Shambles, But Make It Digital

by Miriam Ronzoni on October 1, 2024

I work at a very large University. I am the parent of a child who has just started secondary school, and of one in the middle of their primary school journey. I am currently taking what in the UK is called a Level 2 Adult Education course. In all four of those domains – and more – there is a conspicuous absence of  a streamlined place to access, or input, material – be it learning materials; homework; stuff to mark; lecture slides; lists of students; exam dates; you name it. Things are instead, scattered through a multiplicity of platforms and apps, with no particular rationale, order, or clear chain of command.

Of course, each of them requires different passwords (or more robust login methods still). What is more, most of them are not particularly good: they are not very user friendly; they are all structured in different ways and require roundabout procedures to access information; and finally, many of them have persistent, deal-breaking glitches. The app which my son’s football team requires me to use to register his availability for matches doesn’t actually let me log his availability two times out of three. The platform on which I am supposed to log my homework for the course I am taking fails to save my work on random occasions, but when I try to address the problem by writing everything up in a  word file first…I then find out that it is impossible to copy and paste stuff into it. I could go on, but you get the gist.

Of course, my first thought was that I am just a dinosaur. But some of my younger course attendees are even more helpless than I am, and both my students and the pupils at my oldest child’s secondary school are permanently overwhelmed and confused.

So, CT readers and fellow bloggers, what is the issue? Seriously, I am asking you, there must be some fledgling literature, in sociology and/or organisation studies, addressing this? My clever husband speculated that this might be (hopefully) a transitional phase: there are too many apps; quite a bit of money to make producing them; quite a bit of inertia within large organisations when it comes to transferring stuff that has already been started somewhere else: and generally a lack of investment in streamlining (by allocating staff to this task, for instance). As a result, nobody is really doing the job of making online platforms more user friendly, and there is a lack of incentives to do that anyway. I find this thought quite plausible, but it is just armchair sociology at the end of the day. Is there soome truth in it?

Still, the issue isn’t trivial. It is wasting workers, and users more generally, a lot of time, and it probably contributes to the mushrooming of bullshit jobs and unnecessary menial tasks that technology was going to liberate us from. I haven’t been able to find any work on this so far – if you are aware of any, please share it here.

{ 40 comments… read them below or add one }

1

mkvf 10.01.24 at 10:14 am

I don’t think this is unique to education. I see it at work in media CMSes and other tools, and whenever I try to contact the NHS. The problem, I would guess, is who decides whether to buy a piece of software: typically not employee users, let alone external ones. Instead, senior managers who never use the tool themselves, or interact with customers, buy it based on claimed benefits—particularly, how much it saves on staffing or staff time. Until it breaks so badly you lose business—say a restaurant booking tool that doesn’t record reservations—there’s no incentive to replace it.

2

Miriam Ronzoni 10.01.24 at 10:16 am

Yes yes I was in no way suggesting that it was only happening in education!

3

notGoodenough 10.01.24 at 10:34 am

(apologies for the double post – please do delete the previous as I fluffed the html formatting)

Perhaps an answer may lie in this thought:

it seems that one of the things about treating something as a business is that the primary (and, arguably, only) priority becomes “making money”, with “providing goods and/or services” being merely the means to that end (and, quite possibly, a means which – should it prove inefficient and unnecessary – may be dispensed with)

That is to say, commodification inherently leads to exploitation as profit is more important than providing commodities.

(c.f. the various digital entertainment industries moving towards business models which prioritise maximums extraction from customers over convenience, value, etc.)

4

Jestyn 10.01.24 at 10:53 am

Isn’t it just the demon of corporate IT – systems are bought by purchasing departments not end users, so they go for things like cost and deployability as the main drivers, not ease of use. Anyone who ever tired to use Peoplesoft or SAP know all about that…

5

Doug Muir 10.01.24 at 10:57 am

It’s not only happening in education, but it might be particularly bad in education. Four kids here, and they’ve all passed through multiple schools. And the online experiences have been wildly various but have mostly ranged from mediocre through bad to horrible and destructive.

We’ve had pretty much exactly the experiences you describe, plus some additional awfulness — like the online course that kept mis-grading my son’s submissions, and where the human in the loop agreed that it was graded wrong but wouldn’t (probably couldn’t) change the grades.

