New vs. good

Introduction

Here I intend to gather thoughts on topics like whether what is new is automatically better. (The eventual page could be much longer than the current.)

The topic has featured repeatedly in other parts of my writings, including a much older text on whether what is progressive is automatically better, as well as a number of texts from my Wordpress days (TODO import and link).

There is some overlap with regard to “better for the one, not for the other” between the discussions of announcements and milk cartons below. This is a side-effect of their being written independently with some time between. I have kept both, as they both are also good illustrations of the “new vs. good” theme, which, after all, is the point of this page.

Better for the one, not for the other

A particular complication is that some changes (“newers”) are for the better for some (especially, for the force behind the change) but not for others—and all-too-often this is not given due consideration when making decisions. In the other direction, such complications can make it hard for evaluators, e.g. in a text like this, to judge whether a specific change was a net-improvement or -worsening over the sum of all those affected.

A particular problem is that changes often seem geared at the most stupid, the least informed, whatnot—things are made better for the stupid; everyone else sees things grow worse. This, especially, in the world of software. (The situation is even worse for those of some brightness, as many or most of the remaining changes are geared at the broad masses in a one-size-fits-all manner.)

Another is changes made to favor the changer in a one-sided, ethical disputable, and/or destructive manner. Moreover, such changes often backfire, because side-effects, hidden costs, loss of costumers, etc., are not properly considered in advance.

Good examples can be found in the extreme excesses of loudspeaker announcements and similar in today’s world. For instance:

1. Various commercial messages in stores are an evil to or blended out by most customers, including myself. However, there might be a small minority of the easily manipulated or naive that either benefit or incorrectly believe that they benefit. (“Oh! Something is on offer! I must rush over! Can’t miss the opportunity!”)

   At the same time, the store in charge of these messages obviously perceives a benefit. Whether the perception is correct is another matter. Yes, the aforementioned minority might buy more, on average, but the majority might end up buying less. (I have been known to outright avoid stores with too many or too intrusive announcements, and/or where an attitude of too great contempt for the customers shines through.)

   ---

   Side-note:

   In terms of “new vs. good”, these were quite rare in my childhood, if in doubt because few stores had PA systems. They have, since, grown ever more prevalent and intrusive.

   (Over the last few years, I have the impression of an improvement. It might be that the dissatisfaction among customers grew too large or became noticed by decision makers, causing a partial change of course; it might be that coincidence and choices by individual stores have given me the wrong impression.)

   ---

2. Announcements in railway stations and in trains are, at least in Germany, a horror, through the combination of frequency and the proportion that is pointless. To boot, they are often exceptionally loud—presumably, to make life easier, in a noisy environment, for the minority that is hard of hearing. In some cases, as with the Cologne main station, this is taken to such excesses that the actual pain threshold is crossed for many who have normal hearing. (This potentially creates a vicious circle: the travellers’ hearing is damaged by the announcements and the loudness must be further increased to accommodate the increasing number of half-deaf travellers...)

   ---

   Side-note:

   In terms of “new vs. good”, the frequency was much lower in the past. I cannot give a time scale, but comparing 1997 (my first year in Germany) with 2007, there had definitely been a great increase. Further increases have taken place since, if in doubt because PA systems and whatnots have become more widespread, as older trains (often with no or only a poor PA system) have been replaced with newer trains, as even many smaller stations have been given PA systems, and as use of PA systems have been extended to other means of public transport, e.g. streetcars. To boot, there is a greater amount of automatism, e.g. in that many announcements are pre-recorded, which reduces the need to have onboard staff perform any wished-for announcement.

   ---

   To give an example of pointlessness and undue frequency: I have been on train rides with stops every five minutes, where, within these five minutes, something like the following played out as separate announcements, each applying to just a small minority of the passengers, each grabbing attention:

   1. Pointless greeting of the new passengers.

   2. Announcement of what the next station is.

   3. Announcement that this station will be reached in just a few minutes.

   4. Announcement that “we are now arriving” at this station.

      (Note how the second through fourth item could easily have been reduced to one.)

   5. Various cautions not to forget luggage, to be careful when exiting, and whatnot, that usually bring no value.

   Such a set of messages might have some benefit for very inexperienced travellers; however, the majority of travellers are experienced or outright commuters, who travel through the same stations every weekday. (And imagine hearing what amounts to a message a minute, every minute, for, say, twenty minutes in the morning and twenty minutes in the evening, day in and day out, while e.g. trying to read or to get some work done.)

   A particular issue is the “Various cautions”, where I suspect less of a genuine caution and more of a legal safe-guard for the train company—maybe, good for the train company; bad for the passengers.

