Hey All,
It has been a bit since a GEOS post.
I went into this with a clear plan: spend real time digging into the classic GEOS library, see how the old programs behave on modern hardware (especially the Ultimate 64), and write up practical notes on what still works, what needs workarounds, and how the expanded capabilities of current machiines change the experience.
That plan didn’t survive contact with the actual software.
The core issue is that a lot of the old GEOS programs were written for a very constrained environment, and running them cleanly often means deliberately throttling what the Ultimate can do. Features that make the machine far more usable as something you could actually sit down and work with—better storage access, more RAM without constant disk thrashing, cleaner multi-device setups, the ability to keep more things available at once—end up being limited or turned off so the old code doesn’t choke. You end up choosing between “run this old program the way it expects” and “have a machine that feels less like it’s fighting you.” For anyone hoping to treat the Ultimate as a single coherent environment rather than a collection of compatibility modes, that’s a real friction point.
On top of that, the practical side of getting programs out of the archives and into a usable state is more annoying than I expected. A lot of the GEOS material sits in formats that aren’t straightforward to work with today. Extracting, converting, and getting things into a state where they actually load and run reliably takes more steps than it should, and some of it still carries DRM or copy-protection leftovers that get in the way. None of this is surprising if you’ve spent time in the old GEOS scene, but it does make the “just review the software and document how it works now” approach slower and less rewarding than it looked from the outside.
The end result is that the old software is interesting as historical material, but it actively works against treating the Ultimate as a more capable, all-in-one machine. The compromises required to keep the classic programs happy reduce the advantages of the modern hardware instead of building on them. For casual exploration or nostalgia it’s fine. For anyone thinking about the Ultimate as something you’d use regularly without constantly switching contexts or fighting the software, the old GEOS library starts to look more like a constraint than a feature set.
Where This Actually Led
The next logical step, once the limitations became clear, was to stop trying to force the old software into a role it wasn’t designed for and instead look at rewriting something GEOS-like that is built for the Ultimate from the start. The machine already has more memory, better storage options, and direct access to its own command level and hardware features. The Cyberpunk version coming out only widens that gap with faster processing and still more memory. At that point, carrying forward the old C64 constraints stops making sense. A new OS that isn’t burdened by those requirements can actually use what the hardware offers instead of working around it.
I’ve started down that path with SutehkOS (vanity name but hey I may be the only one who ever uses it so why not). The existing modern OS projects I’ve looked at have clearly had a lot of work put into them, but most of them remain heavily text and command-line driven. That is a different experience from the original GEOS approach of just clicking around and being able to get things done without memorizing key combinations or remembering which function key does what in which program. That style of interaction is not a great user experience if the goal is something you can sit down and use without constant context-switching in your head. The ones that have tried to extend GEOS's capabilities while clever and fun more often than freeze and choke under some ultimate settings and rely too heavily on the old GEOS and the four device swapping architecture. They also still need an original GEOS install to build on which makes the premise of "just using" a new Commodore Ultimate a bit of a burden. Trust me I understand that the modern adders were built for existing aging hardware before the Ultimates came on the seen; so certainly this is not a criticism of them in fact they are awesome and fun to play with.
Those realizations though pushed the focus toward applications and desk accessories that live inside the OS, the same way GEOS handled them, but built to be usable today on the Commodore Ultimate you buy take out of the box and want to use for more than playing games. Less clunky, still carrying some of the nostalgia, and designed around the capabilities that are actually available now. The classic GEOS application library, along with the various augmented and shareware programs that came later, turn out to be a strong idea base for what the OS itself and the new applications should be able to do. Those old programs show the kinds of tasks people wanted to perform; the difference is building the tools so they fit the current hardware instead of the other way around.
That is the direction I’m heading with GEOS-related work. There will still be some discussion of the old software here —it isn’t going away—but the practical path forward is modern Commodore: more memory, faster processing, better mass storage options, and a direct interface with what the Ultimate can actually do at the hardware and command level. The goal is an environment that feels familiar in the GEOS sense without being limited by the assumptions that made sense in 1986.
Comments always welcome and encouraged!
-Sutehk