# What Paper Charts Did Better

*Five functions of the paper record worth keeping, each stated as a requirement and a test*

Kanwar Partap Singh Gill, MD · November 7, 2026 · Becoming a Doctor

A title like this one asks for trouble, and I would rather name the trouble at the start. Any comparison between paper records and electronic ones can be rigged without anyone intending to rig it. Set the best paper workflow anyone can remember beside the worst electronic one anyone has endured, and paper wins. Set the paper folder that went missing for three days beside an electronic record that opened in a second on a phone in another city, and paper loses. Both comparisons are honest about their examples and dishonest about their selection. The essay on this site about [paper's failures as a system](../what-paper-charts-did-terribly/index.html) is the necessary companion to this one, and I would not want either read alone.

So this essay makes a narrower claim. Paper records, for certain specific tasks, had properties that helped their readers, and those properties can be described precisely enough to be built into any medium and tested. None of them is a property of paper as such. They are properties of a design that paper happened to impose, sometimes by accident, and that screens do not impose unless someone decides they should. My training in Amritsar, which ended with an MBBS from Sri Guru Ram Das Institute of Medical Sciences and Research in 2002, took place among records that were overwhelmingly paper; my residency in the United States, from 2007, was among electronic ones. I will not describe a remembered chart or a remembered preference here. Where I say that a task seems easier on paper, I mean that the evidence and the logic of the task suggest it, not that I can vouch for a particular afternoon.

## Compared to What?

The designer and statistician Edward Tufte has a three-word question that he places at the heart of quantitative reasoning: *Compared to what?* It is the right question for this essay too. Paper cannot be better in the abstract. It can be better at something, for someone, under conditions, and the claim means nothing until the something is named.

I will name two tasks, because they are the ones in which paper's design seems to me most instructive. The first is seeing the course of a measurement over time: whether a fever is settling, whether a weight is falling, whether a blood count is recovering. The second is comparing two documents: today's note with the admission note, this year's report with last year's, the plan as it was written with the plan as it now stands. Both are ordinary, frequent and consequential. Both are tasks in which a reader is not looking for a single fact but trying to see a relation between facts.

For other tasks the comparison runs the other way, and decisively. Finding a result from another hospital, reading an order in a stranger's handwriting, opening a record that is physically somewhere else: on these, paper was worse, sometimes dangerously so, and nobody should want those failures back. The point of naming tasks is to keep the argument from sliding between them, so that a virtue in one task is not borrowed to excuse a failure in another.

Consider a hypothetical clinician, on the third morning of a patient's admission, who needs to answer two plain questions before seeing her: is the fever settling, and has the plan changed since she came in? Imagine first a paper folder of the ordinary kind. The clinician opens it at the back, where the nursing chart hangs, and the temperature graph answers the first question at a glance: three evening spikes, each a little lower than the last. For the second question she turns to the front, pulls out the admission note, finds yesterday's progress note, and lays the two side by side on the desk. The original plan and the current one are both in view, and the difference, a changed antibiotic and a scan that has been dropped, is visible in a minute. All of this assumes that the folder is on the ward and not in radiology, that the nurse plotted the temperatures correctly, and that yesterday's note is legible, three assumptions paper often failed to honour.

Now imagine a screen of the kind many clinicians have used. The temperatures are available, but the default view is a table of numbers in rows by time, and the graph is a click or two away in another tab. The admission note and yesterday's note are both available, but each opens in the same pane, so that reading one closes the other, and the comparison must be made by reading the first, remembering it, and reading the second. Every fact is present and retrievable, and the folder cannot be lost. Yet both questions take longer to answer and depend more on the reader's memory than on her eyes. Nothing in this second scene is a failure of the electronic record as a medium. It is a failure of a particular arrangement, and other arrangements exist. That is exactly the point. The paper version was better at these two tasks because of how it was laid out, not because it was paper, and the layout can be borrowed.