   ---

   Side-note:

   A problem is that the human mind seems geared both at detecting changes in environmental sound and at reacting to human voices, which means that each such message grabs attention, disturbs a flow of thought, must be processed sufficiently far to determine whether something of importance is being said, etc. Alternatively, many passengers eventually become so jaded that they block out the messages entirely—and might then miss something actually important. (These are the lucky ones.)

   Also see below for how the same phenomenon once could be used in a positive manner when calling a hotline.

   ---

   ---

   Side-note:

   There is an interesting potential parallel between this type of “announcement bloat” and feature bloat in software. Just like some software companies seem to think that “we could have X as a feature; ergo, we should have X as a feature”, many with access to a PA system seem to think that “we could make announcements about X; ergo, we should make announcements about X”—in both cases, with little thought on whether the feature resp. announcement truly is beneficial or whether it is neutral or outright detrimental, whether it is something actually wanted by the customers, what the side-effects might be (e.g. a drop in maintainability resp. a growing annoyance among customers), and similar. Likewise, I cannot suppress the suspicion that some genuinely believe that “the more features/announcements we make, the better for the users/passengers”, again without considering factors like what is truly beneficial and what detrimental.

   ---

   Other absurdities include air-passengers-going-by-ICE-train, where there then can be two greetings, both indiscriminately pumped out through the entire train, both more wordy than is reasonable: one from the railway company to greet the regular passengers and another from the airline to greet the air passengers. Even if we assume that the air passengers had some benefit from this, no matter how dubious, the annoyance caused to the regular passengers, who far outnumber them, cannot be justified. (The exact reason for the presence of these air passengers is unclear to me. It might be that a greater group of travellers are collected in the one city to go to the airport of another, making this a first leg on a longer journey.)

3. In the days of yore, calling a hotline was reasonably doable: I was put in a queue, instrumental music played, I put the receiver on my desk and waited for a human voice. Once the voice came, I picked up the receiver again. At some point, however, the hotlines decided that the customer had to be continually reminded that he was “in the queue”, “was important”, “had just a few more minutes to wait”, or similar. As the difference between such a message and the voice of an actual human counterpart was much smaller than the difference between a voice and the music, my options were suddenly much more limited. At the same time, annoyances similar to train-travel now appeared, with an ever-recurring distraction from the task at hand.

   This is not just a good example of “new” being worse, but raises interesting questions about the “why”. It might be sheer incompetence (maybe, as a variation of the bloat mentioned in an earlier side-note), but it is also possible that some specific goal is pursued. Many seem to suspect that hotlines are deliberately set up to be as deterring as possible, and this would fit in well, as an additional annoyance. A personal suspicion is a wish to reduce the risk that customers forget the phone, briefly leave the phone unattended, or similar—a customer who does could cost the hotline, maybe, some twenty or thirty seconds of time at minimum wage. (In both cases, the hotline would gain a benefit at the cost of the callers by such messages.) Then again, it could be that there simply are some callers so stupid that they need the constant reminder, and for whose benefit everyone else takes a hit.

   ---

   Side-note:

   Even this was, of course, long ago, before cell phones became dominant. For a variety of reasons, I have not called a hotline in ages, including the discomforts of and time waste from queueing, the low competence of the typical counterpart, and the complete unreliability of claims made by the counterpart, where promises were made on the phone and later disavowed by the company in question.

   ---

Milk cartoons

Milk cartoons provide a number of examples of how what is new does not necessarily bring an improvement over the whole line, but in some areas at the cost of other areas, and might even be a net negative to some.

As far back as I can recall from my Swedish childhood, milk was always bought in one-liter cartons—and ones extremely similar to those that I, today, age 49, encounter when I buy milk in Germany. (Mostly, approximately brick-shaped. Some alternate shapes occur, but are rarer.)

Of the changes that have taken place, the most obvious is the development of how pouring is handled.

1. In the very early days, the customers were supposed to fold up a piece of the cartoon, take a pair of scissors, and cut out a spout. There were even printed markings for how to cut in the optimal way. Resealing the carton, for better durability of the milk, required more work or equipment than it was worth, and the carton was usually left unsealed in the refrigerator. (I have a vague recollection of some type of plastic “shoe” that could be put on the top of the carton for closure. Short of that, duct tape might have worked.)

2. At some point, small plastic spigots (with great reservations for English terminology) arrived on the market. The customer now had the option to buy one of these spigots, push or drill it into the carton, and then pour through the spigot—which was easily un-/resealed through a screw-cap.

   This was in many ways an improvement, but it did bring the disadvantages of having to own one or several additional pieces of equipment and some minor loss of milk, as the outflow of milk could not be brought as close to the edge of the carton as with the cut-out spout. Other disadvantages yet might have been present, e.g. through the fact that the spigot stuck out several centimeters over top of the carton.

