Fwd: The world is broken, will you help me fix it?
Such grand ideals. Such folly. By the same reasoning, Linux should have already taken over the desktop long ago. The same concepts that prevent this from happening is also slowing progress toward your ideals. Too many chefs, not enough cooks, and no agreement on the recipe in the face of limited resources. Think of it like this:
-
Web OS: This pretty much already exists. We call it a standards compliant web browser. However, compared to what you're asking for, this is more low-level. To bring it up to the level you want, one of the many authentication/authorization packages would have to become a standard part of the browser, but who will set that standard? Will the browser manufacturer comply? It's hard enough to get them all to comply with the current HTML spec, CSS spec, and ES spec. There are so many minor variations on how things work that creating a cross-browser compliant piece of software is tantamount to creating a valuable work of art as it is. Now you're asking for an Authentication spec and an Authorization spec to be added to the list...
-
Node OS: Exactly how would a scripting language replace an entire operating system? By that logic (and I use the term "logic" loosely here), the gnu C++ compiler should have long ago replaced Windows. The reality is that node.js is already a platform on which programs can be written targeting virtually any environment where node works. This means node is probably already about as close as it's going to get to being a virtual OS with the node executable as its command shell.
-
Git OS: Git? Complicated? Really? There's only 7 commands I use regularly in Git: clone, fetch, pull, push, commit, add, & set. Those 7 commands do everything I need to do with git for 95% of what I need. Since I don't need the other commands that often, I don't even bother to remember them, just that their function exists in Git. As far as making a UI is concerned, every major and most minor IDEs have integration plugins for Git. So there are already tons of UIs for Git. I haven't even begun to talk about the non-IDE software for the same functionality. ...And to top it all off, a UI for a command-line tool shouldn't be called an OS.
Having said all of that, I've just got 1 question for you: Why are you posting this in ES Discuss when it has NOTHING AT ALL to do with enhancing or correcting ECMAScript/Javascript?
I mostly agree, except the "folly" part. I was hoping to spark some interest (to answer your question), but it appears there is no interest to be sparked.
See below:
On Fri, Jul 26, 2019 at 4:04 PM Ranando King <kingmph at gmail.com> wrote:
Such grand ideals. Such folly. By the same reasoning, Linux should have already taken over the desktop long ago. The same concepts that prevent this from happening is also slowing progress toward your ideals. Too many chefs, not enough cooks, and no agreement on the recipe in the face of limited resources. Think of it like this:
Yes, I've been focusing more time on the marketing (hence, the logos). It will be a leadership challenge, as much as a technical one.
- Web OS: This pretty much already exists. We call it a standards compliant web browser. However, compared to what you're asking for, this is more low-level. To bring it up to the level you want, one of the many authentication/authorization packages would have to become a standard part of the browser, but who will set that standard? Will the browser manufacturer comply? It's hard enough to get them all to comply with the current HTML spec, CSS spec, and ES spec. There are so many minor variations on how things work that creating a cross-browser compliant piece of software is tantamount to creating a valuable work of art as it is. Now you're asking for an Authentication spec and an Authorization spec to be added to the list...
Building Auth into the browser isn't immediately necessary. In its current state, the webOS is just a client-side View and Router. And with a few node endpoints, it doesn't need to be much more than that. Building real apps will obviously require some persistence. That's where node/git comes in.
- Node OS: Exactly how would a scripting language replace an entire operating system? By that logic (and I use the term "logic" loosely here), the gnu C++ compiler should have long ago replaced Windows. The reality is that node.js is already a platform on which programs can be written targeting virtually any environment where node works. This means node is probably already about as close as it's going to get to being a virtual OS with the node executable as its command shell.
In the beginning, the nodeOS needs to be little more than a simple node start/stop UI. Show me what's running, and let me start/stop/configure it. And maybe at some point, browse the file system, execute commands, browse npm, install/manage deps, create web apps. I would want to make a real IDE at this point. Obviously, if we're going to rewrite all the software, we're going to need a badass IDE, with the best debugging, logging, testing, layout, versioning tools imaginable.
- Git OS: Git? Complicated? Really? There's only 7 commands I use regularly in Git: clone, fetch, pull, push, commit, add, & set. Those 7 commands do everything I need to do with git for 95% of what I need. Since I don't need the other commands that often, I don't even bother to remember them, just that their function exists in Git. As far as making a UI is concerned, every major and most minor IDEs have integration plugins for Git. So there are already tons of UIs for Git. I haven't even begun to talk about the non-IDE software for the same functionality. ...And to top it all off, a UI for a command-line tool shouldn't be called an OS.
Git normally works, but sometimes, ohshitgit.com.
End users need version control. Some apps have a revision history (Google Docs, WordPress). But many apps still lose their history when you exit.
Look at how many color schemes I tried for this logo:
[image: image.png]
Git could seamlessly auto-branch behind the scenes when making changes to permanently track everything. End users need version control, but they shouldn't (won't) be tasked with learning git commands. We need a familiar version control UI to rewind, fast-forward, branch, bookmark/tag, etc, across all apps. Whether it's a document, a 3D model, or text/code. Version control is an essential part of the operating system, and should exist not just as a CLI, but also as a user-friendly UI. This UI should be *consistent across apps, *and the UI should be present within the app. Just trying to clarify that version control is more than a CLI.