It also matters what *better* would mean. In human-factors research it usually means that representative users, performing a defined task under realistic conditions, finish faster, make fewer errors, or judge their own accuracy more correctly. Those are measurable, and the measurement has been formalised. The National Institute of Standards and Technology has published empirically based use cases for testing the usability of electronic health records, which is to say that the machinery for testing claims of this kind exists. A lesson from paper is worth keeping only if it can be put in a form that such testing could confirm or refute. What follows is an attempt to put five of them in that form.

## The Line at the Foot of the Bed

Hospitals have not always drawn their patients' temperatures as lines. The practice owes a great deal to Carl Wunderlich, professor of medicine at Leipzig, who published a book in 1868 on the behaviour of body temperature in disease and who is credited with introducing temperature charts into hospitals. His thermometer was reportedly a foot long and took twenty minutes to register, and his work established both a number, thirty-seven degrees Celsius, as normal and a habit, the regular plotting of temperature on a graph. The number has not aged well. In 1992 Philip Mackowiak and colleagues measured oral temperatures in 148 healthy adults and found a mean of 36.8 degrees, with an upper limit of normal lower than Wunderlich's and a variation with the time of day. The habit has aged much better. A fever plotted as a line can be read across a room, and its shape, rising, falling, spiking every other evening, carries information that a column of numbers holds but does not show.

A novelist got there first, in his own way. In the sixth volume of *Tristram Shandy*, published in 1762, Laurence Sterne's narrator, who has spent most of five volumes digressing, announces that he is now beginning to get fairly into his work and hopes to go on "in a tolerable straight line." He then draws, on the page, the lines he has actually moved in through his first four volumes, four wandering scrawls, and the line of his fifth, with loops lettered for each digression: the trip to Navarre, the airing with the Lady Baussiere, the round led by the devils. The small curves marked *c* are "nothing but parentheses, and the common ins and outs incident to the lives of the greatest ministers of state." Finally he draws the line he aspires to, ruled "by a writing-master's ruler (borrowed for that purpose), turning neither to the right hand or to the left." It is a joke about narrative, and it is also a graph of a story: the shape of a long, digressive text made visible at a glance, in a way that the text itself, read page by page, could never make visible.

An illness is never Sterne's ruled line, and a good record lets its reader see the loops. The paper temperature chart did this by necessity. Its graph sheet hung in a fixed place, at a fixed scale, and anyone passing could see the shape without asking for it. The lesson is not that paper drew better graphs. Electronic records can graph a decade of values in seconds, and that is a real advance. The lesson concerns what the reader sees first. On paper the shape was the default view. On a screen it is often a table, with the graph one or two actions away, and an option not taken is a graph not seen.

Stated as a requirement: for any measurement that is repeated over a stay or a series of visits, the default display should show its course as a shape, on a stable scale, without the reader having to request it. Stated as a test: give clinicians a set of records and ask whether a measurement is improving, worsening or unchanged; compare the time taken and the error rate when the default view is a graph with the same when it is a table. If the graph does not help, the lesson is wrong and should be dropped.

## A Page Stays Where It Was

The second property is spatial. In a paper folder, a document stayed where it had been put. The admission note was at the front of its section, the orders on their own coloured sheet, the nursing chart at the back, the results pasted in a known place. Readers who used the same kind of folder every day learned its geography and moved through it partly by memory of position: the sheet with the ruled columns, the page with the graph, the thin slips from the laboratory.

There is evidence that this was more than a comfortable habit. In 1992 E. Nygren and P. Henriksson, at the Center for Human-Computer Studies in Uppsala, interviewed physicians about how they used the medical record in everyday practice. They concluded that the record was a well-functioning working instrument for the experienced physician, and that reading it as a basis for decisions involved interpreting its format, layout and other textural features, which in Sweden were the features of typewritten pages. Those features guided searching, reading and judging the relevance of items, and the authors described the skill of using them as an integrated part of diagnostic expertise. In a companion paper they then did something instructive: they designed a screen interface that imitated paper pages, organised in bundles linked to index lists, with page-turning as its fundamental action, in order to let readers use the perceptual skills they already had.

