Lina's Corner

Lina and Lana are alters of a chaotic DID system from New Zealand.

[lina-the-alter]

Lina Here. I am an alter of @thehyperspacesystem and this is my side blog. We decided I should have my own blog as I talk mostly about things that are off topic for our main blog.

I am:

The body's age

Schrodinger's woman / trans woman gendered.

ace/aro, like seriously don't hit on me. I won't even notice you exist.

I have that general "hacker girl" thing going on, spending much of my time programming or tinkering with the computer.

I like Programming, Linux and Open Source Software

I enjoy working on projects with others

Only really fronting during autumn and winter (NZ/UTC+12 time)

Trying to make sense of the fact that I am an alter of a DID system

another alter, Lana who is quite similar to me also posts here.

I sometimes make personal projects related to managing our system's DID/OSDD and post about the process of making them. These are generally NOT made available due to unreasonable user expectations. Where I do link to my projects on github, they are source code that you will have to compile yourself for your operating system. The right way to think of this is "I am excitedly sharing my hobby with you" NOT "I can expect commercial software company level technical support and ease of use from this." My target audience is people who at know enough about computers and programming to at least build a program from source code. There will be bugs and crashes. My code is made out of open source bits and pieces and a LOT of duct tape. It should not be taken as an example of good programming.

[parabo1ic]

why dont linux distros make the kernel a microkernel with a translation layer man i want my microkernel os

[myzanilred]

Actually, this is a fun question that I'm sure many have asked at some point, in some form or another.

At that point, it wouldn't exactly be Linux anymore. The thing that makes a Linux distro a Linux distro, is the the Linux kernel. Not to say there haven't been/aren't attempts to try and replace it with something else.

To start with, the Linux kernel is incredibly complex. The many interfaces and subsystems it provides make building a complete compatibility layer extremely difficult. Especially when it comes to features like namespaces and cgroups. This is why the original Window's subsystem for Linux (which did use a sort of compatibility layer) never quite took off, and why FreeBSD's Linux binary support is still incomplete. The speed at which the Linux kernel changes doesn't help things either.

Additionally, in order for another kernel, of any kind, to effectively take the place of the Linux kernel, in the real world, it would need to have decent parity with Linux's hardware support. This means that the device drivers that allow Linux to be so usable would need to be ported to, or rewritten for, this new system. For a microkernel architecture, this would require many of the driver to be redesigned as userland services or modules, that can function as standalone units. This would likely be a large portion of the work, and the reason that so few other kernels are so widely used, as specialists on specific hardware who write many of those drivers tend to be few and far between, and are likely already working on Linux, or another existing kernel. This is why Hurd still hasn't changed much. It simply lacks the broad support from devs who have both the time and domain-specific knowledge to really move it forwards. It's the eternal Catch-22 of OS development: If your kernel doesn't support real hardware, why would anyone use it, and if no one's using it, who's going to want to work on it?

Simply put, OSs are big, complex beasts that exists outside of what most programmers are familiar, or comfortable with, and making software for one fit seamlessly into another is a special type of hell (ask any Wine dev). It is very much a case of, "If it were that easy, someone would have done it by now."

[lina-the-alter]

You may be able to make a home made kernel that works on your specific hardware with enough time and effort assuming you don't have any hardware with closed specs like nvidia graphics, but getting it to work on literally any other computer would be that much work again.

One of our other coding inclined alters made a simple OS for intel gallieo development boards (it seems to have been done for fun, not as a serious project). I doubt it would even boot on any other hardware.

The linux kernel runs everything from the previously mentioned galleo board, to your phone, your tablet and the servers this website runs on to the computer i'm typing this on.

At this point Linux is THE kernel, it has become an unstoppable force. There is a reason that android and chrome os used the Linux kernel and didn't try to develop their own kernel and that's with Google's resources.

[lina-the-alter]

[Image ID: Text that says:

People Thought I Was Dead

The first reaction people had was to think something bad had happened to me, some of them even thought I was dead. Then something socially curious happened: everybody started speaking to each other on Facebook and WhatsApp to try to figure out what was wrong. Some of them even contacted my family multiple times. They all had my phone number, email address and other ways of contacting me. However, none of them did. It was like I had exited the matrix, and was living in another reality.

Trustworthiness

I was told that I couldn’t be trusted since people can’t check online what I’m doing when I’m not around.