   ---

   Side-note:

   Why “or several”? Apart from the possibility that someone might want to have several cartons open at the same time, there was the issue of switching cartons, where the spigot had to be exchanged or washed at least occasionally for reasons of hygiene. Having several spigots made such switches easier. (Here we also see a potential, but more subtle, disadvantage: what if someone failed to perform such occasional washings and eventually caught some type of illness?)

   Why would someone want to have several cartons open at the same time? There are at least two reasons: Firstly, milk in Sweden was sold with different fat percentages, many (especially, among children) had a near-religious preference, and these preferences were not necessarily the same even in the same family. Secondly, while such cartons were mainly associated with milk, other drinkables also used them (occasionally, in the early days; comparatively often, as time went by).

   ---

   ---

   Side-note:

   Of course, at this stage, there was a choice—no-one needed to buy a spigot and scissors still worked. The spigot could be argued as a net-improvement already through this choice: those preferring the old way could stick to the old way; those who preferred the spigot could use the spigot. (I preferred the spigot.)

   This is an important lesson: that a good way to make things better is to provide more options, as opposed to finding something new and forcing everyone else to switch. The world of software is particularly lamentable in this regard: in the days of yore, software makers often sought to give options and to enable; today, they try to force everyone into a one-size-fits-all model with as few options as possible, even at the cost of disabling instead of enabling.

   On the detail level, we have to consider whether flexibility comes with an additional cost that reduces the value of the flexibility. For instance, someone who used a spigot could ultimately cut open a spout to get at those last few drops in the carton. This would remove one of the disadvantages of the spigot, but it would do so at an additional cost in form of work, few would bother (leading to a waste of milk, even be it a small one), and those who did bother might have lost more value in terms of time than they gained in form of milk.

   ---

3. As the spigots grew more popular, cartons increasingly came with a little cross (or other marking) to indicate the optimal position for a spigot, and, eventually, even came pre-weakened at this spot for easier insertion.

   This was, again, likely a net improvement; and, certainly, made life easier for spigot users. However, it did come at the cost of a less durable carton, with a greater (if still small) risk of accidental breaking or penetration during transport or whatnot—and it did so without any benefit for those who preferred to cut a spout. At the same time, changes to manufacturing machines were needed and chances are that manufacturing was now more complicated.

   (I do not remember when the indications for an optimal spout-cut were removed, but chances are that it was at this time or not much later . If so, the choice mentioned above was weakened, to the additional disadvantage of “spouters”.)

4. In a next step, cartons increasingly came with built-in spigots. This was a further step in customer comfort for spigot users, but, again, not one without additional costs. Consider e.g. the now considerably more complicated manufacture, the environmental impact of using some additional quantity of plastic for each and every carton (as opposed to the old spigots that could be used for several years), and how there was now some space wasted during transport. (As the spigot stuck out slightly above the end of the carton, cartons could no longer be packed quite as densely.)

   From another point of view, users were now stuck with the one specific spigot provided by the milk carton, without the options that (potentially) existed in the past. Yes, of course, they could still put in a spigot of their own, but these largely disappeared off the market, the markings and pre-weakening were now gone, and the use of such a second spigot seems like overkill—even when considering the superior make of the old spigots. The last brings us to another disadvantage: the built-in spigots were much shorter and more prone to mis-pouring than the old spigots. (Presumably, to reduce the aforementioned disadvantages of packing density and use of plastic, as well as to reduce the risk that something would break before the carton reached the customer.)

5. This remained the state for a long time—maybe, even, since the late 1980s. Come 2024, there appears to be a new development: the screw-cap that tops the spigot no longer detaches properly from the spigot. This might have some advantage in terms of not losing the screw-cap—but I cannot even recall the last time that happened. Instead there is the disadvantage that opening and closing the spigot is a bit harder and that more attention has to be paid to the screw-cap, lest it gets in the way of the milk while pouring. Alternatively, the screw-cap has to be detached with a bit more force or e.g. a pair of scissors, but the former risks a breaking and the latter reduces comfort by several decades.

   Here, we see a change that will likely be a net worsening for most. (While the other instances have been net improvements, even while demonstrating that they are not improvements for everyone and/or in every regard.)

   ---

   Side-note:

   I am uncertain how pervasive this change is at the moment. So far I have noticed it while buying milk at Aldi, which prompted me to prefer Akzenta for my milk purchases. Last time around, however, the same issue appeared with the Akzenta milk. (Aldi and Akzenta are the grocery stores that I visit the most often. The respective milk brands are the cheapest in the store.) Ditto some other, rarer, purchases of other drinkables at Aldi, sold in a similar type of carton and with a similar type of spigot.

   For the time being, I hope that this is an experiment that will fail and that the earlier, easily detachable, screw-caps will be reintroduced (or, on the outside, that some compromise will be found where the “new” screw-caps are more easily detachable).

