draft0 - a shared blog by just some people

Go To Navigation
Show/Hide Navigation

Entries tagged 'cat:incomplete'

This category is for posts that are not quite finished, yet. This content may or may not make sense depending on how severly a post is not finished, yet. You may find here abandoned projects, outdated information, empty posts, posts without their most crucial part, or generally incomplete p

Old Cherry Keyboard Repair

I found a Cherry G81-1000 in the scrap recently. It's one of those old 101 key models with a metal board and oldish mechanical switched. Model M fans would probably disagree. But it is comparible with an IBM Model M in many ways. Maybe it's becasue Model Ms are still produced that, from what I saw, Cherry G81-1000s are offered for more on eBay. So they do seem to be of interest to other vintage computer fans, not just me.

To test it I first had to replace the cable. I decided for a cheap, black audio cable with two 5-pin DIN plugs. I say audio cable because hat's what it was advertised as. It has no shielding at all. Anyway, just becausew they used proper shielding back then doesn't mean I need it. Because I'm not involved with the computer keyboard fan bubble I had no idea where to find the pinout on ther PCB. Web search engines were of no help. But since a stub of the original cable was still on it, I knew which colour of the original cable goes where. I then checkerd three other (a bit newer) Cherry keybards and found them to use the same colours. I tried the same pinout and it works. I'll attach a diagram just in case anybody needs to do the same thing.

The keybnoard is generally working. But typing does not feel nice and some keys stopped working after a while. From their arrangement it is pretty obvious that one trace on the board must be broken. Here starts the bigger problem. Because of the way the board is constructed, the most of the traces aren't accessible. They are covered by a metal shet from the back and the switches from the front, which are connected to the metal sheet. I know how to repair traces of a keyboard matrix on such a flexible PCB. But if I can't reach them without destroying either be PCB or the switches, that knowledge is not worth anything.

During my research on those old boards I also found out why the switches look so different from MX switches and why they feel so weird. They aren't really mechanical switches, but rather Cherry MY switches, which use indivisual membranes inside. This explains why typing on it feels so bad. I did not find a way to repait the trace, yet. But knowing now that nor cleaning nor using the switches for a while will make them feel better, it does not feel worth putting more time into this.

Comment via email
Alternative Operating System: Essence
This entry is referencing the entry 'Alternative Operating Systems'.

Essence

This is one I'm continuasly disappointed to not have been able yet to get running on real hardware. I like what I've seen. But I can't get it to boot, as do others. I don't know too much about the internals of Essence. But it seems to be relatively far in develpment. There is a sleek GUI with tabbing windows in the look of early Chromoium browsers, which looks very inviting, if only I could get it to even try to boot on any computer. The focus has not been on making the OS actually boot on real hardware so far. And unfortunately there has been no release since 2022 and no update to the code for over a year. So I stopped hoping that it might be working soon. I was looking forward to getting to know a knew OS that doesn't take a Unix-like approach and has nice tabbing windows.

(tba:screenshots)

Other new OSs not for real hardware

While Essence is what I created this entry for because I really hoped to be able to try this GUI on real hardware, I'll use the spacce left by the lack of an experience report to mention a few other PSs that look promising but aren't (yet) made to run on actual hardware.

Munal OS

An experimental operating system fully written in Rust, with a unikernel design, cooperative scheduling and a security model based on WASM sandboxing.

TacOS

My from-scratch OS with its own kernel written in C and assembly

TacOS is a UNIX-like kernel which is able to run DOOM, among various other smaller userspace programs. It has things like a VFS, scheduler, TempFS, devices, context switching, virtual memory management, physical page frame allocation, and a port of Doom. It runs both on real hardware (tested on my laptop) and in the Qemu emulator.

Comment via email
Web Pages About The Web Worth Reading

My thoughts about the state of the web and how to shape it are not the most interesting ones. They're fueled by articles and other pages I've read. I decided to link to a few here. Many more I've read and closed without saving the URL. But maybe I'll extend this list in the future so it becomes a curated reading list on the topic.

Web Browsers and Engines

  • grazer - grazer is like a web browser but trees are shorter
  • Dillo - fast and small graphical web browser
  • Servo - The embeddable, independent, memory-safe, modular, parallel web rendering engine
  • Verso - web browser built on top of the Servo
  • The Argonaut Constellation - range of software projects aiming to illustrate the potential for a more private JavaScript-free web
  • NetSurf - multi-platform web browser - small as a mouse, fast as a cheetah
  • -

Comment via email
Alternative Operating System: Sortix
This entry is referencing the entry 'Alternative Operating Systems'.

sortix

Sortiix is one of the more mature hobby OSs. It feels well-rounded in the features that it already has. Stable and relatively bug-free. Sortix is a Unix-like, POSIX compliant OS with it's own kernel, system tools and libraries. It contains packages that have been ported from Debian. But all the most interesting core components are implemented anew.

