Today I shipped 20 apps and a screensaver, and 18 of the 21 items are small enough to fit on a floppy disk. So I thought it would be fun to make this approach to software development into a manifesto!
I like this manifesto a lot, short and simple, and a cause I admire in a world of hopelessly bloated software! I have an odd request: many Web 1.0-style "small web" sites which embrace technical...
I like this manifesto a lot, short and simple, and a cause I admire in a world of hopelessly bloated software!
I have an odd request: many Web 1.0-style "small web" sites which embrace technical simplicity – and which I therefore think your manifesto might have a decent level of ideological overlap with – still embrace these 88x31 animated Gif icons that were all the rage on Geocities and such a couple of decades ago. Would you be willing to make an "official" 'fits on a floppy' icon in this format? It would be really cool to have one which, when clicked, links to your manifesto!
Even though of course anyone could make one for this purpose, I feel like the creator should get right of first refusal at making such a thing!
Fun idea! I'll put it on my list. Back in the day (~1996/7) I created some of those for the burgeoning emulation scene using, if memory serves, Microsoft GIF Animator....
Fun idea! I'll put it on my list. Back in the day (~1996/7) I created some of those for the burgeoning emulation scene using, if memory serves, Microsoft GIF Animator. https://en.wikipedia.org/wiki/Microsoft_GIF_Animator
I've tangentially thought about this recently. Although a smaller binary does not necessarily mean lower requirements, it is a good eval I feel. With agentic coding lowering software dev costs, I...
I've tangentially thought about this recently. Although a smaller binary does not necessarily mean lower requirements, it is a good eval I feel.
With agentic coding lowering software dev costs, I think open-source developer effort should be directing their agent focus on optimization and minimizing bloat. We've always had this mindset that hardware outpaces software but I think we are at the opposite point now. We will have shortages of hardware, energy, and compute in general, so let's tweak the other side of the equation. Make existing hardware more capable via software.
I'm thinking exploiting efficiency gains from specificity of applications. There will be more software overall but the coding agents would make it easier to execute.
The externalities are also great: less power consumption, less e-waste, more jobs (IT is one function, not THE function of an org), and less tech mega corps (cost barrier of entry for in-house dev is lowered)
Just need better open weight models that can run on consumer hardware and we are free finally
I hope people and/or machines start doing it. It's not that difficult once you decide you want to do it. What I would say is that it should be a relatively easy transfer of knowledge from "which...
I hope people and/or machines start doing it. It's not that difficult once you decide you want to do it. What I would say is that it should be a relatively easy transfer of knowledge from "which package can do that" to "which system frameworks can do that". Anybody or anything can glean that easily from the docs.
The single biggest win with me creating apps to replace some old cruft in my login items was that my computer boots to desktop in about half the time.
It's called specialism. I picked Apple 25 years ago and have developing for their technology ever since. I have made the odd Android and Linux thing, but they never made any money. Given that this...
It's called specialism. I picked Apple 25 years ago and have developing for their technology ever since. I have made the odd Android and Linux thing, but they never made any money. Given that this is my job, that's a pretty important thing.
I didn't mean you - you are free to choose your biases based on any rational, commercial and whatever other reason yourself. I was just surprised to see "a manifesto for small software" - title...
I didn't mean you - you are free to choose your biases based on any rational, commercial and whatever other reason yourself. I was just surprised to see "a manifesto for small software" - title sounding quite general - focusing strictly on apple.
Today I shipped 20 apps and a screensaver, and 18 of the 21 items are small enough to fit on a floppy disk.
So I thought it would be fun to make this approach to software development into a manifesto!
I like this manifesto a lot, short and simple, and a cause I admire in a world of hopelessly bloated software!
I have an odd request: many Web 1.0-style "small web" sites which embrace technical simplicity – and which I therefore think your manifesto might have a decent level of ideological overlap with – still embrace these 88x31 animated Gif icons that were all the rage on Geocities and such a couple of decades ago. Would you be willing to make an "official" 'fits on a floppy' icon in this format? It would be really cool to have one which, when clicked, links to your manifesto!
Even though of course anyone could make one for this purpose, I feel like the creator should get right of first refusal at making such a thing!
Fun idea! I'll put it on my list. Back in the day (~1996/7) I created some of those for the burgeoning emulation scene using, if memory serves, Microsoft GIF Animator. https://en.wikipedia.org/wiki/Microsoft_GIF_Animator
I've tangentially thought about this recently. Although a smaller binary does not necessarily mean lower requirements, it is a good eval I feel.
With agentic coding lowering software dev costs, I think open-source developer effort should be directing their agent focus on optimization and minimizing bloat. We've always had this mindset that hardware outpaces software but I think we are at the opposite point now. We will have shortages of hardware, energy, and compute in general, so let's tweak the other side of the equation. Make existing hardware more capable via software.
I'm thinking exploiting efficiency gains from specificity of applications. There will be more software overall but the coding agents would make it easier to execute.
The externalities are also great: less power consumption, less e-waste, more jobs (IT is one function, not THE function of an org), and less tech mega corps (cost barrier of entry for in-house dev is lowered)
Just need better open weight models that can run on consumer hardware and we are free finally
I hope people and/or machines start doing it. It's not that difficult once you decide you want to do it. What I would say is that it should be a relatively easy transfer of knowledge from "which package can do that" to "which system frameworks can do that". Anybody or anything can glean that easily from the docs.
The single biggest win with me creating apps to replace some old cruft in my login items was that my computer boots to desktop in about half the time.
Why the apple bias? Small Linux or Android apps don't count?
It's called specialism. I picked Apple 25 years ago and have developing for their technology ever since. I have made the odd Android and Linux thing, but they never made any money. Given that this is my job, that's a pretty important thing.
I didn't mean you - you are free to choose your biases based on any rational, commercial and whatever other reason yourself. I was just surprised to see "a manifesto for small software" - title sounding quite general - focusing strictly on apple.
I see, thank you.
Every byte matters! Thanks for this. It's so hard to make something great in a world where pretty much everyone will settle for "good enough".