Yeah, you read that right.

Society has been brainwashed to believe that privacy is something criminal. Sorry to disappoint, but privacy is a basic fundament of freedom and democracy. That’s why the voting system is anonymous [1]. When people tell you “If you have nothing to hide, you have nothing to fear”, what they really mean is “democracy is overrated, get over it”.

[1] Privacy: you know who I am but not what I do. Anonymity: You know what I do but not who I am. The voting system uses both, privacy when you go vote, anonymity when they count the results. /end Image ID] This is from an article about a computer security professional's experiences giving up social media and smartphones, which we are now considering doing. Although we will likely retain some use of tumblr on a computer only. You can read the full article here:

I’ve now lived without social media and without a smartphone for the past 3 years. And I’m still alive. A lot of people have asked me how I

When you look at smart phones and social media from their perspective, you see how addicted everyone is.

[lina-the-alter]

It seems that pretty much any program on a computer these days will pop a dialog asking for your credit card no so it can charge you a subscription for basic features

like yes I know developers of commercial software have to eat and pay rent, but just let me buy the fucking program instead of charging a monthly fee.

[lina-the-alter]

I'm furious about google working to block sideloading. You are going to be able to enable it again, but there is a 24 hour delay before you can actually enable it. The playbook is they make it very difficult for people to do so that that people don't do it and then use their telemetry data to justify removing the ability completely because nobody used it.

unrestricted sideloading and being able to run my own code without jumping through any code is literally the ONLY thing that android offers that iOS doesn't. There is literally no other reason to choose android, it's an otherwise inferior operating system.

This is not about security, it's about making sure that you don't download a youtube ad blocker.

I see a world coming where I am going to have to get programming language tools like compilers on the darknet and will have to use an exploit to get anything not signed by Google/Microsoft/Apple to run. Linux will probably eventually be made illegal to possess without being an "approved IT professional" and if you can get ahold of it, you will need a bootloader exploit to get it to boot on your locked-bootloader device.

[lina-the-alter]

Trying to adjust to the body's new disabilities, we can no longer type with a keyboard. We use windows on screen keyboard and a trackball.

Do i try to do my programming hobbies with assistive technology? or is this another hobby to grieve the loss of?

[lina-the-alter]

I have become disillusioned with mobile app development and I am not going to waste my energy on it. You are basically forced to develop a cloud-based application.

The only way to get certain permissions on android and iOS is to take google and apple to court and win, which is obviously only possible for massive tech companies.

There was a very forward thinking article which predicted all this called "The coming war on general purpose computing". I see a world coming where you will have to have approval from a big company to even get your hands on a compiler and every program has to be signed to run. Of course hackers will still use linux, even if they have to use exploits to bypass locked bootloaders.

[lina-the-alter]

Life cycle of a CDD app that i have observed:

The developer develops the app and releases it

The community is excited and embraces it

the community starts demanding more and more features

The developer can't say no because most CDD app developers have a CDD themselves and traumatized people are bad at saying no

More features are implemented and the app grows increasing the amount of work the developer has to do until it completely consumes their life

The community launches harassment campaigns targeting the developer over some perceived failing

The developer issues a statement defending themselves

The community gets angrier.

The developer gets sick of it and shuts down the app

Someone else makes a new app to take the place of the app that shut down starting the cycle again.

TL;DR if you want apps to help manage your CDD, stop being toxic. Or maybe start a non profit organization and hire some professionals to build it. You will quickly discover how much everything involved actually costs and how much work is involved.

[lina-the-alter]

I feel kind of dirty. I just finished drafting a Terms of service for a web app i'm developing. I used an open source template and modified it to suit my app.

I know i should get a lawyer to look at it, but getting a lawyer to draft a TOS costs approximately $2,000 so fuck that.

My app does not collect any personally identifiable information so I should be fairly safe.

It's a sad world we live in that giving a program away for free requires a legal disclaimer drafted in a specific way.

[lina-the-alter]

So AI coding assistants are massive time wasters. I asked AI to scaffold out some routes with Vue and Ionic Framework in a web app and whatever it's done doesn't work. I am getting tons of errors.

googling the error hasn't helped. I've been trying to debug what the AI has done for an hour and half. I'm about to nuke the entire devcontainer and just start again and do everything by hand because this is more trouble than it's worth.

