Short-term memory and productivity: a theory of hold time

Over the past few months, I have observed some curious things about my workflow. They are not rigorously measured; in fact the only way I have noticed these is mainly by looking at the system clock, or wall clock. I don’t have the kind of free time that would allow me to really set up ways to verify this, but if there are any published papers on this phenomenon (or lack thereof) I would love a pointer to them.

  1. Simple tasks (sending emails, collating names) take about the same time both on the desktop at home, and on the laptop at work.
  2. Complex tasks (timetabling, collating multiple columns of student information, searching for images and videos) take longer on the work laptop tan on the desktop. And I mean sometimes as much as 3 or 4 times longer.
  3. I lose my train of thought at work much more easily, and hence fall out of a state of flow more frequently.

A simple explanation for this is that my internet connection is faster at home, my desktop is faster (though the effects of this are pretty negligible except for niche tasks like image processing), and my screen is bigger. But how does this lead to a three- or four-fold change in productivity? I don’t think speed differences in hardware alone can sufficiently account for this.

Observation #3 is particularly disturbing, because it has a dampening effect on my work morale. Not being able to sustain a state of flow readily leads to an unproductive rest-of-the-day, and I end up pushing a lot of work till I am home and have access to the desktop. There are parallels here to computer engineering: in a processor, things happen in cycles, which are paced and measured by clocks. The time taken by various tasks (addition, multiplication, reading and writing to memory) is measured in cycles, not in seconds, because a modern processor’s frequency is scaleable, which means a task that took 3 nanoseconds (ns) at 1 GHz would take 1 ns at 3 GHz. Things happen at the rise and fall of the clock edge, like steps in a waltz.

Between these edges, during the intervals, the state must remain steady ( or at least end up being the same at the end of the cycle). A voltage that goes high must remain high until the end of the cycle, when it can transfer its state; in a waltz the leading partner’s posture must be maintained until the next step; a fall must be quickly recovered from, lest it interrupts the dance! This steady-state interval is known as “hold time”.

There is no neurological evidence that the brain operates in a similar manner, with a “brain clock” regulating cycles, but my experience of Observation #3 echoes this idea of hold time. In a state of flow, if I focus too hard on memorising a 6-digit number (usually a 2-factor authentication PIN), I end up dropping other things that were sitting in my short-term memory. Often this number is going to be used within the next 5 seconds, and over time I have learned to tune the effort I expend to remember this number so that I don’t forget anything else in my head.

PINs don’t offer much trouble in the way of use though, because they usually arrive after the authentication page is done loading. But many other tasks which involve temporary holding and transfer of short snippets involve waiting for a page, application, or file to finish loading, and quite often by the time they do, I have forgotten the number—whether a mobile number, ID number, or some other value—and must go back to remember it anew. (Before you tell me to use the digital clipboard—there’s something else there already. I tend to hold many things in my head …)

So just spend longer remembering the number, what’s the problem? That’s a question I’ve been asking myself as well, but I think these things are not as consciously controlled as we imagine. Mental workflows take conscious effort to train, but subsequently become unconscious. How many productivity masters cognise every single shortcut or hotkey or keystroke that they use on a daily basis? The key to productivity is precisely not to think about the low-level tasks involved, otherwise one quickly trips over one’s feet like in the centipede’s dilemma. Having gotten used to a much faster workflow, I am trying to adapt myself to a slower one, for the sake of sanity, but … it goes poorly.

An enterprise-encumbered laptop, together with spotty wifi that deauthenticates me or drop the connection on a regular basis, does a lot more than just take more time to process data. The hold time required for such a workflow is a lot longer than I am used to working with my desktop at home, with the result that I work much more inefficiently, having to constantly adjust between two forms of workflow. Needless to say, I have inhabited my desktop at home much more thoroughly than I have my laptop at work.

Assuming this theory holds water, what does that mean for productivity-enhancing policies? Firstly, it means trying as much as possible to reduce mental hold time for bits of information. Even in the most tightly integrated operations, there is still a need for people to manually transfer some data, and the longer the hold time required, the longer the task will take, all else being equal. This means not being stingy with hardware upgrades that significantly improve system performance, such as solid-state disks.

More importantly it also means upgrading network systems, especially wireless ones; reducing user friction (time taken to get onto a network) and improving network latency. On bad days up to a third of my time is spent just waiting on or fiddling with the wifi.

There are ways of working around this, and I have had to resort to these often, but these workarounds all lead to a longer cycle length. I’m not sure there is a way to have my cake and eat it here, but I am constantly on the lookout for such promises.