Why education should be particularly bad in this regard I don’t know. But it kinda seems to be.

Doug M.

6

B 10.01.24 at 11:18 am

This basically: https://www.wheresyoured.at/saaspocalypse-now/

Organisational resistance to switching costs leads to software companies having no retention issues and therefore no incentive to make their software better, and plenty of incentives to make it worse so they can upsell a premium version which bypasses the problems they put into their basic software (and also increases sunk cost and switching costs in the process). The people who have to work with the software are rarely the ones who decide which software to use, so sales tactics are supreme and having an actually good product is secondary if that (and even then “looks good” takes priority over “functions well”).

7

Mike Huben 10.01.24 at 11:22 am

Enshittification?

Are there free software projects that address this problem?

The real solution would probably be to empower users to perform the selection and require fixes and upgrades.

8

Dan Weiskopf 10.01.24 at 11:39 am

Seconding what everyone above has said. It’s the SaaS business model. Ed Zitron has a long discussion of this in a recent issue of his newsletter. He argues that SaaS depends on making sales to managers who are not the primary users of the platform. But it only needs to “work” in some vague sense, and once in place it becomes impossible to dislodge since there is no incentive to fix bugs or provide off-ramps for clients to migrate to other options. So the software ends up low-quality, unresponsive to user needs, and impossible to integrate into a coherently functioning system (although other SaaS providers specialize in that, too). The whole discussion is here:

https://www.wheresyoured.at/saaspocalypse-now/

That’s probably the type of institutional analysis you’re looking for.

9

Matt 10.01.24 at 11:45 am

I guess this makes me feel slightly better about the system we use here where I teach (which, I think, is a modified version of Blackboard, though I’m not 100% sure.) It’s clunky and not pretty, but… it does all the things you mention tolerably well. Some things are annoying (I have small sections as well as a larger lecture, but the “notice” feature doesn’t let me send emails to just certain sections, for example) but it mostly works. It’s actually generally worse than the somewhat similar systems I used back when I used to teach at Penn, 7 years ago, so some such things that work tolerably, if not great, have been around for a while.
(We, and many Australian universities, do use some awful software designed by Elsivar that turns our “research outputs” into awful, ugly, but bureaucratically accepted faculty web pages, but I mostly try to not look at those. They are not needed day-to-day.)

I’ll also say I’ve mostly been pleasantly surprised with the software that the Australian government uses for taxes, medicare and some other things. To my mind, at least, it’s fairly user-friendly (though maybe I’m just starting with the very low standard of interacting the the IRS in the US), “clean”, and not hard to access.

This shouldn’t make you feel better. Maybe it should make you feel worse! But, it does seem that some even not that impressive institutions can get and use at least pretty okay software.

10

Bradley C Kuszmaul 10.01.24 at 12:26 pm

The purpose of a system is what it does.
https://en.m.wikipedia.org/wiki/The_purpose_of_a_system_is_what_it_does

The system is very cleverly constructed to waste your time and confuse you.

11

TF79 10.01.24 at 12:51 pm

My sense is that while the medium and mode have changed, people had an incentive to half-ass things in the past, half-ass things in the present, and will half-ass things in the future. So while it may be annoying to click through 3 links that may or may not work to get to some flier that doesn’t bother to actually say what time my kid’s event actually starts, presumably 25 years ago that would have (maybe) come home as a flier in the backpack missing the same information.

12

Dave Timoney 10.01.24 at 12:57 pm

The problem is exacerbated in the public sector by:

1) A lack of money, so old systems are retained beyond their useful life, not properly maintained (hence the ransomware attacks) and not adequately supported.

2) Organisational fragmentation (often the result of marketisation) leads to a lack of interoperability and failures in the adoption of best practice.

3) Not only are buyers usually clueless managers, but systems are rarely developed with the actual end users (e.g. benefit claimants, patients etc) in mind, and they certainly aren’t involved in testing or ongoing improvement.

4) Many suppliers have a talent for winning contracts, but not for actually writing software (the Post Office Horizon system, developed by Fujitsu/ICL, a hardware company, being a case in point).