Sortix has made steady progress over like 15 years. That's probably the biggest difference between it and similar hobby OSs. Version 1.0 was released in 2016. Even in that form it is really usable. It boots without any issues on real hardware. It's stable, comes with all the core unix tools you'd expect, but it didn't have a GUI and network support, yet. It switched to nighly releases after that. Since then, features have been added to the point that not many things separate it from being as useful for daily usage as HaikuOS.

(tba:screenshots)

Comment via email
Alternative Operating System: MikeOS
This entry is referencing the entry 'Alternative Operating Systems'.

MikeOS

A very limited and simple operating system written in assembler for 286 computers (and newer compatible architechtures, obviously). Pretty much what I would imagine as a successful outcome if I would write one to see that I can do it. It works, there's a text editor, a game, you can list, edit and execute files. Not much more though at first glance. I didn't look into writing additional software for it myself, yet. There are many forks of MikeOS. Most of them named after the forker and not under active development. It's a project I'd look into if I'd want to learn x86 assembler. Simple, not looking like any other OS I've seen.

MikeOS is neither UNIX-like in any way nor is it similar to CP/M or DOS. It is closer to being a BASIC interpreter with a program menu. But there is a bit more to it. Everything is in the same text mode in VGA resolution. After booting, you get a box with list of menu items in pointing to submenus or executable programs. Among those programs are a BASIC interpreter, some simple games, a text editor and a file browser. That's about the gist of what you get. It's all very simple and fast. Enough to write your own applications or scripts. And it all fits on a single 3.5" HD floppy. There are third-party applications to be found on the internet. But almost all that I stumbled upon were part of MikeOS forks.

Next to the menu there is also a command line with a very small list of commands. To be honest, from a user point of view, I don't know what to do with it.

There are quite a few forks. It seems to be a great hobby project. And because it's all relatively simple it's a project you can actually finish after a while, not like writing your own UNIX clone with system tools, glibc port and modern GUI. MikeOS forks usually come with some additional apps and scripts and some changes in UI. Some add to the few system calls, add their own menu. Some even started to add network support. One MikeOS fork I have to mention is MichalOS because it overhauled the UI to make it more pleasing and added quite a few simple apps and games that all seem to be very stable. It has an image viewer and a music program. TomOS is a fork that adds support for directories. ShoockOS seems to be about simplifying things even more.

On the MikeOS web site there are handbooks, resources for development and links to software projects for MiikeOS. MikeOS is a great help if you want to learn x86 assembly in a practical way without starting from zero. But it is also useful as an OS for embedded applications or simple hobby projects (interacting with Arduino, other serial communication tasks, …). And it could also be considered useful for everyday tasks if you found a home computer from the mid 80s useful.

(tba:screenshots)

Comment via email
A Document-centric Web

I've been thinking about what exactly it is that is wrong with the WWW and what to do about it. It feels like it became worse over the years. It's easy to make out individual reasons: Walled gardens, ad-centric web sites, bad mass-generated and LLM-generated at the top of all web searches, the lack of search results from small, personal web sites without an SEO budget or the malicious will to implement all the SEO tricks for the sake of it, egoistic algorithms (that push harmful and hateful content because it makes them feel good). Thinking about how these things became the way they are is sometimes interesting and may help by teaching a lesson. But it doesn't really bring us closer to a solution. I've had a few thoughts about what does, though. It's not as complicated as I make it sound here. It's just that I needed to think about it to be sure what I even want. And that is for the web to be made of documents with hyperlinks. If it is a text document, a searchable database, a set of image, a list of things, an article with pictures and audio samples, a video file that I can download and play, a tree of links to other documents or something similar to those, then it is something for the world wide web. If it is an application, it's not a document for the web. It's impossible to draw ab exact line here. But that's the general rule for me now. It's not that web apps shouldn't exist. People make then and people like using them. So that's fine. But that's not what the web was made for and not what I personally want when browsing the web. It's one thing that photo editors, t-shirt designers, CAD software, action games and all those things exist. You can use them or not. But it has become so normal that web sites are huge and require JavaScript to even load that sites started being huge and require JS even though it's not necessary to serve their purpose. And that has become normal, too, by now.

What can you do?