In the beginning, the gitOS would just be a part of the nodeOS file explorer. Eventually it could provide an API and UI for apps in the manner described above.
I've used GitKraken, SourceTree, and GitHub desktop. None of them were spectacular experiences.
The first phase of this system is stupid simple. But by itself, it's not enough to excite anyone. Unfortunately, there is no creative energy. Nobody gets excited anymore...
Reread the original email. I mean, honestly, what more could anyone be looking for? It's the whole world on a silver platter. Jesus Fucking Christ.
Having said all of that, I've just got 1 question for you: Why are you
posting this in ES Discuss when it has NOTHING AT ALL to do with enhancing or correcting ECMAScript/Javascript?
I'm desperately seeking help. It's quite clear from my original email:
- I'm drowning in debt and despair
- I have an incredible opportunity to climb out of this hole, fix a broken system, improve the world, and make a fortune
- I need help
I have no following, and this list is a particularly qualified group of techies, the type that might understand my vision.
It's ok, I'll build a free webOS camp that doesn't take 2000 hours, but more like 2 hours to get started. My View class is basically like 50 lines of code. Windows is 50 million. I'll train an army of designers and developers, and we'll fucking run this shit.
As indicated before, and I believe agreed to by you, this has nothing to do with ECMAscript. This is nothing less than spam on a public mailing list. We are not here to receive unsolicited emails with business proposals.
---------- Forwarded message --------- From: Michael Lewis <mike at lew42.com>
Date: Fri, Jul 26, 2019 at 2:34 PM Subject: The world is broken, will you help me fix it? To: <peter at foundersfund.com>, <neil at foundersfund.com>, <ken at foundersfund.com>, <geoff at foundersfund.com>, <napoleon at foundersfund.com>, <
scott at foundersfund.com>, <bruce at foundersfund.com>, <toby at foundersfund.com>, <kate at foundersfund.com>, <brian at foundersfund.com>, <
inquiries at eric-weinstein.net>
Hello world,
I have a vision. It is the Truth.
[image: image.png]
The short version: we need better design, development, data, versioning, collaboration, management, and decision making tools. If we rebuild the software design and development experience from the bottom up, we'll be able to create 10x better software, 10x faster, 10x cheaper.
Michael Lewis (me)@mlew42 twitter.com/mlew42
twitter.com/mlew42/status/1154782897242550274
It's doubtful he will ever see the tweet.
In the "podcast," Peter Thiel and Eric Weinstein make the case that we're not actually doing a good job, despite what some would claim. It's hard to understand, because we're so hyper-competitive. Competitively incompetent. What cannot be argued, is that we can do better.
The Long Vision
We need better decision making software. Email is garbage. We need a tool to identify all options, discuss, rank, plan and execute. Something like Reddit, that results in real productivity, government decisions, and allocating resources. Plan, then build. Design, then develop. CAD and CAM.
There's no single recipe for what a "decision making system" would look like. Maybe it's just a Google doc. Maybe it looks like Reddit. Or maybe it's like WordPress, where you can install plugins? We need a solid platform (like a webOS) where it's relatively easy to build new features.
I could try to speak more to the absolute mess that currently exists. Go to Product Hunt, and start counting. "Ridiculous" hardly explains it.
There are 700 web-based project management platforms listed on capterra.com. And not a single one of them is doing a great job. Every single web app in the world has to build its own user auth, database, user interface, workflows, features, etc. This is sort of like having a separate DMV for every state in the USA. Competitive incompetence.
webOS
Hold onto your pants: We need to take all our software, and rebuild it on a single tech stack. Enter the webOS. It's like Webflow + WordPress + Wikipedia + Google + Facebook + Twitter + LinkedIn + YouTube + Khan Academy + Windows + Android + Gmail + Reddit + ....
Rewriting so many sophisticated systems wouldn't be easy. Unless you realize that they're all basically the same thing: a database of users, posts, and apps.
The webOS has to be a hybrid of open and closed source, with advanced licensing controls. It must be monetized, so people are incentivized to contribute (open source lacks this). There needs to be clear licensing terms, and advanced licensing controls, so developers can license their code on a per user per month basis, for example.
A real Operating System builds all the critical productivity features (like threaded discussions, project management, "email", version control, collaboration, UI, plugins, etc) directly into the operating system itself, so that it's automatically integrated throughout every app.
gitOS
The core of all of this, is the file system. The data. The meat and potatoes. Git is an incredible technology (that powers GitHub), is cross platform, handles versioning, branching, merging, conflict resolution, etc. This technology should be built into every database, document, design, image, email, application, film, etc. Someday soon, *every person who uses a computer will be an expert with versioning. *But git, as a command line tool, is insanely hard to use. We need a git UI. Let's call it the gitOS. This will provide an incredibly powerful feature to every application built with the webOS.
node OS
Ever since I jumped ship as a WordPress developer, I've been waiting for node.js to catch up. It hasn't. JavaScript and node.js should have replaced Windows by now. Show me a WordPress competitor built on node.js, and you'll find a couple (sails, keystone, gatsby?). There are literally a million competitors in this field (SquareSpace, Weebly, Wix, etc). Why can't anyone fix this problem? Node.js is capable of so much more:* nodeOS.*
[image: image.png]
Can you help me?