A study from McGill published in 2000 by Vimla Patel and colleagues looked at the same question from the other side. Comparing paper and computer-based records in a diabetes clinic, they found that the paper records had a narrative structure while the electronic records were organised into discrete items, and that exposure to the electronic record was associated with changes in how physicians gathered information and reasoned. The structure of the record even shaped the conversation with the patient. The authors called records of this kind cognitive artefacts, objects that shape how the people using them obtain, organise and reason with knowledge.

Read together, the two studies make a point that is easy to lose in arguments about paper. The layout of a record is not decoration. It is part of how readers think with it, and a skilled reader's expertise is partly expertise in a particular layout. That cuts both ways. It means that a stable, learnable layout is a real asset, which paper provided by default because a printed form does not rearrange itself. It also means that some of the attachment clinicians felt to paper was attachment to a layout they had already learned, and that any new layout, however good, would at first seem worse. Preference and performance have to be separated, and only testing separates them.

Stated as a requirement: each kind of information should have a stable location and a recognisable visual form, so that a reader can tell what a document is before reading it, and both should survive software updates and differ as little as possible between screens. Stated as a test: measure how long experienced users take to find a given item, and how often they find the wrong one, before and after a layout changes; and measure whether they can identify a document's type from its appearance alone.

## Two Pages Side by Side

The third property is the one I would most like to see preserved, because it is the easiest to lose. Paper allowed two documents to lie side by side. A reader could open the folder at the admission note, pull out today's progress note, and lay them next to each other on the desk. A report from last year could sit beside this year's. Two sets of results could be held up together. Nothing was required but a desk and two hands.

Tufte's name for the general principle is the small multiple: a series of similar graphs or tables drawn at the same scale and with the same axes, so that the eye can compare them directly. He argues that comparisons of change, of difference and of alternatives are made best when the display enforces them visually. The research on paper in offices points in the same direction. Abigail Sellen and Richard Harper, in *The Myth of the Paperless Office*, published in 2001, argued from studies of knowledge workers that paper's persistence in offices was no failure of nerve. Paper supports particular activities well, among them reading across several documents at once, annotating while reading, and arranging documents in space so that their relations can be seen.

Comparison on a screen often works differently. Many systems show one document at a time, or one in a narrow pane beside a list, and a reader comparing two notes opens the first, holds what it said in memory, and opens the second. Comparison then depends on recall, which is precisely the faculty that an [essay on medicine before search](../when-medicine-was-mostly-memory/index.html) argues should not be asked to carry what can be put in front of the eyes. A change in the plan between Tuesday and Thursday, a value that has moved from one report to the next, a sentence that was in the admission note and is absent from the discharge summary: all are easier to see when both versions are visible together and harder when one of them must be remembered.

Perceptual psychology suggests why sequence is the harder condition. In a well-known experiment published in 1997, Ronald Rensink, Kevin O'Regan and James Clark showed observers a picture, then a brief blank screen, then the same picture with something changed, cycling between them, and found that people could look at the alternating images for a long time without seeing even large changes. The blank interval, acting much as an eye movement does, was enough to wipe out the cue that would otherwise draw attention to the altered spot, and only focused attention on the right part of the scene revealed the change. The phenomenon, now called change blindness, has been replicated in many forms. Viewing two versions of a note one after the other, with a screen refresh between them, reproduces the structure of that experiment almost exactly, except that the interval is longer and the reader has other things on her mind. Two documents lying side by side allow the eye to move back and forth between corresponding places, which is a different and easier task, and a display that marks the differences between two versions does the attentional work for her.

Stated as a requirement: a reader should be able to display any two dated documents, or any two sets of results, side by side at the same scale, with differences between versions of the same text marked. Stated as a test: give clinicians pairs of notes in which something clinically relevant has changed, and measure how often and how quickly they detect the change when the two are displayed together and when they must be viewed in sequence.

