giwake

game developer, I think?

  • they/them

i make games and music, sometimes.

profile picture by @thewaether!!!

moving to https://bsky.app/profile/giwake.bsky.social


hkr
@hkr

A computer is a tool chest, programs are individual tools.

There are many ways to build a chair. You can use all hand tools to carve and cut the individual components of a chair, hide glue to join them together, and milk paint to apply a finish. You can also use power tools, screws and lacquer to build one. Many choose to buy a flat pack of precut and drilled pieces held together with a combination of screws and dowels. There are collectors of vintage chairs who argue that the old methods are the best, and there are people who buy mass produced monstrosities of leather, steel and foam, chairs that no individual human could build on their own.

There is no correct method of building a chair. Each of the methods above have thousands of variations and approaches. There are thousands of tools that could accomplish the job, thousands of ways a tool could be used to build the chair, and thousands of variants of the same tool but customized to fit a particular need. Some tools can be used for purposes that they weren't designed for with surprising results, both good and bad.

Some chairmakers are set in their ways; over the years they've honed their methodology and have learned every secret of every tool in their chest to accomplish any task in front of them; to suggest a new tool or a different approach is akin to asking them to cut off their hands and learn to use new hands. For them the craft of chairmaking is about honing their abilities over long years; it doesn't matter that their tool chest is full of tools from before world war II, they've been in service for 80 years and can be in service for 80 more.

Other chairmakers are excited by tools themselves and are constantly on the hunt for new tools; planes of various sizes, blades of various sharpness and material, new methodologies for joining pieces of wood together and new stain formulas to try. For them, the craft of chairmaking is about using the best and latest tools and methodologies, finely tweaked for specific scenarios. If they feel their tool in their tool chest is not up to snuff for a particular task, they have no problem buying a new one if it promises a solution to their problem.

In social circles of people who make things, the two types above are usually at odds with each other. The chairmaker set in their way thinks the tool chaser chair maker is spending too much time and money fussing with their tools and buying new ones. The tool chaser thinks the chairmaker set in their way are doing things slow and inefficiently, and if they would only take the time and spend some money to learn the new ways they'd be better off.

A software update that changes the UI and how things get done in the software is akin to Craftsman opening your tool chest, rearranging all your tools, and replacing tools with brand new tools made of different materials and different sizes.

The tool chaser is excited about this; they automatically get updated to the latest and greatest* and they dive right in to learning the ins and outs of their refreshed tool chest. They don't get too attached because they know it's ephemeral and that the next time Craftsman swings around things could be completely different again.

However, the chairmaker who was set in their ways, who has spent years honing their craft and learning the intricacies of every tool in their chest has had their world turned upside down. Instead of being able to start working on their next chair, they have to spend hours, days, weeks learning where their tools went. They're confused why their hole borer now only fits certain types of bits when their old one could fit any. They're upset that they have to use biscuit joints and screws now, because Craftsman took their hide glue and chisels and replaced them with zinc screws and a contraption that can make biscuit joints, but only for specific types of biscuits that have to be purchased from Craftsman direct. They would do anything to get their old tools back, their tool chest re-arranged to what it was before. Craftsman however doesn't want this, so even if the chairmaker takes the time to find their old tools and rearrange things back to where they were, Craftsman will be back to change it when they're least expecting it. Perhaps Craftsman will decide that trying to change it back violates their terms of service and takes the entire tool chest, tools and all, away from the chairmaker as punishment.

In the end the chairmaker set in their ways could learn the new tools, could get used to how things are arranged, but their chairs will still come out differently than they used to, and will likely be inferior to what they used to build. Not because the new tools are necessarily inferior, but because the time taken to hone a craft had to be tossed out. It'll take years for the chairmaker to rehone their craft, but only if Craftsman doesn't decide to update their tool chest again next year.


You must log in to comment.

in reply to @hkr's post:

Except everyone's one rapid release cycle away from changing camps because their brain shatters, pretty much at all times

in the analogy, you marry the Craftsman employee who's fucking with you, i think is how that works

The Craftsman employee then spends their time lowering the brightness of all your gas lamps but insists you're going crazy and that nothing has changed when you bring it up.

so the interesting thing about this is that, like...

it's only sometimes accurate to compare user-facing software to the contents of the tool chest, as seen above. that's certainly accurate with software for creative professionals, people for whom the real creation is music, or visual art, or something of that nature. it's not at all accurate with software that has a function unrelated to creation, such as a chat app.

the sense of disruption is real and important regardless, but when the software we're talking about isn't meant for use in creation, what we're seeing is more like swapping out the actual chair. still disruptive! still a problem! but it's a mistake, we think, to see conflict over that kind of change as being about different ways of making stuff.

this doesn't really change much about what you're saying, it just kinda felt a bit wrong to us and we wanted to clarify.