In the commercial sector, a growing problem is the ease with which systems can be built by the technically limited using CMSs, app builders and now LLMs to write code. Rather than an efficiency gain, this leads to software that lacks robustness, fails outside of tightly-circumscribed use cases and lowers the incentive to employ actual software engineers, so reinforcing the cycle.

13

Dave 10.01.24 at 1:11 pm

I have a child in early elementary school. There are approximately a dozen platforms in use: for aftercare communications, school lunch, grades, two for math homework, one for language arts; plus assignments in e.g. google slides. Then there is the parents’ WhatsApp chat group, full of very nice people who are usually confused and frustrated. It never stops.

The school software is junk and is sometimes downright invasive: it wants access to the computer’s camera, and in the case of school-sanctioned games, they want money—actual dollars!—for plus ups. I find it exploitative and enraging. Fortunately, my child is very reasonable when we say we’re not paying for that stuff.

I assume these apps come from vendors who are friendly with the administration and maybe the state I live in. Once something is implemented, it can stay in the budget and be accounted for forever. I wish there were the slightest bit of quality control, and I wish it didn’t have to be a political fight to rationalize schoolwork and communications.

14

Rj 10.01.24 at 1:23 pm

As a software developer, everyone that I know that works on edu complains about how hard it is to get funding to do anything useful. It seems common for margins to be cut so thin that they barely cover the costs of installing the systems, with no allowance for integration or upgrading.

Having said that, integration of software stacks is a very difficult problem, since each package has their own idea of the world, their own release schedule and their own incentives. “Glue” code to bind a particular stack together is custom, finicky and thankless work that requires ongoing maintenance and shrewd prioritisation because there is never enough resources to implement all of the sensible features and users often want old bugs to remain when they have built their workflow around them.

The best software tends to solve a quite narrow problem and be very opinionated, representing one person’s ideal way of solving the problem and if that doesn’t match the way you think about it, things get difficult quickly. Managing this sort of complexity requires skills and long-term planning that is certainly not universally available. As software diffuses through the economy and our culture, presumably the skills will too, although it is likely to remain unevenly distributed.

Basically, it’s a hard problem.

15

Aardvark Cheeselog 10.01.24 at 2:41 pm

The solution is strict liability for software developers (no more “this product is sold with no warranty or representation of suitability for any particular purpose”). You have to be able to show that you made a good-faith effort to make it work, and that you respond to reports that it doesn’t, or the State takes all of your money and gives it to the people you fucked over.

Because the problem is for-profit software development, the resulting pressure to be “first to market.” Being “first to market” with something that doesn’t work needs to be punished, not rewarded.

16

qwerty 10.01.24 at 3:28 pm

Indian IT companies are the cheapest. I’m familiar with HCL, in particular. It’s a huge company. And I read somewhere that they developed software for 737 Max. ’nuff said.

17

Dan Strong 10.01.24 at 4:06 pm

Other commenters make very good points about market and organisation level dynamics which lead to this issue. At a higher level, I think this is an outcome of the piecemeal privatisation of public services – which is itself arguably a second phase of the waves of privatisation that have been eroding the UK’s public infrastructure for decades.

The piecemeal approach is a reaction to previous models of IT outsourcing which saw central government sell off public sector computing infrastructure to huge multinational providers, then pay those providers for the privilege of using the expropriated assets. The government’s predictable reaction to this was not less privatisation but more – divide up public services into tiny slices and dish them out to a set of (at least nominally) competing firms, and hope that the invisible hand disciplines the capitalists into providing better quality services for cheap. Well, we all know how that’s going.

The only way out of this mess is to exorcise the Thatcherite vampires who have had their fangs in our public services for far too long. Unfortunately, the current ruling party seems to exist largely as a vehicle for private interests, so I’m not holding my breath for a legislative solution.

