[delichon]

> I’d never considered the kinship a mason might feel, taking up the work of restoring or repairing the work of a predecessor who may have died hundreds of years before.

I wish I could feel that kinship with the programmer who wrote the code that I now have to maintain and extend, who quit and got a job writing spaghetti somewhere else.

[foobarian]

Maybe that's how it is with masons too? You never know what you might find behind the facade :-)

[Bluestein]

> I wish I could feel that kinship with the programmer who wrote the code that I now have to maintain and extend

By all means it could, should (must?) be that way.-

> Maybe that's how it is with masons too? You never know what you might find behind the facade :-)

To that, indeed. Who knows what "debt" or "enshittifcation" is hiding behind obscure technical details of any craft, seen by experts alone ...

That said, by an large, cathedrals have endured the test of time. I so wish we could say the same about the systems we are building nowadays.-

[nxobject]

The cathedrals in our industry pop up when least expected. I doubt the first IRS, Sabre, or American Express programmers working on IBM mainframes in the 60s/70s thought their programs would last so long, even at great cost.

[Bluestein]

Indeed. Cathedrals built with COBOL.-

[tlb]

In software, most people dislike most other people's work. I wonder how much that varies between fields, or if it's been studied systematically. It might be pleasant to work in a field where you'd enjoy most other people's work.

[short_sells_poo]

People aren't incentivised to write code that somehow carries true legacy the same way that a master mason will want their work to be a legacy for decades, if not centuries.

This creates a positive feedback loop where people just shit out software as fast as possible, which then means the next person to come around will be met with maintaining a pile of poo. This person will also not have any time (or mental energy) to do anything to be proud of and so the circle continues.

And it's entirely understandable. Clueless PM ABC wants to increase some barely relevant KPI so they can get a pat on their head from Clueless VP XYZ. PM ABC thus invents yet another user hostile feature in Windows 11, and thus Programmer 123 gets the thankless task to integrate ads in the right-click context menu. The programmer might be an artisan at their craft, but ultimately they just want to get their 9-5 done so they can do some actually exciting stuff in their free time. Then 1 year later Programmer 124 comes around and is horrified to see the boondoggle that 123 left behind.

[Bluestein]

> People aren't incentivised to write code that somehow carries true legacy the same way that a master mason will want their work to be a legacy for decades, if not centuries.

You raise a great point. Methinks there's a pernicious "assumed ephemerality" of software systems that has a "flywheel" sort of effect in that loop, creating sinks, which, thus, "enshittify".-

I honestly wonder what it would look like if software were "set in stone", and the assumptions were reversed. You were building systems for the ages, and ever and ever ...

[short_sells_poo]

This is really it. There's no permanence and little to no personal consequence for bad code. In fact, everyone wants everything done ASAP. Time is money, and very few places have the culture and the resource to afford to build long term stuff.

And yet there is software that has stood the test of time. It might not be pretty from a purely ideological or aesthetic perspective, but it lays the foundation of many great things. Consider something like Numpy. The internals are a multi-decade boondoggle of graduate level python code, Lovecraftian Fortran that causes you to take psychic damage just by looking at it, and random hacks that some disgruntled researcher submitted at 10pm on a Sunday after an all weekend marathon to get his stuff finished on time.

But so much good stuff was built on this pile of bad code, because it is not the concrete code that is ultimately of value and worthy of admiration, but the concept. The implementation is sort of secondary. So perhaps it's the ideas and the concepts that have legacy in computer science, code is just a carrier. The same way that genes are the definition of the human race, an individual human is just an ephemeral implementation of those genes. It is the genes where the miracle happened really and which gave rise to beings that are self aware (to quote Richard Dawkins).

[Bluestein]

> Lovecraftian Fortran

This is great. "Abandon all hope all ye ..." :)

> perhaps it's the ideas and the concepts that have legacy in computer science, code is just a carrier.

Interesting thought. Only that implementation has consequences on resulting system quality, durability, performance. In other word, it is what we have live (and contend with) in the "flesh" - to follow the genetic analogy ...

[Bluestein]

> or if it's been studied systematically

I'd honestly be very curious to know.-