   ---

   ---

   Addendum:

   (2024-07-03)

   Roughly two weeks later, German news sources claim that this, beginning “today”, is a legal mandate, which removes any hope of an improvement. Specifically, and entirely unrelated to usability issues, some EU directive has been implemented in order to avoid loose screw-caps landing on the beach. Whether this mandate will have any significant effect on the environment, I leave unstated—but I do express the strong suspicion that this is yet another case of taking environmentalist action for the sake of taking action or of over-prioritizing details of the environment while achieving little in the big picture. (In both cases, without a proper cost–benefit analysis and a holistic view of pros and cons.) In this, it is a good example of why we need evidence-based politics: if the decision is a poor one, this would have been more obvious and the decision easier to avoid; if it is a good one, the sceptic could have convinced himself of this, instead of speculating about further incompetence from politicians.

   It does point to something interesting around the theme of this text, however, namely how easy it is to miss pros and cons that go beyond our own experiences, patterns of use, or whatever might apply: I cannot recall the last time that I drank any type of screw-capped beverage outdoors, let alone at the beach (nor the last time that I was at a beach at all), and complications like screw-caps being lost or forgotten by beach visitors did not even occur to me during the original writing.

   My purpose with this text is, of course, not to analyse the pros and cons of undetachable screw-caps (but to illustrate “new vs. good”), so I will give myself a pass. (And, no, I did not consider hypothetical explanations like “a detachable screw-cap could land in the throat of a curious toddler” either.) In a bigger picture, however, those who do set out to analyse the pros and cons in a reasonably holistic manner (or who should do so) often fail spectacularly, which is a major contributor to problems with e.g. software usability, because software is all too often directed at the weakest users with no thought for anyone else, and all to often targets the preconceived opinions (preferred usage patterns, ingrown habits, or similar) of some decision maker without concern for those who might have other opinions (whatnot). Analogous problems exist elsewhere, e.g. in politics. (And, again, we do need evidence-based politics.)

   ---

   ---

   Addendum:

   (2025-08-22)

   I can now look back at a little more than a year of experiences. So far, I have only rarely had reason to complain, be it because I soon picked up the knack of how to handle milk cartons without detaching the “new” caps or because minor early issues with the design/production/whatnot were soon ironed out. (For want of a direct comparison, and in light of the likely smallness of adaptions needed on either side, it is hard to say what applies.)

   However, when I buy something with a new cap other than milk, there can be minor problems—and I cannot rule out that other brands of milk also have issues, that others have greater issues than I do with the same milk cartons, or similar.

   For instance, earlier during the summer, I picked up a six-pack of very cheap beer in plastic bottles. (“Karls Krone” bought at Aldi—I do not recommend it.) At first, I attempted to drink straight from a bottle—and immediately came into conflict with the cap, which got in the way of my mouth. I now had two main choices, viz. (a) to fetch a glass, cup, or similar, and (b) to detach the cap. I picked the latter and detached the cap, thereby defeating the entire idea of the new caps. To boot, this cost me some extra effort and introduced a risk of spilling some bear. (This effort might have been almost negligible per bottle, but it could accumulate for a bigger drinker and the force needed might have been too much for some old lady. Likewise, more force needed relative maximal strength, as with that old lady, could considerably increase the risk of a spill. Likewise, someone already a little drunk or with poor coordination through age might run into trouble.)

   Glasses and whatnots, in turn, are not necessarily available at all—in particular on the beach (cf. the previous addendum). Ditto scissors and knives that might be used to reduce the force for someone weak. Of course, if such implements are used, they might pose a risk of injury (again, especially among the already drunk)—and what if someone remembers to bring the entire bottle, cap included, to recycling while forgetting a knife on that beach? Chances are that the bottle cap would have been the far lesser evil.

   Looking at beer more generally, I have had a heavy-for-me-year, this six-pack being my third. One of the previous two used traditional glass bottles with a metal cap—which was not attached in a similar manner to the plastic caps. I grabbed a bottle opener, opened a bottle, and the cap was completely free. Now, is a plastic cap truly so much worse than a metal cap that the difference in treatment makes sense? Let us say that it is: Is the sum of all plastic caps brought to the beach still worse than the sum of all metal caps on the beach? (Soft drinks likely mostly have plastic caps, but this is not universal. Plastic caps on beers are the exception. Milk is not a typical beach drink—if in doubt, because of the greater risk that it goes bad in the heat.) Let us say that this too holds: If the plastic caps are worse than the metal ones, should not the same apply to plastic bottles vs. metal cans, making banning plastic bottles a more constructive step than mandating the new caps? (I do not know the answers, but these are examples of questions that should be asked—and of questions that the likes of politicians and environmentalists tend to skip, in my impression.)

