3 ms·

«If you believe that the number of people wanting a file manager and office suite is decreasing, what do you think the desktop environment actually will consist

«If you believe that the number of people wanting a file manager and office suite is decreasing, what do you think the desktop environment actually will consist of?» Look at ChromeOS and tablets. But, if you want to skip to a decentralized, user innovation and privacy-first future, look at KDE Nepomuk (which was an ambitious decentralized semantic desktop that collapsed), MS Graph (which is a centralized semantic system that probably won't), Plan 9, Nix, Qubes. «Isn't that precisely what the system is??» It's not really networked, in fact it's going backwards. X lets you run distrbibuted application on a fine level, as do other decentralized ideas like WS and HATEOAS, but we're reverting to a thin client model where the user is dependant on the browser and ???. But getting to the fine level requires more isolation, data & identify consciousness. «What does this even mean? What is a "web based conversation" in terms of the user interface??» Think of an inherently local, yet potentially complex control; a volume control. I can simply use it to control volume, but I might also want to mix volume with networked devices, which spirals into complexity, especially as we get into spatial audio. It's all possible on a Linux desktop, but doing it takes you into open ended complexity, compounded by the fact there are so many versions and combinations of software. A post-90s desktop could add a new control to the context menu which opened a live discussion about that version of that control and combination of desktop options.

72deluxe 4y ago
You are using lots of words that do not mean a lot to me. It is indistinguishable from marketing-speak. "decentralized, user innovation and privacy-first future" Why would you want a decentralized desktop? Are you somehow referring to a syncing of files to a new desktop when you use a different machine? If so, see Windows Profiles, VNC, RDP or NoMachine. If you're somehow referring to something different, I cannot see the use-case for it. The desktop follows the concept of a desk surface in real life. When I go to work elsewhere, I have a new desk but I can put familiar things on it. I don't need the things I put on it to live in the ether above me. "Decentralized" is mubmojumbo - the data has to live somewhere. It can certainly have copies on different machines and perhaps that's what you mean by "decentralized". My desktop under RedHat 6.2 was "privacy-first". It was even this way on my Windows 3.11 machine... And I am able to use a volume control fine - I've been able to use one for 40 years without needing it to be "networked". I even manage to pipe audio from my Windows PC to Pulse on a Linux laptop so I can use Presonus Studio One remotely. It seems you are inventing scenarios that are looking for problems to solve. This seems endemic in the modern software world, in an attempt to "stand out" from the crowd, in my opinion. I am honestly very confused by your response.
- davidy123 4y ago
  I'm glad you got your volume control networked. I kind of gave up on that, since everything kept changing with every release. "Decentralized" in the developing sense, for distributed consumer, business and individual services, is about identity and trust. In contrast to "federated," the services do not have to have a specific agreements between them, they just have to adhere to standards with identity and privacy baked in. As an example, this was how OpenID Connect was supposed to work, but for simplicity, most logins only permit a few providers. The modern standards around these ideas, if you really want to explore, are Verifiable Credentials and Decentralized Identifiers, sometimes collectively referred to as "self-sovereign identity."[1] These standards are designed to enable deep exchange of information between any services (with user permission), while preserving privacy. Each participating service just gets a random identifier, with no personal information. The idea of "files," where all the information required for a "document," is disappearing for many purposes. A document is a moment in time. Instead, information is an aggregate panel from distributed (decentralized) information. This is evident from how federated identity works today, and how specialized services people access from a web site work with increasing ease and utility, though today it's about pushing the envelope with obscurity and plunder. If you only work in files, and you "even manage" to do complex things with your volume control, and you're content, then you're basically part of the problem. You have a 1990s vision of a desktop, which you expect every other user to climb to identically, even though many of the steps constantly shift for arbitrary reasons. The steps you used to get your volume control to work that way were about a moment in time configuration, built on all kinds of underlying knowledge that is probably deeply entrenched in your life, and it should have been simple. Do you expect every user to go through that, or does it take a special user? Consumer solutions, based on the web and closed operating systems and ecosystems, are providing these features, based on increasingly sophisticated ideas of "what most people want," and based on deep integration with distributed (but not decentralized) web services. These approaches have risks, but also huge advantages, but they truly need more consciousness from end users. If Linux is going to be relevant to more than a small cohort of desktop users, and a Linux desktop is meaningful as a true open, comprehensible platform rather than a way to access a browser which does black magic, it has to highlight these features. These features are available in completely coherent forms (see Matter standards, IoT, Solid, VC, DID, etc) to allow services like getting a drivers license or booking a complex international trip online in minutes without relying on a black box web site, but because so many people can only see a 1990s desktop, efforts behind Linux Desktops are basically a dog chasing after a 1990s car. KDE was way ahead of the curve with Nepomuk. It provided a local semantic database of fine information, built around standards with networked schemas that support entailment. So rather than typing documents from memory, local and distributed information is integrated. But it was too ambitious, and most technical users got distracted by the money to be made selling pet food online or the challenge of getting their volume control networked. So, today, all the centralized providers, and open projects like Wikidata, are forming around these ideas. But Linux desktops are too busy trying to re-add live backgrounds to pay much attention. Finally, notice I am only now mentioning "blockchain?" Blockchain can be used in some schemes, for example to host public keys and revocations, but this is not a side effect of blockchain hype, it's about where everything is headed, if computers are more than fancy typewriters. 1. https://en.wikipedia.org/wiki/Self-sovereign_identity https://en.wikipedia.org/wiki/Self-sovereign_identity