I've came to believe that there is no route to turning the web into the web that resembles my vision of a good web. It's not even just that I don't think that I (together with similar-mimded people) can't obtain the power to force the usage of certain software or to form habits in others. I don't think it would be a sustainable way with a lasting positive effect on the web. That doesn't mean that there aren't ways to bring others to use better alternatives to walled gardens, closed networks, sites with lots of ads and only 10 % content, sites that use dark patterns and search engines of the oligopositic type. You can promote small projects, share links to useful and interesting sites, talk about how you use the web, make it normal to use a search engine that doesn't only find bloated, commercial, SEO optimised sites. You can start your own projects, enrich the market with libre and other non-commercial software, submit pages to search engines with a curated index, create good content for alternative search engines to find. In this entry I'd like to talk about what I do to change my own experience of the web to the better, though, not about making the web better for everybody.

What I do for now

I needed to think about web things a lot before I realised how much of my own web experience I can change by making certain choices and how viable it can be to simply not use certain services. It's not a new idea to me. I don't do Windows, I use alternative front-ends to YouTube, I've used various unconventional operating systems on PCs and phones. But in regards to the web I thought that it's not that simple. Web sites are how they are and even nice web sites link to bloated pages with megabytes of unhelpful CSS and megabytes of maliscious JavaScript. If I use a browser that is fast and doesn't do JavaScript, my web experience is worse than it is with Firefox (I prefer LibreWolf, btw). Whenever I tried Dillo, Nersurf or something similar, or when I disabled JavaScript in Firefox, I didn't get along with at least some pages. It doesn't appear to be easy to simply decide accept that some pages don't work and just go on to the next one. There surely are use cases where this is not acceptable. But for the usual usees - everyday browsing and casual research - that shouldn't be a problem considering I already accept that some pages aren't accessible because I don't want to register with them. In a sense it's my fault that I don't like how the web is today. It's me who keeps visiting web sites that are like that. And I can stop that by doing some simple changes: Disable JavaScript and use a search engine that prefers non-commercial web sites in its search results.

Really?

Don't get me wrong: I doesn't feel like an improvement to disable JavaScript entirely. There are browser extensions that let you control what page is allowed to serve scripts and what you want to execute. I've tried that, but it's complicated to get it right and frustrating because you always have to configure stuff while browsing and that never stops. But maybe it is an improvement despite not feeling like one at first. I mean, quitting to take drugs to which your brain has developed a strom chemical dependency also is often a worse experience than continuing to take them. But also often it's worth getting used to not taking them anymore. It also doesn't seem like an improvement to only use search engines with tiny indexes that rarely return with the ideal search result you hoped for. Maybe this isn't a viable choice. I think there is no search engine of that type with a large enough index to recomment it for daily use. Those projects just aren't there, yet. But that may just be one more reason to use and support them more. And since they do get rid of all the sites that do things that I don't like, it could be an improvement to get used to using them.

What does that mean in practice?

Some web sites don't have any images anymore, some web sites only load ads and recommendations, but not the actual article, modern closed-plattform chat apps don't work, just as most other sites that can be calles web apps, keyboard focus doesn't start at the main input field, some sites aren't readable because all their styles are missing, burrying the content between or under thousands of things that should have started out hidden, on some sites certain links aren't working anymore, many audio and video players don't work because of attempts to prevent permanent file downloads, there are pretty much no ads. If you use a browser with a less than very popular rendering engine, add misaligned elements on many to almost all web sites, unreadable elements on some sites with unconventional styling and missing elements if they use unusual positioning options. If you only use a search engine that doesn't do commercial sites or whitelists desirable sites, add to that the feeling of trying out the web in 1995 unless you navigate to specific sites that you know contain what you are looking for. The web feels relatively small with a search engine like that. But even then it's huge. Maybe it's a quiestion of what you expect. If you really don't know what site you are looking for, use a universal search engine. If you have an idea where to find the information you are looking for, start at that site. It might be Wikipedia, Slashdot, Toms Hardware, an Invidious instance. The web is totally usable if you don't enter everything in the same search engine as a lefrex. I thought it was great at a time where we didn't do that. And for the rest (missing content, non-working sites): Those tend to be the sites that I wanted to filter out in the first place. So the endeavour seems to work as intended.