   The remaining six-pack consisted of cans with pull-tabs (reservations for terminology). The pull-tabs did remain connected to the cans (as they have for decades), but differ in that they were not in the way in the slightest, even when drinking directly from the can.

   From another angle: My problem with the plastic beer bottle arose from the choice of drinking from the bottle, while I cannot recall the last time that I drank directly from a milk carton. Vice versa, if someone does drink directly from a milk carton, he might well run into similar problems. Here we see that whether someone has or does not have a problem can depend very much on individual actions and choices, which makes it the more dangerous to apply a “works for me” thinking. (As a counterpoint, a “does not work for me” thinking can also be problematic. There might e.g. be a knack to handling the beer bottle that I had not yet mastered.)

   ---

   ---

   Addendum:

   (2025-09-09)

   Yesterday, I bought a bottle of Coca-Cola (plastic, 1 l)—my first since long before the original entry in 2024. I did not reflect on how the cap was attached or not attached at the time of opening, but it detached completely without problems and I cannot, after the fact, see any signs that the means of attachment would have been changed in a manner similar to the original milk cartons or, in the previous addendum, plastic beer bottles.

   Exactly such plastic soft-drink bottles, however, are among the likeliest to be found on the beach (or, as case may have it, in other scenarios where the caps can land “in nature”)—much more so than milk cartons.

   Why there was no change for this bottle, I cannot say, but there is reason to suspect incompetent politicians having, yet again, gone after the wrong target. (As opposed to “just” making a mostly-for-show effort that brings little benefit while causing inconveniences, costs, or other problems out of proportion to the miniscule gains—which seemed likely from the begin.)

   ---

Another interesting change is an increasing switch, likely around the mid-1980s, from a greyish or brownish type of carton to a whiter one. The latter was more aesthetically pleasing, but came with greater processing (likely, some type of bleaching), was environmentally controversial, and was, likely, slightly more expensive. While I have not seen one of these greyish or brownish cartons in a very long time, they were certainly present in at least Sweden until at least the early 1990s (maybe longer) as a minority option for those who feared unnecessary damage to the environment.

However, this is in my lifetime and Sweden/Germany. At other times and in other places, other types of packaging have been used, and these have their own advantages and disadvantages. For instance, the approximately brick-shaped cartons, to which I am accustomed, have advantages like easy stacking and dense transportation, a low cost of manufacture, and similar, relative the glass bottles that preceded them. Glass bottles, however, at the time had the great advantage of re-use; as did other types of earlier containers. We might then even have scenarios like a milk buyer going to the store with his own container, having it filled with milk, and, thus, having a smaller effect on the environment. Indeed, in past times, this was often the case and I have seen the suggestion by some environmentalists that we should return to this approach. Other approaches yet have other advantages and disadvantages, like the plastic one-gallon jugs of milk that seem to be popular in the U.S. and the Canadian bags (!) of milk, and might have different sets of advantages and disadvantages relative their predecessors.

A false appearance of objective improvement

An interesting special case is when there appears to be an objective improvement, e.g. by an improvement in some numerical proxy, but where this improvement in numerical value does not match an improvement in what truly matters.

A large example family is when various economic/business/whatnot indicators are not adjusted for price inflation (and/or value of currency, and/or purchasing power) and similar distorting changes. Consider GDP with and without adjustment and how a few years of high inflation can create the impression of great growth, while actual growth has been small or, even, negative. (More generally, GDP and similar measures are very dubious for reasons like being weak and potentially misleading proxies.)

Another, when so much effort is spent on improving the numerical value of some proxy indicator that it no longer is a good proxy, implying that an improvement in numerical value does not reflect an improvement in underlying “good”. At an extreme, the underlying “good” can even deteriorate as the numerical proxy improves. (Alternatively, that the proxy remains strong but that there is a shift in value ranges or whatnot so that (a) indicators from different times are not truly comparable, (b) there is a temporary loss of connection during a transitional phase. Ditto some similar variations.)

Yet another, when circumstances of other types weaken a proxy (shift value ranges, whatnot). A particularly perfidious special case is when someone replaces one proxy with another without clearly signalling this, e.g. in that a different measure of unemployment is used while both the new and the old measure are referred to as “unemployment”, without further qualifiers.

Athletics and the apparent ability to make numerical comparisons over time is a good source of illustration (and one that is not very political). For instance, the immediate trigger for this section was the news of an extraordinary-seeming discus world record by Mykolas Alekna. Alekna is justifiably considered a truly exceptional talent, who was mentioned as a potential future GOAT even as a teenager (similar to Mondo Duplantis, cf. below). As of the time of writing (2025-04-14), he has now twice improved an “unbreakable” record, and has done so by a great total margin. (Jürgen Schult threw 74.08 in 1986 (!), which remained a record until last year, when Alekna improved it to 74.35, with a second improvement to 75.56 (!) yesterday. Here and elsewhere, I draw on wikipedia on discus throw^w (oldid=1285483390) for statistics. All distances in meter.)