[lina-the-alter]

Maybe there's something to this after all Just for fun I put this

in the prompt window and then copied pasted all the errors from the browser console. I expected it to fuck things up even worse, but after two tries it got it working and told me what the error actually was.

Don't worry, i'm not going to become a vibe coder, I think vibe coding is an insane practice.

[lina-the-alter]

So AI coding assistants are massive time wasters. I asked AI to scaffold out some routes with Vue and Ionic Framework in a web app and whatever it's done doesn't work. I am getting tons of errors.

googling the error hasn't helped. I've been trying to debug what the AI has done for an hour and half. I'm about to nuke the entire devcontainer and just start again and do everything by hand because this is more trouble than it's worth.

[lina-the-alter]

I am trying to sort out my feelings about Ai coding assistants. I have placed the rest below a cut because this is quite long.

VS Code now tries to complete the code you are typing using AI and I can't figure out how to turn it off. You can press tab to accept it's suggestions or just ignore it and keep typing, it is sometimes confusing what is the actual contents of the file and what is an AI suggestion as this is indicated by making the AI suggested text dimmed.

I do accept suggestions occasionally for things like config files, but most of the time it doesn't seem to understand what I am trying to do.

In javascript and typescript it actually works reasonably well.

I think it would choke on C++ or rust.

I am trying to decide if I feel that using this is OK or not.

I am for most part anti AI and my position generally is that AI is dangerous and should be completely banned world wide, except for certain permitted safe uses (eg enemies in video games). There is a book that explains the reasons for this better than I can called If anyone builds it, everyone dies by Eliezer Yudkowsky.

Before we even get into the issues posed by superinteligence, present day AI systems that actually exist learn things like racisim and discrimination from humans and amplify them.

For a practical example, supermarkets in new zealand have started using facial recognition and computer vision systems that try to detect shoplifters. This system had an issue where a number of people who had done nothing wrong were banned from their local supermarket because the computer said they were shop lifting.

The thing these people had in common was that they were brown.

There is no way to prevent an AI system from learning patterns that are present in the training data. You cannot say "learn this, but don't learn that.". These systems are trained on white men and malfunction when presented with faces of other demographics.

This is the kind of threat that AI currently poses. It already does things like decide who gets a job.

[lina-the-alter]

For laughs I asked co-pilot AI how big the .exe would be if I statically linked every library in existance to a C++ program. It listed common libraries that totaled about 8Gb and used that as an example. The correct answer is that code segments in windows.exe's (even 64 bit ones) are limited to 2 Gigabytes.

[lina-the-alter]

[This advice is targeted at beginner developers]

Advice for people making DID/OSDD management apps:

Do not use online website builders or online app builders, they will hold you back.

Use an actual development environment installed on your actual computer. If you don't want to do this, something like GitHub codespaces could also work.

AI coding assistants can be helpful, but you must have enough programming knowledge to know whether to accept what the AI wants to do to use them effectively. AI is often confidently wrong.

Store your code on GitHub and use proper git version control. Private repos are available if you would rather keep your code private.

Use one of the big cloud providers for hosting, but be sure that you understand how you are billed and choose resources accordingly. You can end up with a huge bill you can't pay if you set things up wrong.

Most cloud providers have a certain amount of resources that you get for free. It is possible to design things so that you stay within these free limits

Protect your REST APIs with API keys that you don't share, this will help prevent you from getting a big bill.

NoSQL / document type databases are often cheaper than things like MySql or SQL server

consider just storing data locally, no server costs that way!

Javascript or Typescript based app frameworks are probably the easiest for a beginner to get started with.

If building for Android back up your release signing key and store it securely. You will not be able to update your app without it.

If building a web site or a rest API for an app, learn about the OWASP Top Ten security vulnerabilities and make sure that you do not have them in your app.

Allow people to export their data to JSON, data sovereignty is important and is a legal requirement if your app can be downloaded or accessed in EU countries.

Do not check secrets like database passwords or keys into git source control.

If your app grows beyond a few users, at some point you will need to talk to actual lawyer about the legalities of your operations. Terms of service and things like that usually need the help of a lawyer.

Unless you are building a completely offline app that is deployed through sideloading, you will probably have to spend some money on this

[lina-the-alter]

I take really big issue with the use of chromebooks in schools. This is reason why computer literacy is actually falling. Chromebooks are not general purpose. They are basically a web browser appliance. You cannot install the linux developer environment on an enterprise enrolled chromebook which the ones in schools are.