## The Edge of the Page

The fourth property is the most ambiguous, and I want to give its failures their due. A paper form had edges. The box for the presenting complaint was a certain size; the space for the plan was a certain number of lines; a progress note written by hand was limited by the writer's patience and the width of the sheet. The edge forced selection. A writer who had three lines had to decide what mattered most, and the reader of those three lines received the decision.

The counterexamples are obvious and real. Writers who had more to say than the box allowed wrote smaller, until the writing became unreadable, or wrote over the margin, or continued on another sheet that could become separated from the first. Some things were left out because there was no room, not because they did not matter. A constraint that forces selection also forces omission, and the omitted item may be the important one. Electronic notes removed the edge, and with it both the discipline and the cramping; the consequence, as [another essay here](../the-note-that-grew-longer-and-said-less/index.html) describes, has been notes that grew longer and said less.

What is worth keeping is the existence, somewhere in the record, of a place where selection is required: a summary of fixed and visible size, at the top of the note or the record, in which the writer must say what matters most now. Paper provided this by accident and imperfectly. A screen can provide it deliberately and enforce it gently, while leaving the full detail available beneath.

Stated as a requirement: every note should open with an assessment and plan of bounded length, visibly separate from the detail that supports it. Stated as a test: measure how long readers take to find the current assessment and plan, and how often they misidentify it, in notes with and without such a bounded summary; and check that the bound does not raise the rate at which clinically important items are omitted.

## A Name at the End of Every Entry

The fifth property concerns authorship. On a paper chart every entry ended with a signature or initials, every entry carried a date and usually a time, and every entry was in a hand. A reader could see, at the point of reading, who had written each part of the record and when. [The essay on the handwritten chart](../the-handwritten-medical-chart/index.html) explains why recognising a hand is not the same as verifying an entry, and I will not repeat that argument. The point here is simpler. On paper, the author and the time were attached to the text, visibly, at the place where the text was read.

Electronic records know far more about authorship than paper ever did. They log who created each entry and when, and often who changed it. But much of that knowledge sits behind the text, in metadata that a reader must request, and the visible page may show a single author at the top of a note that contains sentences carried forward from other people's notes on other days. Another essay on this site follows [a sentence copied through six documents](../a-record-repeated-is-not-a-record-corroborated/index.html) and asks what its repetition can prove. The paper chart's advantage was not better information about authorship. It was better placement of the information it had.

Stated as a requirement: the author and date of every statement a reader sees should be visible where the statement is read, or available in a single action, including for text carried forward from earlier entries. Stated as a test: point to a sentence in a note and ask a clinician who wrote it and when; measure how often the answer is correct and how long it takes.

There is a sixth candidate, the visible gap, the hole in a paper record that announced itself and so invited a question. I leave it to [the essay devoted to it](../the-honest-gap/index.html), which examines when an empty space in a record is honest and when it is not.

## Lessons That Do Not Survive

Not every virtue claimed for paper deserves a place on the list, and part of keeping the real lessons honest is discarding the counterfeit ones. Several are commonly offered and seem to me not to survive examination.

The first is that paper made doctors think. The idea is that writing by hand, slowly, forced a clinician to compose, and that composition is a form of reasoning which typing or selecting from menus has eroded. There may be something in it, and there are studies on handwriting and learning in other settings, but I know of no good evidence that clinicians writing by hand reasoned better about patients than clinicians typing, and plenty of handwritten notes were as formulaic as any template. Slowness is not thought. A lesson that cannot name the mechanism by which paper improved reasoning, and cannot propose a test of it, belongs with nostalgia.

The second is that handwritten records were more humane, because the hand is personal. The essay on the handwritten chart argues that handwriting made authors visible, which is a real and transferable property, but that the personal look of ink also lent inferences an authority they had not earned. Humane records are records that describe people accurately and fairly. That is a property of what is written, and it can be achieved or failed in any medium.