Superficially, Alekna has simply fulfilled the immense potential ascribed to him. In reality, there is great reason to doubt the value of these results:

Firstly, discus is a discipline highly affected by location, wind, and whatnot, and one where, unlike e.g. sprinting, there are no wind limits, which makes it susceptible to extreme outliers that do not reflect the ability of the thrower. (To make matters more complicated, discus throwers usually benefit from a head wind and sprinters from a tail wind.) The aforementioned Schult very rarely broke 70 and never came close to his own record in any other competition. Some others have/had a reputation as “wind chasers”, of searching for the right conditions to throw far, even be it in a low-status competition.

Secondly, both world records came in the same location and at the same time of year—just a year apart.

Thirdly, this location and these times have seen other extraordinary results. Last year did not only see Alekna’s first world record but also e.g. an astounding women’s 73.09 by Yaime Pérez—the longest that any woman had managed since the 1980s. (Note the great suspicions of extreme doping and the great successes of East-block women in the throws during the 1980s.) This year? Among other results we find a 73.52 by Valarie Allman among the women (another 43 centimeter beyond the extraordinary result from 2024) and, among the men, a 74.78 (!), above even Alekna’s first record, by Matthew Denny—backed by a 74.25 (beyond Schult but not Alekna) from a few days earlier. Now, assume that Alekna had been out of the equation. We would then have Matthew Denny as world-record holder—a far, far lesser talent than Alekna. Denny’s career high so far, already 28, is a bronze in the 2024 Olympics; Alekna, at 22, already has an Olympic silver (also 2024, missing the gold by an inch and through bad luck) and a handful of other international medals. Alekna was mentioned as a (legitimate) potential breaker of that “unbreakable” record for years, based on his rapid development, great potential, and the successes of his father (two-time Olympic champion Virgilijus Alekna, at 73.88, is the thrower who came closest to the record before 2024); Denny had shown nothing to imply that level of potential.

In conclusion, these results tell us very little about, say, where Alekna stands relative other greats and how today’s discus elite compares to those of past times. Moreover, they show that the world record in discus might be something of a curiosity relative even many other track-and-field events.

An interesting contrast is Mondo Duplantis, who has repeated world records in the pole vault (and a similar background of being hailed as a future GOAT even as a teenager): He has set his world records in a variety of venues, under much more controlled conditions, and at different times of year. Moreover, to date, none of the other current greats are even remotely on the same level of personal best (6.27 to 6.07 for KC Lightfoot; excepting the long-past-his prime Renauld Lavillenie at 6.16 from 2014). Moreover, if we look at second, third, fourth best results, his superiority often increases.

However, even here, personal bests (for comparisons between athletes) and world records (for comparisons between eras) are imperfect metrics. The most natural comparison for Duplantis is Sergey Bubka, but with Bubka’s personal best being set more than 30 years ago, comparisons become tricky. Consider e.g. the development in shoes that have had such a large effect on running events (it self a good example of changes to circumstances)—what if Duplantis has a speed advantage through better shoes that gives him an unfair edge over Bubka? Going back farther, pole vault has seen a number of changes in pole construction and materials. The poles of Bubka and Duplantis are reasonably comparable (but not necessarily to the degree that a Duplantis-advantage can be ruled out), but those of Bubka/Duplantis and Warmerdam (the third GOAT candidate among those in the know) are two generations (bamboo -> steel -> fiber glass) apart and Warmerdam’s results are on an apples-to-oranges basis relative Bubka/Duplantis—and a fair historical comparison between different vaulters requires that results with too different poles not be grouped together. (Much like, say, comparisons of unemployment numbers must consider the respective measure of unemployment that is used.)

From another angle, we have issues like training methods, ability to compete as a professional, ability to draw on past experiences made by others, etc. Duplantis had the benefit of getting off to an earlier start than Bubka and is e.g. known to have watched videos of other jumpers even as a child—and chances are that both had incomparably better conditions than Warmerdam. Even after adjusting for poles, shoes, and whatnots, Duplantis would likely out jump Warmerdam—but does this still apply in an alternate reality where they grew up and trained in comparable circumstances? In a next step, this raises the question what changes in circumstance should resp. should not be considered in a fair comparison. (Something that can be of great importance when we move back to e.g. politics.) That result comparisons require comparable poles is a given in almost all cases (ditto wind and whatnot conditions in the discus), but whether factors like training opportunities are properly considered depends on exactly what we want to compare and what we want to achieve.