72deluxe 4y ago

You are using lots of words that do not mean a lot to me. It is indistinguishable from marketing-speak. "decentralized, user innovation and privacy-first future" Why would you want a decentralized desktop? Are you somehow referring to a syncing of files to a new desktop when you use a different machine? If so, see Windows Profiles, VNC, RDP or NoMachine. If you're somehow referring to something different, I cannot see the use-case for it. The desktop follows the concept of a desk surface in real life. When I go to work elsewhere, I have a new desk but I can put familiar things on it. I don't need the things I put on it to live in the ether above me. "Decentralized" is mubmojumbo - the data has to live somewhere. It can certainly have copies on different machines and perhaps that's what you mean by "decentralized". My desktop under RedHat 6.2 was "privacy-first". It was even this way on my Windows 3.11 machine... And I am able to use a volume control fine - I've been able to use one for 40 years without needing it to be "networked". I even manage to pipe audio from my Windows PC to Pulse on a Linux laptop so I can use Presonus Studio One remotely. It seems you are inventing scenarios that are looking for problems to solve. This seems endemic in the modern software world, in an attempt to "stand out" from the crowd, in my opinion. I am honestly very confused by your response.

davidy123 4y ago
I'm glad you got your volume control networked. I kind of gave up on that, since everything kept changing with every release. "Decentralized" in the developing sense, for distributed consumer, business and individual services, is about identity and trust. In contrast to "federated," the services do not have to have a specific agreements between them, they just have to adhere to standards with identity and privacy baked in. As an example, this was how OpenID Connect was supposed to work, but for simplicity, most logins only permit a few providers. The modern standards around these ideas, if you really want to explore, are Verifiable Credentials and Decentralized Identifiers, sometimes collectively referred to as "self-sovereign identity."[1] These standards are designed to enable deep exchange of information between any services (with user permission), while preserving privacy. Each participating service just gets a random identifier, with no personal information. The idea of "files," where all the information required for a "document," is disappearing for many purposes. A document is a moment in time. Instead, information is an aggregate panel from distributed (decentralized) information. This is evident from how federated identity works today, and how specialized services people access from a web site work with increasing ease and utility, though today it's about pushing the envelope with obscurity and plunder. If you only work in files, and you "even manage" to do complex things with your volume control, and you're content, then you're basically part of the problem. You have a 1990s vision of a desktop, which you expect every other user to climb to identically, even though many of the steps constantly shift for arbitrary reasons. The steps you used to get your volume control to work that way were about a moment in time configuration, built on all kinds of underlying knowledge that is probably deeply entrenched in your life, and it should have been simple. Do you expect every user to go through that, or does it take a special user? Consumer solutions, based on the web and closed operating systems and ecosystems, are providing these features, based on increasingly sophisticated ideas of "what most people want," and based on deep integration with distributed (but not decentralized) web services. These approaches have risks, but also huge advantages, but they truly need more consciousness from end users. If Linux is going to be relevant to more than a small cohort of desktop users, and a Linux desktop is meaningful as a true open, comprehensible platform rather than a way to access a browser which does black magic, it has to highlight these features. These features are available in completely coherent forms (see Matter standards, IoT, Solid, VC, DID, etc) to allow services like getting a drivers license or booking a complex international trip online in minutes without relying on a black box web site, but because so many people can only see a 1990s desktop, efforts behind Linux Desktops are basically a dog chasing after a 1990s car. KDE was way ahead of the curve with Nepomuk. It provided a local semantic database of fine information, built around standards with networked schemas that support entailment. So rather than typing documents from memory, local and distributed information is integrated. But it was too ambitious, and most technical users got distracted by the money to be made selling pet food online or the challenge of getting their volume control networked. So, today, all the centralized providers, and open projects like Wikidata, are forming around these ideas. But Linux desktops are too busy trying to re-add live backgrounds to pay much attention. Finally, notice I am only now mentioning "blockchain?" Blockchain can be used in some schemes, for example to host public keys and revocations, but this is not a side effect of blockchain hype, it's about where everything is headed, if computers are more than fancy typewriters. 1. https://en.wikipedia.org/wiki/Self-sovereign_identity https://en.wikipedia.org/wiki/Self-sovereign_identity