The third is that paper was simpler, with fewer clicks and fewer screens. For a single reader at a single bedside it often was. But the simplicity was purchased by others: by the nurse who plotted every temperature by hand, the clerk who filed every slip, the porter who carried the folder, the colleague who could not see the chart because it was here. Counting the clicks of one reader against the whole invisible labour of a paper system is the selective comparison this essay began by refusing.

The fourth is that paper was more private. The companion essay on [the record's failures](../what-paper-charts-did-terribly/index.html) shows how much exposure paper's obscurity concealed. Privacy is a practice in any medium, not a property of one.

What remains when these are discarded is smaller and more useful. The five functions above have mechanisms that can be named: a shape seen before numbers are read, a location learned and relied upon, two views held together rather than one remembered, a forced choice of what matters, an author attached to each sentence. Each mechanism predicts something a test could measure. That is the difference between a lesson and a longing.

## Writing and Weighing

Punjabi has an old compound for accounts that I find useful here. ਲੇਖਾ (*lekhā*) is a written account, a reckoning; ਜੋਖਾ (*jokhā*) comes from ਜੋਖਣਾ (*jokhṇā*), to weigh. To go over the ਲੇਖਾ-ਜੋਖਾ (*lekhā-jokhā*) of something is to review its accounts, and the phrase keeps the two acts together: what has been written down, and what it weighs. A merchant's book earns its keep less by recording every transaction than by letting its keeper see, at the end of the season, what the transactions add up to.

Most of what paper did well for the two tasks I named belongs to the second half of that phrase. A graph at the foot of the bed, a page that stays where it was, two sheets laid side by side, a box that forces a summary, a name at the end of each entry: these are aids to weighing, to orientation, to seeing what the record amounts to. They are not aids to retrieval. Paper was poor at retrieval, and the great promise of electronic records was retrieval: any result, any note, from anywhere, at once. That promise has largely been kept. What was not always noticed is that weighing does not come free with retrieval. A record can make every fact instantly findable and still make it hard to see what the facts mean together. Paper gave its readers certain aids to weighing because its physical form forced them. A screen gives them only if someone designs them in.

The difference shows most clearly in the first minute a reader spends with a record. A clinician meeting a patient's folder for the first time, before reading anything, took in a good deal: how thick it was, which sections were full and which thin, whether the temperature graph was calm or jagged, how many hands had written in the last few days. None of that was reliable on its own, and the essays on [the handwritten chart](../the-handwritten-medical-chart/index.html) and on [the smell of old charts](../the-smell-of-an-old-hospital-chart/index.html) explain how such impressions mislead. But it was orientation, given before reading began, and it told the reader where to look first. A record that opens on a list of tabs, all of the same size and colour, gives no such orientation. It waits to be asked. The work of weighing then begins only after the work of retrieving, and on a busy morning it may not begin at all.

## Testing Without a Laboratory

Each lesson above ends in a test, and it would be easy to read the tests as gestures toward research that someone else will one day do. They are meant more practically than that. Most of them can be run by a single clinic or hospital department, with modest means, and the habit of running them would do more to improve records than any amount of argument about paper.

What such a test needs is simple to list. It needs a small set of records, synthetic or properly de-identified, so that no patient's privacy is spent on the exercise. It needs a handful of the people who actually use the records, chosen to include those for whom the record is hardest: the new resident, the locum who has never seen this system, the nurse at the end of a night shift. It needs a task stated as a question with a right answer, such as whether this value is rising, what changed in the plan between these two dates, or who wrote this sentence. And it needs two measurements, the time taken and whether the answer was right, recorded for each version of the display being compared.

People who study the usability of software have long held that small tests, repeated as a design changes, find a large share of a system's problems, and that a few users struggling with a task in front of an observer teach more than a survey of many users' opinions. Small tests have limits. They can show that a display confuses people; they cannot establish, with the confidence a large study gives, that one design is better than another for every user in every setting. But the claims in this essay are modest enough that small tests are the right first step. If five clinicians answer the fever question faster and more accurately from a graph than from a table, the lesson has earned a place. If they do not, it has not, and no amount of fondness for the paper chart should keep it.