Excursion on software

As is clear from the above, I have negative feelings about modern software (and modern websites).

The categories on software development and web design already contain some examples and discussions, if not necessarily with a mention of “new vs. good”. Ditto, some texts from my Wordpress days (TODO import and link), e.g. on the continual and severe worsening of Firefox.

I might or might not add further examples here in due time.

Excursion on the reverse error

It is easy to jump into the reverse error of taking the new to be worse than the old, to mistake a poor implementation of a new paradigm for a poor paradigm, to only see the disadvantages or weaknesses of the new and not the advantages and strengths, or to otherwise judge the new unfairly.

Consider digital photography and the mass-market boom that began in, maybe, the late 1990s:

1. Many users (in particular, professional photographers) condemned digital photography in general as inferior to “analog” photography when they should have condemned the then state of digital photography. Not only had far less time and money been spent on developing digital photography than on analog photography, but digital photography also depended on the state of supporting technologies that were rapidly improving.

   Most of the legitimate criticism of that then state grew less relevant with each year, and we are now (2025) in a reverse state, where digital cameras are usually better than combinations of analog camera and film—and even on points, e.g. effective resolution, that once were considerable weaknesses. (Different types of film can lead to different results just as different cameras can.) If we look at a “bang for the buck” measure, digital pulls ahead even further. Indeed, the built-in cameras on cell-/smartphones are usually enough for most photographic purposes. (And similar statements have already applied for quite a few years before 2025.)

   ---

   Side-note:

   I do not have the depth of knowledge to judge where analog cameras might still have an advantage, but chances are that some points exist, as there might be characteristic of the physical film or typical functioning that are unnatural for a digital camera and/or are hard to replicate, should one want to, with a digital camera and/or later digital processing. Likewise, there might be things doable with film exposure or the development process that are similarly hard to do digitally.

   ---

2. Even early digital photography brought a different set of advantages and disadvantages, and the impression created could have depended strongly on the needs, habits, and whatnots of various users—with the added complication that existing habits were often geared at the analog, because that was what everyone had used in the past.

   Consider the ability to view an image on a computer without the prior need to digitize the image: To someone not already in the habit of viewing images on a computer, this might be a mere nice-to-have, while e.g. having a sufficient level of resolution could be a must-have. This, again, in particular for professionals.

3. Some advantages were not immediately usable but required investment of time or money.

   The likely most notable case is the ability to edit an image digitally on a computer: For starters, go back far enough and there was no guarantee that the camera owner also owned a computer. (And, in a special case of the first item, any computer owned might have been painfully slow by today’s standards.) If he did own a computer, he still needed to invest in software, as cost-free alternatives were rarely competitive at the time (if they, at all existed)—and I do not rule out that more advanced users have to resort to commercial programs even today. And then there was the issue of being able to use the program in a proficient manner, which could be an obstacle even at a fairly mechanical level to an unskilled computer user, while even a skilled computer user could require serious effort to learn aspects as when to use what feature to get what effect, when what effect makes sense, and similar—just like there is much, much more to taking a good photo with an analog camera than the mechanics of pointing it in the right direction and pressing a button.

4. Other advantages might only have arisen over time. For instance, a smartphone with a digital camera is far more plausible than one with an analog camera, but smartphones did not exist back then. Regular cellphones did soon add cameras, but this too was not an immediate step and the advantages of having a camera in a pre-smart cellphone are much smaller than in a smartphone.

However, when it comes to changes imposed by others, it is important to remember that even a transitional phase (e.g. to learn new skills) is a disadvantage in its own right—and potentially a hefty one. Ditto, when the transition creates a new need for equipment or a need to modify or outright replace existing equipment.

When it comes to strengths and weaknesses, a relevant question is also whether a strength was lost, or a weakness incurred, unnecessarily or for a spurious reason. (And, from another angle, whether a legitimate reason was actually made public. If it was not, the reactions of those affected by the change might be unfair towards the change—but they might also still be fair towards the changer, who failed in proper communications.) A common scenario with (at least software) products is that a decision maker has a mentality of “I do not use this feature; ergo, it is pointless and should be removed” or “the majority of users do not use this feature; ergo, it does not bring enough value and should be removed”. (To make matters worse, many decision makers, e.g. typical product managers, have a poor understanding of topics like software and software development, usability, and what workflows and functionality make sense.)

Looking more closely at feature removal can be instructive even when it comes to more general points:

Removing features can have advantages in terms of e.g. lesser complexity and easier maintenance, and there is an upside to removing features. However, great care must be taken, lest more harm than good is done, lest unexpected harm is done, and lest the baby be thrown out with the bath water. (And those who constantly add new features to a product to see what sticks are hypocrites when they want to remove old and established features.) Among things to consider, even for features supposed to have few users, we have:

1. The actual amount of code needed for the feature, the likelihood of future changes, and similar.

   If the code is short, has not seen a non-trivial change in years, is well covered by unit tests, whatnot, there might not be much to gain by removing it—even in the unlikely event that it goes entirely unused. (And the less so, when there are users of the feature, possibly, some who positively need it.)