18

PaddyinIowa 10.01.24 at 8:07 pm

To follow up on “the purpose of the system is what it does.” As they used to say about Facebook, “You’re the commodity, not the consumer.”

https://www.hrw.org/report/2022/05/25/how-dare-they-peep-my-private-life/childrens-rights-violations-governments

19

somebody who remembers getting their first e-mail address at the university through the unix user's group 10.01.24 at 8:48 pm

the article linked by Dan Weiskopf @ #8 tells a great deal of this story. however the real issue is a disinvestment in the labor of education. you could have a system where you just type up (for example) the essay you need and e-mail it to the professor, who reads it and grades it. but that requires the school to pay the professor to take some time with their e-mail, to give the professor the leeway to let them decide which submissions are valid, and so on. and perhaps a few bad actors would fall through the cracks. but at least you could copy and paste your work from one to another.

the profit motive is even more broadly at work as Dan Strong points out at #17. the professor could obtain course materials in easily accessible, searchable, copyable and transmissible formats, and send them to the class by e-mail, and they could read it and copy/paste it when and how they liked. but that requires the people producing the materials to understand that perhaps a student might send some of that to a friend when trying to learn, without paying them extra. this would be fine if the people producing the materials were paid to produce them to be used freely in education. but you don’t make millionaires into billionaires by doing that. you only make educated, happy students and teachers. the nightmare of the privatizer

20

Alex SL 10.01.24 at 8:55 pm

Nothing new to add, but seconding what others have already written. Based on my experience with corporate software being introduced, replaced, upgraded, and very rarely reverted after a change that was too catastrophic to ignore, it seems to be a combination of:

The people making the purchasing decisions are not the end-users. Somebody on Mastodon suggested to me the other day that Excel may be working okay (except for the date issue) because even high-level managers use that, whereas they never sully themselves by interacting with many corporate IT systems except through their personal assistants.

To the degree that high-level managers care about functionality, it is things like how nice the report looks that the system spits out about organisation-wide activities rather than user-friendliness for the staff who have to put things into the system.

The people coding the software often care more about having it easy to code than whether it works well for the end-user, and they tend to view end-user testing as a box-ticking exercise rather than something that should trigger changes if it goes poorly. I have had the misfortune to work with one developer who bragged about how lazy he was, and another story was related to me where the developers tried to pressure stakeholders into signing off on end-user testing as successful despite everybody providing very negative feedback. Over the years I have repeatedly been shocked by how limited the skillset of most programmers is and how little they care to understand the world of their end-users – with the exception of some developers working for small companies, who have to deliver well to succeed.

Vendor lock-in. Because it is so difficult to migrate to a new system, once your organisation has been sold on a system, it sticks, and poof goes any incentive for the provider to make it user-friendly or functional.

None of this is likely to improve without a fundamental change in corporate culture, as in CEOs suddenly starting to care deeply about whether their thousands of front-line staff have good user interfaces and search algorithms. (Even then, I honestly have no idea how the search function on Confluence can be as broken as it is without it being intentional.)

21

Kevin A. Carson 10.01.24 at 9:04 pm

Two extremely relevant concepts:
“Captive clienteles,” which are described in Edgar Friedenberg’s book The Disposal of Liberty and Other Industrial Wastes. The gist of the idea, updated for the digital age, is that most on-the-job productivity software is designed by stovepiped R&D bureaucracies for sale to the stovepiped IT departments of other corporations, and the latter acquire it for use for workers whose feedback is irrelevant because they have no choice but to use it or quit.
“Enshittification,” a concept developed by Cory Doctorow and Ed Zitron in their blogs.
I tied the two concepts together here: https://c4ss.org/content/59818
Somebody noted years ago — I forget who — that the quality of work people could perform at home on their own open-source, browser-based or desktop software was orders of magnitude using the software tools supplied by their employer.

22

Andrew Morris 10.01.24 at 9:05 pm

In the days before personal PCs were available (yes,I’m that old), I had a friend who sold accounting software to large companies. When his clients asked for additional capabilities from the software, he said that that would be a “mod,” and would quote the cost for modification. Then, after some time, the modification would be added to the basic package with the R & D costs picked up by the client. Same as it always was.

23

Bruce.desertrat 10.01.24 at 9:12 pm

Came here to link Ed Zitron’s blog, https://www.wheresyoured.at/ which multiole people have already. It’s an extension of the general ‘enshittification’ process and the rot economy.

Exponentially growing levels of companies trying to squeeze profits from inserting their own SAAS into ever more parts of the business processes that the customer needs. I have run into SAAS companies that integrate into other companies’ SAAS offerings.