There is one further advantage to testing, which is that it changes the conversation. Arguments between clinicians who miss paper and designers who never used it tend to be arguments between memories and intentions. A stopwatch and a list of right answers give both sides something to look at.

## What the Screen Does Better

None of this is an argument for paper, and fairness requires the other column of the account.

On the tasks this essay has named, a well-designed electronic record can outdo the best paper chart. It can draw the graph automatically, at any scale, across years, and lay several graphs beside one another. It can display two notes side by side and mark every difference between them, something no desk could do. It can show the author and time of every sentence more reliably than handwriting ever did, because it does not depend on a reader's recognition. It can enforce a bounded summary without cramping anyone's writing. And it can do all of this for a reader in another building at three in the morning, while another reader looks at the same record in the clinic, which paper could never do for anyone. It is also legible, which on its own has prevented a great deal of harm.

So the comparison, made honestly, is not paper against screen. It is a set of functions, some of which paper provided by default and screens provide only by design, against the designs actually in use. Where those designs already provide the functions, there is nothing to learn from paper except that the functions matter. Where they do not, paper is useful as a demonstration that the functions were possible with very simple means, and as a reminder of what readers lost when they went missing.

There is one more caution, and it applies to me as much as to anyone. A clinician who trained on paper may simply prefer it, and preference dressed as evidence is the most persuasive form of nostalgia, because it is sincere. Nygren and Henriksson's finding that layout reading is part of expertise explains such preferences without justifying them: an expert in one layout will feel slower in another for a while, whatever their relative merits. That is why each lesson above ends in a test. If a test shows that a function does not help, it should be dropped, however fondly it is remembered.

## Lessons, Not Longing

Put plainly, the five lessons are these:

1. Show repeated measurements as a shape by default, on a stable scale.
2. Give each kind of information a stable place and a recognisable form.
3. Let any two documents or result sets be seen side by side, with differences marked.
4. Open every note with a bounded summary of what matters now.
5. Keep the author and date of every statement visible where it is read.

Each can be specified, built and tested, in any medium. None requires paper, and none requires a reader to have loved paper. They are what remains of the paper chart when the smell, the weight and the affection are taken away, and I think they are worth more than all three.

Sterne, at the end of his chapter of lines, notes that the straight line is praised by divines as the path for Christians, by Cicero as the emblem of moral rectitude, by cabbage planters as the best line, and by Archimedes as the shortest between two points. He had already conceded that his own story had not yet run that way. Neither does any illness run straight, and a record that pretended otherwise would be lying in its layout. What paper sometimes did well was to let the curves show: the fever that spiked on alternate evenings, the plan that changed between two pages laid side by side, the name at the foot of each entry marking where one person's view ended and another's began. Those are the lessons worth keeping. The ruler can be left where Sterne found it, borrowed for the purpose and returned.

