of, a new message.

This is not going to turn into a physics / philosophy discussion about whether two events can actually be truly simultaneous.  Our focus is on whether the delay between events is meaningful compared to the time involved in the task requiring the information flow.  This depends on the task itself, and not just the total amount of delay.  If I’m listening to a stereo CD, or watching a video with audio track, a few tens of milliseconds makes a huge difference in my perceptual system’s experience of synchronous behavior.  However, if I send out documents before going on a week-long vacation “off the grid,” it doesn’t matter to me whether the responses take an hour or six days.  The GROUPER emphasis is on whether delay interrupts one’s cycles of task performance, with time spent waiting rather than doing.

On occasion, we experience a shift in the media, for example when sending an e-mail at work, we might receive a very quick reply.  We then reply quickly ourselves, and, for a short period, the e-mail medium becomes more like an instant-message (synchronous) medium.  However, GROUPER still classifies e-mail as asynchronous in general because at any moment the two communicators may leave the communication without the other being aware of it.  We’ve experienced both the unexpected reply to a late-night email, and the frustration and change in expectation when we realize that work time for us is either sleep or holiday time for someone else, and that is why they’re not available.

Neither synchronous media nor asynchronous media is always better than the other, but each have specific situations in which they are better and should be used.  For example, if the goal of the news channel is to deliver news to viewers faster than their competitors can, then they will want to forgo the option of editing the satellite news feeds from the reporter on-location and tolerate the delay in the live feed.  If the goal is to provide a company-wide notice of a new corporate policy, it may be best to not interrupt daily tasks, but to leave an e-mail for all employees for when they have a moment to really absorb it—and not require them to drop their tools and listen, for example.

The challenge that GROUPER has identified with regard to synchronous and asynchronous media is that, when communication has enough delay, it is not only disruptively long, but the receiver must nevertheless wait for the message to come through the medium before he can continue his work.  The receiver can also unknowingly continue to work in the environment with information that is either incorrect or incomplete.  For example, telemedicine can be both synchronous and asynchronous.  Telemedicine, or distance medicine, involves the use of technology to enable patients to communicate with their physicians and physicians to communicate with other healthcare professionals.  Telemedicine is synchronous when a patient communicates with his physician in real time via videoconferencing.  All of the information shared between the two parties is both timely and accurate and does not hinder the physician’s capacity to make informed decisions regarding the patient’s care—unless the patient is withholding certain information.  (See Vallette et al. (2011) for more information.)  Telemedicine can also be asynchronous, for example when two physicians, a primary care physician and a specialist, are communicating regarding a patient’s care and using email-like store-and-forward technology in order to share the patient’s medical history documents.  While using this technology allows the specialist to view the documents on her own time, the patient might come into the primary care physician’s office with an emergency, and the specialist might not see any change in the patient’s medical documents in time for the patient’s next visit.  Then she might make a diagnosis based on outdated information.  The specialist would never intend, of course, to make a misdiagnosis, but, unfortunately, in this case she may not know the whole story because of the nonroutine visit by the patient.  Without respect for the medical-document medium as an asynchronous medium, the specialist is more likely to forget that she may not always have the most up-to-date information.

Further reading:

Caldwell, B. S. (2008).  Knowledge sharing and expertise coordination of event response in organizations.  Applied Ergonomics 39, pp 427-438.

Vallette, M. A., Chafac, M. N., Benedict, A. J. & Caldwell, B. S. (2011). Reducing barriers to knowledge sharing among healthcare professionals and patients. Proceedings of the Industrial Engineering Research Conference. Reno, Nevada: Institute of Industrial Engineers, Inc.