4 ms·

> integration The way this minilanguage works is that all contacts app can push/pull a `contacts.scroll` in a git repo. That's the database: a plain text file

by breck 2y ago

> integration The way this minilanguage works is that all contacts app can push/pull a `contacts.scroll` in a git repo. That's the database: a plain text file in a git repo.

kkfx 2y ago
Yes, but any other app need to explicitly integrate it, and it the biggest issue in modern software. Let's imaging such project is an Emacs package, you do not need any change from notmuch, mu4e, gnus, meow, ... to use your package, you can just put a bunch of advises, hooks, some functions wrapping some others etc, it's YOU the one who can integrate. In modern software it's the opposite and that's why anything fails because you can convince some apps to support this projects while some others will do not and to communicate you need some import/export solution/bridge and so on. Modern software is a series of walled gardens with an enormous amount of resources to plumb something in between as a result only some projects who get enough popularity have chances of integrations, the other will remain standalone, and most tend to try implementing any possible features becoming monsters of bloats after a bit of time. There is no escape in such model, the modern web is not a valid answer, we simply need the classic OS as a single app/framework but this being incompatible with commercial software (at least it can only be mono-vendor) is a thing nobody in modern industry accept. It's simply about time to admin that software can only be FLOSS and the competition can only be in hw. Meaning things we need to build as products, with raw materials not simply "selling some copies/bit", or "physical economy" vs "paper/finance one".
- breck 2y ago
  > it's YOU the one who can integrate. In modern software it's the opposite Ah I see what you mean. Yes I agree. > It's simply about time to admin that software can only be FLOSS and the competition can only be in hw. Meaning things we need to build as products, with raw materials not simply "selling some copies/bit", or "physical economy" vs "paper/finance one". Agreed!
  - kkfx 2y ago
    Sorry, in general I comment quickly and my English is still poor enough to be not so clear in some cases, I'll try to do better when I see issues, but still sometimes I remain unclear :-) Anyway really dream a modern system like Emacs where you can simply "plug some code" in someone else "app" on your system, how easy it could be extending and also evolving thanks to idea circulation because well, a small "package" is easy to read and understand and being not "another whole app" but just a kind of "plug-in" it's far easier to try. Choosing the current pro-commercial model we have already lost decades of potential evolution only for a bad base design... You can extend that to anything else: a washing machine? Imaging if it's control panel is just and open interface with a common standard communication protocol: you can integrate it in your p.v. and make your wash in self-consumption issueless, it would be so simple most geek can do showing a potential evolutionary path to others. From this to a simple "smart home" where all appliances run one after another to maximize self consumption, than a different design for new machines (let's say instead of one big induction coil many little to being able to tune the load) and so on, with innovation happening at almost all levels like in classic FLOSS but IRL, not just in some geek projects.

by breck 2y ago

> integration The way this minilanguage works is that all contacts app can push/pull a `contacts.scroll` in a git repo. That's the database: a plain text file in a git repo.

kkfx 2y ago

Yes, but any other app need to explicitly integrate it, and it the biggest issue in modern software. Let's imaging such project is an Emacs package, you do not need any change from notmuch, mu4e, gnus, meow, ... to use your package, you can just put a bunch of advises, hooks, some functions wrapping some others etc, it's YOU the one who can integrate. In modern software it's the opposite and that's why anything fails because you can convince some apps to support this projects while some others will do not and to communicate you need some import/export solution/bridge and so on. Modern software is a series of walled gardens with an enormous amount of resources to plumb something in between as a result only some projects who get enough popularity have chances of integrations, the other will remain standalone, and most tend to try implementing any possible features becoming monsters of bloats after a bit of time. There is no escape in such model, the modern web is not a valid answer, we simply need the classic OS as a single app/framework but this being incompatible with commercial software (at least it can only be mono-vendor) is a thing nobody in modern industry accept. It's simply about time to admin that software can only be FLOSS and the competition can only be in hw. Meaning things we need to build as products, with raw materials not simply "selling some copies/bit", or "physical economy" vs "paper/finance one".

breck 2y ago
> it's YOU the one who can integrate. In modern software it's the opposite Ah I see what you mean. Yes I agree. > It's simply about time to admin that software can only be FLOSS and the competition can only be in hw. Meaning things we need to build as products, with raw materials not simply "selling some copies/bit", or "physical economy" vs "paper/finance one". Agreed!
- kkfx 2y ago
  Sorry, in general I comment quickly and my English is still poor enough to be not so clear in some cases, I'll try to do better when I see issues, but still sometimes I remain unclear :-) Anyway really dream a modern system like Emacs where you can simply "plug some code" in someone else "app" on your system, how easy it could be extending and also evolving thanks to idea circulation because well, a small "package" is easy to read and understand and being not "another whole app" but just a kind of "plug-in" it's far easier to try. Choosing the current pro-commercial model we have already lost decades of potential evolution only for a bad base design... You can extend that to anything else: a washing machine? Imaging if it's control panel is just and open interface with a common standard communication protocol: you can integrate it in your p.v. and make your wash in self-consumption issueless, it would be so simple most geek can do showing a potential evolutionary path to others. From this to a simple "smart home" where all appliances run one after another to maximize self consumption, than a different design for new machines (let's say instead of one big induction coil many little to being able to tune the load) and so on, with innovation happening at almost all levels like in classic FLOSS but IRL, not just in some geek projects.