A tool is a tool, no matter how trivial it seems. People use discord as a tool for community management, professional correspondence, and other tasks. You might not think there's a difference between zinc and brass screws, but to the right people they're worlds apart. We shouldn't approach changes with "oh this is trivial, this won't have any large important effects on someone," especially when we don't know the full breadth of how the tool is used in every shop.

that is fair, it's just... do people really have that kind of relationship with it? the thing you're talking about doesn't cleanly map into either of the maker subcategories you described, as far as we can see

we are not advocating for more change, we fucking hate what Discord has been doing, we just think it's important that the ways we talk about this stuff correspond to reality

The conversation has shifted to beyond discord and my post was more about general UI changes in software rather than just discord. I'm not going to go hunt for examples of people using discord as seriously as photoshop. You're just going to have to believe me when I say that a tool is a tool and there are going to be people who use that tool in a specific way that get hampered by any UI changes.

So like

firstly, yes we're also trying to talk about the general case. we kinda regret mentioning chat apps because it did kinda invite the mention of Discord, which perhaps obscures more than it clarifies, but .... if we're trying to generalize, we have to make sure the specific instances we're basing it on are actually examples of the general pattern, which requires thinking them through not skipping over it

secondly, we don't think you and we are actually disagreeing on anything, and we encourage taking time to go over what we said again. we're in full agreement on everything you said in this latest reply, in particular.

You reconciled this downthread but just to answer the direct question: Yes, Discord is, at least in part, a serious tool for community management, and a significant amount of its UI is for server owners or moderators. This is understood both by users and internally at the company.

It's less explicitly thought of as a tool for non-moderators but the various use cases that fall under that description are generally known and considered.

This is a well thought-out post, thanks for writing it!

Consider Craftsman wanting to make their tools more usable for more chairmakers. They do a ton of interviews and research with newer chairmakers because they are avoiding Craftsman tools in droves. They learn that their designs are considered outdated and unfamiliar, or they struggle to figure out how to use them in the first place. Within the physical tool metaphor, Craftsman can simply make newly designed tools and not modify the old tools. In software, that's not so easy.

There's a maintenance burden for software that doesn't exist for physical tools. Allowing users to stick with an old design can, at worst, double the amount of work any new addition requires until the old design is retired. Freezing the old design so people can keep using it won't win you any friends once you do add a feature that everyone wants, because now the old design users are furious that you're neglecting them by not porting it. And someday you'll want to change the remote service that the tool relies on in a non-backwards-compatible way, which forces your hand.

In that sense I think the analogy of your tools being shuffled around by a Craftsman employee doesn't actually reflect the reality of making and maintaining software that relies on a service. It relies on the assumption that they could've just left your tools as they were and they would've stayed working for you, but for software services the maintenance burden is prohibitively large to enable something approaching that.

I think it does apply pretty well to software that doesn't rely on a service, though. Without compatibility concerns it's much more feasible to maintain older versions of software just enough to enable basic user safety.

Uo until this year I did have multiple versions of my really good photo editors/cad programs running concurrently, because a newer version might have some nice features but I still wanted the plugins that were only made for the old one, or the old one was just more efficient for some process, and neither version had been fiddled with by developers for years because at some point each of the companies involved had started making updates something to be avoided.

They worked fine, great even, for 8-12 years from the day of their last patch to the day that computer and its backups were physically destroyed, and I'd pay quite a lot to get those decade-obsolete unmaintained programs back, but that isn't an option any software company is inclined to offer and I'm not interested in what they do want to sell. I can pretty much guarantee from the number of untouchable early-2000s computers I've seen lurking at the heart of businesses that this is an orders of magnitude more common scenario for software anyone actually depends upon than the tiny range of cases where it'll be a legitimate disaster if something doesn't get an annual UI overhaul; the difference is that Craftsman lacks the means to casually fuck over their customers from afar, not a rationale.

For things like Adobe where there's an existing or obvious design that doesn't rely on a service, and instead a service is introduced unnecessarily, sure. You're probably right that lock in was at least one of the goals with the shift to a service.

But I was talking in the context of software where a service reliance is an unavoidable part of the software purpose itself, like Discord or Cohost. There's no escaping the service reliance and compatibility problem for those.

Pinned Tags