And EVERYTHING runs on the basis of ‘Move fast and break things’

24

Sandwichman 10.01.24 at 9:18 pm

“…there must be some fledgling literature, in sociology and/or organisation studies, addressing this?”

Or maybe someone wrote something but couldn’t get it past the ELSEVIER submissions discouragement program?

25

Alex SL 10.01.24 at 10:22 pm

Seeing the comment by Sandwichman: oh yes, editorial manuscript management software is a perfect example of the problem. I was once corresponding author for a manuscript with more than sixty files, most of them figures. (It was a taxonomic monograph of a genus with tens of species, and each species needed a ‘plate’, i.e., a large figure illustrating it, and bundles of 2-4 species each had a map showing their distributions. That adds up.)

The main manuscript file is required to be the first in order, or you can’t proceed. I had to change something, so delete the main file, upload the updated one, so it is now last in order because it is the most recently uploaded. This then meant going through over sixty drop-down menus, one for each file, adjusting them individually to “you are now #11 instead of #10”, “you are now #12 instead of #11”, and so on.

A competing software lets the user drag files up and down to change order, so this could have been done in two seconds, tops. And just as an aside, an even smarter software could automatically put the main manuscript file first in order the moment the submitter flags it as such. But the problem is that journals find it difficult and disruptive to migrate from one system to another, and the problem is the manuscript submitter’s, not the editor-in-chief’s, and much less the developer’s, so…

26

Matt 10.01.24 at 10:30 pm

you could have a system where you just type up (for example) the essay you need and e-mail it to the professor, who reads it and grades it. but that requires the school to pay the professor to take some time with their e-mail, to give the professor the leeway to let them decide which submissions are valid, and so on. and perhaps a few bad actors would fall through the cracks. but at least you could copy and paste your work from one to another.

Having been fairly positive about the system at my current university above, I’ll mention that, this last term, I taught a fairly small jurisprudence class. Because it was small, I decided that I didn’t need to make drop boxes for the final paper and have students submit via that, and give feedback via that, because it’s not very convenient for giving the more substantial feedback I wanted to give these students. (In particular, it’s bad for giving comments on particular bits of the paper itself.) So, I told them to just email me the paper, I marked them, wrote significant comments, emailed them back, and put marks in the gradebook. The other assignment for the class was a presentation, and of course there wasn’t anything to “submit” for that, so I also just put in marks on the gradebook, and emailed comments to students after they presented. At the end of the term I was subjected to a spanking when I was told that things couldn’t be done this way – that I had to create submission folders for everything, submit that way, etc. So, I had to do this retrospectively. This was, supposedly, for “objectivity” or something – some typically Australian psedudo-value coming from being over-bureaucratized. In this case, the problem wasn’t the system, but those setting it up having the wrong values. I do wonder how far that’s the biggest problem.

27

JoeinCO 10.02.24 at 3:57 am

It’s not worth overthinking this. This is the first wave of enshittificaiton where software companies screw over the end users. I think the argument that the capitalistic incentives (market + network effects + asymmetrical information) inevitably lead to this are pretty strong. Pretty soon they will screw over the middlemen (the institutions that purchase the software). And then we are back to the Stone Age, but with really bad software that we have to use. Cuneiform tablets would be more efficient (I know, Bronze Age).

These rules for living might be relevant:

Tech eventually ruins everything it touches
Tech just wants to put you out of work and appropriate your salary
In Silicon Valley, tech monetizes YOU!

28

Chris Bertram 10.02.24 at 6:52 am

@matt “Because it was small, I decided that I didn’t need to make drop boxes for the final paper and have students submit via that, and give feedback via that, because it’s not very convenient for giving the more substantial feedback I wanted to give these students.”

At many British universities you would not be permitted, as a teacher, to deviate from the approved system in this way. Using something like Blackboard is mandatory. There are no opt-outs.

29

Matt 10.02.24 at 7:01 am

At many British universities you would not be permitted, as a teacher, to deviate from the approved system in this way. Using something like Blackboard is mandatory. There are no opt-outs.

That turned out to be so for me, too! I guess I shouldn’t be surprised by this point, but this is still so strange and stupid to me, and counter to what was the case everywhere I taught in the US, that it’s sometimes hard for me to remember. I sometimes make the mistake of asking why something like this must be done, but do these days mostly just try to remind myself that it won’t help me, I won’t get a real answer (other than “that’s what the rules are” or “it’s important that everything is done the same way”, with no explanation as to why that’s important), and that it will do no good to push. So, I try to teach myself acceptance, even though the whole thing is dumb.