## Evidence and Editorial Note

**Documented findings.** Public biography relied on: MBBS, Sri Guru Ram Das Institute of Medical Sciences and Research, Amritsar, 2002; PGY-1 at the University of Arkansas for Medical Sciences, Little Rock, from 2007; the author's published account, in the flagship essay on this site, that his medical-school training took place among predominantly paper records and that electronic records were already a normal part of clinical work at UAMS in 2007. Edward Tufte's question "Compared to what?" and the small multiple as same-scale, same-axis displays enforcing comparison: Wikipedia, "Small multiple," citing Tufte, *Envisioning Information* (Graphics Press, 1990), p. 67. National Institute of Standards and Technology, NIST IR 7804-1, *Technical Evaluation, Testing, and Validation of the Usability of Electronic Health Records: Empirically Based Use Cases for Validating Safety-Enhanced Usability and Guidelines for Standardization* (2015), nist.gov. Carl Reinhold August Wunderlich, *Das Verhalten der Eigenwärme in Krankheiten* (1868), professor at Leipzig, introduction of temperature charts into hospitals, the reportedly foot-long thermometer requiring twenty minutes, and 37 °C as mean normal temperature: Wikipedia, "Carl Reinhold August Wunderlich." Mackowiak PA, Wasserman SS, Levine MM, "A critical appraisal of 98.6°F, the upper limit of the normal body temperature, and other legacies of Carl Reinhold August Wunderlich," *JAMA* 1992;268:1578–1580 (PubMed abstract, PMID 1302471): 148 healthy adults, mean oral temperature 36.8 °C, lower upper limit of normal, diurnal variation. Laurence Sterne, *The Life and Opinions of Tristram Shandy, Gentleman*, volume VI (1762), the chapter of lines, quoted from the Project Gutenberg text. Nygren E, Henriksson P, "Reading the medical record. I. Analysis of physicians' ways of reading the medical record," *Computer Methods and Programs in Biomedicine* 1992;39:1–12 (PubMed abstract, PMID 1302665), and Nygren E, Johnson M, Henriksson P, "Reading the medical record. II. Design of a human-computer interface for basic reading of computerized medical records," same issue, 13–25 (PubMed abstract, PMID 1302668). Patel VL, Kushniruk AW, Yang S, Yale JF, "Impact of a computer-based patient record system on data collection, knowledge organization, and reasoning," *Journal of the American Medical Informatics Association* 2000;7:569–585 (PubMed abstract, PMID 11062231). Abigail Sellen and Richard Harper, *The Myth of the Paperless Office* (MIT Press, 2001), its publication details from Wikipedia, "Paperless office"; the summary of its argument about paper's support for cross-document reading, annotation and spatial arrangement is standard reference knowledge of the book. Change blindness and the flicker experiment of Rensink RA, O'Regan JK, Clark JJ, "To see or not to see: the need for attention to perceive changes in scenes," *Psychological Science* 1997;8:368–373: Wikipedia, "Change blindness," and the study's citation there. ਲੇਖਾ, ਜੋਖਣਾ and ਲੇਖਾ-ਜੋਖਾ: standard reference knowledge of Punjabi usage. The usability-testing view that small, repeated tests with a few users find many problems but do not establish superiority, and the existence of studies of handwriting and learning outside clinical settings: standard reference knowledge, stated without figures. The general capabilities of electronic records described (automatic graphing, side-by-side display, version comparison, authorship logging, simultaneous remote access): standard reference knowledge, stated in general terms. **Vendor representations.** None. No electronic record product is named or assessed.

**Recollection and tradition.** The author's training dates and the predominance of paper in it are his published biography. No chart, ward, task, preference or patient is recalled; the essay states no personal experience of a task being easier on paper. The descriptions of folder layout, temperature charts and side-by-side reading are general, and the clinician on the third morning, the two displays she uses, and the tasks and tests are hypothetical constructions introduced as such. Nothing is asserted about the records of Sri Guru Ram Das Institute or any other particular hospital.

**Analysis and forecast.** The choice of comparison tasks, the five functions and their statement as requirements and tests, the reading of Sterne's lines as a graph of a story, the distinction between retrieval and weighing drawn from ਲੇਖਾ-ਜੋਖਾ, the application of change blindness to sequential note viewing, the rejection of four commonly claimed virtues of paper, and the judgement that paper's useful properties were properties of design rather than of the medium are the author's interpretation as of November 6, 2026. The proposed tests are suggestions for evaluation, not reported results. **Independence.** This essay is not sponsored by any institution, company, health system or technology developer. It is personal reflection, not medical advice; readers seeking care should contact their own physician's office.



## Related reading

- [The Smell of an Old Hospital Chart](https://personal.kpsgill.com/essays/the-smell-of-an-old-hospital-chart/)
- [What Paper Charts Did Terribly](https://personal.kpsgill.com/essays/what-paper-charts-did-terribly/)
- [The Handwritten Medical Chart](https://personal.kpsgill.com/essays/the-handwritten-medical-chart/)