2. The actual number of users. (Also note the end of the previous item.)

   Two particular factors to consider:

   Firstly, decision makers are not necessarily themselves users, let alone skilled or heavy users, of the product at hand. They are, then, not good judges of what features are likely to be used by whom or how often. Other confounding factors exist, e.g. in that feature use might be estimated based on support requests or bug reports, while advanced features are more likely to be used by advanced users, who are more likely to resolve matters on their own, implying that advanced features might be greatly underrepresented relative the number of actual users (which can create an impression of “too little” use when combined with the naturally lower use of advanced features, because advanced users are in a minority). Similarly, some software makers rely on unethical phone-home functionality, where the software sends data about use and whatnot to the software maker behind the users’ back, but many advanced users turn this functionality off, skewing the user basis of the collected data (with additional side-effects on characteristics that correlate with the “level” of the user, like who uses what OS, e.g. in that a Linux user is more likely to be advanced than a Windows user and, therefore, more likely to have turned phone-home functionality off). Likewise, some Linux distributions (that usually deliver software through a centralized mechanism and/or in a modified-to-fit-the-distribution version) have changed a default “on” for phone-home into a default “off”, notably, with the excessive data gathering of Firefox. Bug reports, in turn, suffer from the additional flaw that a better/worse piece of code is less/more likely to result in bug reports at a certain level of use.

   Secondly, if the overall user base is very large (which it often is), then a small percentage of users can still amount to a great absolute number. One percent of one million is still ten thousand—and do we really want to piss off ten thousand paying users because they are “only” one percent?

   (To boot, such users are also often particularly important, a topic discussed further below.)

3. How much effort it would take to remove the feature and what might break by doing so.

   If the code is good, the effort might be small, but there will always be some effort—and exactly good code is a poor candidate for removal (cf. above). For poor code, with e.g. poor separation of concerns, data sharing, intermingling of functionality from several features in the same piece of code, or a great spread of the code needed for the feature, the effort could be quite considerable and the risk that something breaks is correspondingly increased.

   Such costs and risks must be put against the intended gains from the removal.

   I have even seen the approach to just declare a particular feature as unsupported/use-at-your-own-risk or to simply silently remove it from documentation, while keeping the code unchanged.

   ---

   Side-note:

   Whether this is a good approach, I leave unstated, but it does show how large the costs/risks can be relative the gain.

   This approach is not to be confused with deprecation, which is normally taken to imply that the feature is still fully functioning and maintained, but that it may be removed in a later release with no further warning and/or that there is a better way for the user to reach a certain goal. Deprecation might, however, be a sensible first step towards removal or use-at-your-own-risk—and certainly beats just killing a feature without previous deprecation.

   ---

4. What indirect benefits a feature might have.

   Some features can have great indirect benefits. Consider command-line flags and parameters: even on the off chance that no end user actually uses them (a very dangerous assumption—but for the sake of argument), they could still bring considerable value to the developers by making both manual and automatic testing easier.

   Indeed, such indirect benefits can be an outright cause to add some features. For instance, to add an internal scripting and/or plug-in ability can be quite expensive, but it will not just provide something for very advanced users (which might not outweigh the cost)—it might also make both testing and the creation of new features by developers easier, saving future time. To boot, it might give third-parties incentives to write and publicize their own scripts/plug-ins and, thereby, make the product better even for less advanced users.

A particular complication is that features rarely used by the broad masses might be very valuable to particularly important users, including heavy and skilled users who love the product and promote it by word of mouth, users who bother to contribute bug reports or (in FOSS) submit patches, and corporate users who often pay considerable amounts of money for use. For instance, say that a corporate user has used a product remotely for years by logging in to a server per SSH and working on the command line. Now a product manager decides that various command-line parameters should be stripped, or a config file removed, because it would be enough that the corresponding functionality is available in the GUI. This would be a PITA for the corporate user, who might want fast, easy, and consistent control—but is now faced with extra work and a greater risk of mistakes even for manual actions. Scripting/batching, of course, might break entirely (or a user be forced to interact with what should be a batch run). Etc. Then we have the issue of how to at all get to the GUI. Plain SSH is out. X-over-SSH might work if we have a Unix-verse server (maybe, after the installation of additional software on the user-side), but this is far from a given and might over-task many weak users. Chances, then, are that a heavyweight solution like VNC must be used. Such complications might tip the scales towards the product of a competitor.