30

Tim Worstall 10.02.24 at 9:16 am

Two not wholly serious thoughts:

1) In the early days of the motor vehicle the reason to have a chauffeur was not to have a driver but to carry your necessary mechanic with you. Early stages of techs are often pretty bad.

2) The video recorder solution. The only people who ever did work out how to set a time, channel and so on to record a specific show were 8 year olds. To the point that borrowing one to do so was the only rational solution. “in the middle of their primary school journey” there’s the solution, they do all the online data entry.

31

Station 10.02.24 at 2:59 pm

In regards to the simpler means of email for collecting work and communicating, I believe this might technically be a regulatory issue. Email is by nature insecure and poses risks, at least in the US, in regards to privacy. It strikes me that much of the egregious software in education is generated to assuage liability woes of the institutions (for instance the wave of “accessibility” vapor-ware hitting the scene). It’s almost as if there’s a collusion between regulatory bodies and industry to shake down public institutions or something.

32

superdestroyer 10.02.24 at 6:15 pm

This is an expression of the problem that every app/program/automation has to take steps to allow for all of the exceptions to the general rule. Designers can make a great system for a computer/phone/device configured the same way that their own device is configured. But once lots of different apps/programs/operations systems/devices are encountered then lots of coding has to be added to deal with all of it. In addition, all educational code has to be ADA complaint or the equivalent in other countries.

33

John Q 10.03.24 at 12:19 am

It’s important not to underestimate the role of simple incompetence, a horse that is always limping.

For example, Elsevier’s (I think) journal management system requires a separate login and password for every journal, when they could have set up a single login for all their publications. And, the way they’ve set it up means you can’t easily use a password manager. I don’t think there’s any secret plan here, just sloppiness.

Another example, not primarily technical. The Springer publications have started aggressively “unsubmitting” articles because of minor deviations from their formatting rules, which also differ from journal to journal. That’s in part because the job has been handed off to admin staff who have no idea what matters and what doesn’t. In the short run this saves work for the editors, but eventually authors get sick of it and either attempt end-runs, approaching the editors directly, or just go elsewhere.

34

Alex SL 10.03.24 at 5:27 am

John Q,

I didn’t read the discussion here to imply that incompetence wasn’t the main issue. In fact, it is probably behind all of these cases, and the most intentional aspect is vendors locking customers into their software ecosystems by making it difficult to migrate to another. But failing to make decent user interfaces, search functions, and systems that don’t save entries properly are squarely issues of competence. Even the intentional vendor lock-in is most important in how it removes incentives to become more competent.

The unsubmitting of articles over trivialities is something I have also noticed, e.g. with justifications such as that tables should be a separate file instead of part of the main manuscript file or that acknowledgements should be after the abstract instead of after the main text. But that isn’t a software functionality issue. Editors could simply change culture to review manuscripts first and only enforce citation styles etc. after tentative acceptance based on the quality of the science.

Tim Worstall,

There is a difference between a technology being in its infancy and incentives being so misaligned that nobody has to deliver well-functioning products now or in the future. In my personal experience I am dealing with different software where variously:

You need to confirm participation in something by clicking a button bizarrely labelled “[+] add” in the upper-rightmost corner of the screen.

Tax invoices from suppliers are regularly rejected because the file name of the PDF is too long. I assume that part of the code hasn’t been touched since 1990.

The search function is entirely incapable of finding a document that I know exists and for which I know the first few words of the name and its ID number, so that the only way for me to get to it is to enter a long URL with that ID number at the end into the browser.

I also once had to submit a lengthy grant proposal through a web form that did not accept characters like colon, square brackets, or (wait for it) line breaks. Those proposals must have been fun to read for the review panel, given the complete absence of paragraphs.

These are all easily solvable problems. Search is a solved problem. Others can be resolved by asking one (!) end-user to try the system, listening to their feedback, and making a few trivial changes to the interface. The equivalent would be your early cars having the steering wheel at a 90 degree angle to the direction of travel, and manufacturers thinking, nah, redesigning the car is too much hassle, and what’s the end-user gonna do? Their CEO decides to buy our cars, and he doesn’t drive himself, so we are good.