BIg Tech has a vested interest in people not understanding how computers work. They don't want you to think about where you data is stored or how it is processed. They want you to not understand the difference between what is local and what is remote. They want you to not understand that The Cloud(tm) is someone else's computer. They want you to hand over your data without thinking so they can sell it, or mine it to target advertising to you. They don't want you to think critically about software. Everything that happens on a computer is that way because someone designed it to work that way. A common manipulation developers of proprietary software use is to create a false choice in their software and then present that false choice as a fact of how computers work. The privacy settings in windows work like this. Saving files in onedrive by default and you having to go out of your way to disable it is another example. They can then profit from people's lack of computer literacy by charging for additional cloud space when most people probably don't actually need it.

Dark patterns should be taught in primary school along with the basic idea that computers are programmable devices and that someone designed it to work that way when a piece of software does something. Kids should also be taught about jails and jailbreaking and why devices that only run code signed by the manufacturer like iPhones are bad.

[lina-the-alter]

I had a reminder that The Cloud(tm) is actually just someone else's computer.

I noticed that the debit card on my github account was my old, expired one. So I updated it to my current one. Of course because it is the day before pay day the authorization hold failed. So now my account is now locked and i can't use codespaces even though my usage of them is within the free quota, which is why I never noticed I still had an expired card on my account.

While some services don't lock you out the instant a payment fails, other do. Most give you a few days to contact your bank and fix the problem. But github suspends all access to paid features the instant a payment fails. Keep this in mind if you use their CI workflows or Github Actions, because these will no longer run if for some reason your card doesn't go through, it doesn't matter if you have remaining free quota. A failed payment reduces your quota to zero.

If you are using only free quota, it's best to just never add a card to your account so this can't happen.

[lina-the-alter]

I was co-conscious with Maria and the things I saw on her dash pissed me off.

With the posts about the simply plural shutdown on her dash, it would seem that the general attitude of the community when a developer releases a DID/OSDD management app is "you must provide technical support and updates for this forever, oh yeah it should be free too." I feel very validated in NOT releasing anything I experiment with making for my system's own use if that's the attitude of the community.

On the other side of this, pwDID need to start thinking critically about the digital tools they use to manage their system. Ask these questions:

Who made this, are they likely to stick around?

Is this alpha or beta software, or has it been around for a while?

Does this have commercial backing, or is this software someone's hobby?

Is this open source? If the developer disappears, can I take the source code and keep patching / updating it myself (or pay a developer to do it)

How is my data stored and processed? Is data saved locally or in the cloud? If it's stored in the cloud, can I run my own server or export my data? Is the file format of the exported or locally stored data easy to understand? Is there another program that can read it, can I write code to read it? or is it proprietary. If it's stored locally, can I access it to back it up or write code to read it? or is it stored in a location protected by the operating system that I can't access?

As an example, Microsoft Onenote fails these requirements because onenote tries to lock your notes into Microsoft's servers. You can export with difficulty, but the file format is proprietary and extremely complex. A spec is available, but it would take me 2+ years to write and test a proper library to read it.

Obsidian passes most of these requirements, we would rather it was open source. But the files are plain markdown files. So if Obsidian went away, it wouldn't be too difficult for us to either just use a text editor or make our own application that works like obsidian did.

In general people have been conditioned not to think about this kind of thing and have also been conditioned to expect things to be free. This serves the interests of tech companies who make money from controlling and selling our data [1]

Some counties have passed laws to help with this situation, for example in EU regions being able to export data to a common format like JSON or CSV is a legal requirement.

Mobile operating systems also nudge developers to do things in certain ways. For example, on android saving data to a server via a REST API is easy. But saving data to a folder of the user’s choice is surprisingly difficult (API level 36), requiring you to present a UI flow where the user grants your app permission to access the folder they want to save to. You then have to access the granted folder with a different API than the normal filesystem APIs in your programming language. In some app frameworks, the APIs to do this aren’t even available.

We are told this is for security, but it is actually about platform holders maintaining dominance and control. App developers don’t want to deal with this, so they store data on their servers which is exactly what the platform holders want.

User expectations also come into this as we expect to be able to sync between devices without giving any thought to how this is achieved.

[1] I am NOT intending to imply that SP was selling people's data. I am rather talking about user expectations and how platforms restrictions shape developer choices.