35

Trader Joe 10.04.24 at 10:47 am

The flip side of this problem is that when software is really good (thing Apple, Microsoft, Google et al) ….it often becomes very dominant to the point where the next thing you know is regulators are howling about monopoly power and walled gardens and looking for ways to break it up.

Unfortunately for us poor users you’re left with being either exploited by monopolists or being exploited by incompetence…..the choice isn’t even yours.

36

Zamfir 10.04.24 at 12:29 pm

Sometimes at work, I encounter work from the pre-PC age. And you know, it’s typically pretty good? I don’t think I am doing much better work, or much faster.

A common pattern there: computers made something much easier for me than for the Olds – and as result they concentrated on the core items and did those well.

For example, hand-typed contracts and specifications are much shorter than they are today. They skip details that are considered Absolutely Vital today, and get copy pasted in. Calculations and graphs are done by hand, and only show approximations for a few typical points where I can generate exact results for every automated measurement point of the last umpteeth years. The internet gives me instant information on many topics, but people used to have some cabinets of paper reference works, that cover 95% of the stuff that matters for the job. Old hand drawings are closer to what I would call a sketch today, but they are really good sketches – every detail, every dimension was added for a reason, not just because the CAD system could spit it out.

For all those things, I have software that works genuinely well and would be the envy of the olds. I am still not 100% sure how great the real productivity benefit is. If the software is not good (or is good but the seller too powerful, as Trade Joe notes), we quickly go negative

37

clew 10.05.24 at 10:29 pm

as a techie so old as to be nearly ex, I am repeatedly bewildered that the common systems seem to be much worse than “SVN with directory structure and per-directory access” plus “submit in Markdown or maybe HTML” would be. And I would expect LaTeX’s fiddliness to be unmatched, but Alex SL makes me wonder.

Some of it is how much some people hate hate hate even looking at plaintext…. anything. This feels like we solved it for HTML and surely we could manage for other things?… I don’t know. I don’t know why public key cryptography never took off either. People complained about having two passwords and look how many we have to have now.

38

Alex SL 10.07.24 at 2:06 am

clew,

If you are referring to the proposal submission form in particular, the theory among colleagues was that it had only an extremely short white-list of permitted characters because somebody was worried about database injection attacks. I am not a professional coder, only doing some scripting for bioinformatics and other analyses, and I certainly don’t know how that form actually worked behind the scenes, but my uneducated guess is that there should be some way of avoiding such attacks while allowing line breaks.

39

Raven Onthill 10.07.24 at 10:38 pm

This has been an on-going problem in software engineering for decades. Pioneering software engineer Brian Kernighan (yes, the one who co-wrote the C book):

The interesting, sorry, I shouldn’t be saying “interesting” – the areas that are difficult are only two: one that it’s too hard to write programs that work, and the other that it’s too hard to use computers. So if you want things to work on, these are two that you could try. – Brian Kernighan interview, July 2000

And this was true, 25 years before that interview, and it is still true, nearly 25 years later. Web applications are what we used to call “teleprocessing” applications, and those have always been difficult – there are multiple systems and many layers of software involved. There are solutions, but they involve commitment, more staff, and more work, and so they aren’t implemented. Add to that the contempt that managers and executives have for the line developers and the “folk art” status of software development – these days many people know how to solve small problems, and so think large problems are easy – and a lot of web apps are of low quality.

40

Moz of Yarramulla 10.08.24 at 8:25 am

I remain convinced that much of the problem is the holy trinity: buyers, users and developers. Who must stay as far apart from each other as possibly and ideally never communicate at all. Or alternatively, they have different interests and different goals.

Worth noting that the same problems bedevil other formal rules systems, from those governing football to statue law (judicial law is probably more akin to the templates and formulae so beloved of ‘office’ software users (‘horribly bug-ridden’ is charitable!))

An iron rule of software is it either fails completely or grows so complex that the organisation that produced it can no longer understand it, and thus can no longer expand it or even maintain it. This is true both of software made by a single individual and of massive corporations.

Leave a Comment

You can use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>