There are also sites that I would like to read that just happen to use a CSS trick or JavaScript that isn't supported by all browsers. Those are sites that don't pay a lot of attention to accessability design guides but don't have any bad intentions. I've made sites like this before. This article is being posted to sites that fall in this category. Feel free to contact makers of those sites to let them know that you would appreciate being able to read the pages. I know I should test a site in text browsers before publishing them. I never do. Nowadays I don't even test in any other browser than the main one that I'm using (except when using engine-specific style rules). We came to accept that it just looks the same in all browsers. That is something that web designers always wished for. And when Microsoft's browsers improved in this regard it felt like we were there. But it is also true that most users of the web use a browser with one of two/three engines (depends on where you make the cut and call it a new engine after a fork). I think it does still make sense to test a web site in different browser engines. It doesn't have to look great in a text browser (although that would be the best), but maybe Dillo and NetSurf. If you cover those two, you cover pretty much everybody and you don't even have to test the site in Firefox or Chromium.

So, what did I change? For work: Nothing. Corporate dictates what software I use for what. On my private laptop, I currently use NetSurf as my main web browser. I use LibreWolf for two purposes: Copy individual bookmarks or URLs from open tabs, and go to sites that don't work in NetSurf when I don't have the time to find an alternative solution. For chat apps I use their "native apps" although I suspect that they are all just the web app shipped with their own browser. For social media I'm trying out different mastodon/fediverse clients for Linux, which I wanted to do for a while anyway. For search I'm currently using various Searx/SearXNG instances. (I know, not that alternative. I don't want to ruin everything at once for me.) More than recently I deliberately navigate to a specific site instead of using a search engine and ending up on a site that I already knew. When looking for something on eBay, I don't find as many interesting things like before because the pictures are missing and I don't needlessly buy things as much now. When searching for some random information or doing some curiosity research I close many search results directly or very soom after opening them because they aren't displayed properly. So far that doesn't bother me much. I'm already used to having to close tabs again right away because of cookie banners and other popups that make it impossible to get to the content without finishing a maze and reading a bunch of things for at least a minute. Not I open and close more search results, but get my ansers anyway. On video platforms, I open the video in an external player. It's nicer to have the player of my choice with my prefered UI and my custom configuration anyway. Some sites simulaniously look worse and better at the same time. I may have to scroll a bit to the content and it is obvious that the page wasn't designed to look exactly like that. But at least I don't get any grafical animations, lots of side-loaded unrelated content or ads. For shopping my options are very narrow. I already stopped using Amazon for other reasons a while ago. It's really not as much of a hassle as people seem to think. But much more shops than I expected rely on JavaScript for purchasing or logging in nowadays. (Probably at least for a CAPTHA.) It's pretty much all, actually. According to my rule from earlier, those are apps though, and there would be better ways to implement those. So, I don't have a solution other than switching back to LibreWolf or an app on my phone when I need to buy something online. So far, I didn't actually need anything, though. For online banking, sending a message to my insurance, using the Wayback machine and I predict much more, it is the same. For some things I will try to find alternatives. For others I will realise I don't have to. For some sites that I want to consume for enternainment it's disappointing when they don't work. There are so many alternaives for entertainment in all categories. I have so many ebooks, web books, audio books, lecture recordings, podcasts, … that I would like to consume when I get the time and energy to, I really don't need whatever interesting thing I've just found or somebody has just recommended. But now that I know it's there, I don't want to miss out. So far, this has been largest part of my negative experience after switching. But I haven't been at it for long. I'm curious to see how this will go for me.

(tba:links to previous entries, external links)

Comment via email
Case Modding

I used to be somewhat of a casemodder in the early to mid 2000s. That's kind of the style of casemodding that I still like today. We used to shun people who buy ready parts or cases instead of building parts and modding cases themselves. I especially was an advocat of building things from materials that could be found on the street or in scrap containers or were otherwise free to aquire. Maybe just because i didn't have any money, didn't know many people who had enough money to simply buy materials and tools and didn't foresee a future where I was able to simply buy anything I needed to build something. To this day I like using leftovers, scrap and otherwise free materials to build things. I think I wasn't aware of that in the 2000s; but which case mods I like how much is to a large degree determined by how scrappy the building materials were and how simple the tools were. I wasn't usually trying to build something that looks slick and exactly like planned, but something that looks unique and cool, and maybe extraordinary.

Here are a few examples of things that I did to cases of mine that I liked.

(I'll pick out photos of these examples at some point, maybe. I haven't yet.)

Coloured foil window

Making a whole in the left wall of a tower case is probably the most common case mod. There were various window kits to make it easy to get to a clean-looking result. I didn't care for those for a long time. Instead of bying any material to create a window in my case, I used what I had: a dremel-like tool with a cutting disk and some red polymere foil an old text book used to be covered with. If I remember correctly, my mother bought this book used one year for my new school year instead of getting the current version. The previous owner had a red protectie cover around it. Eventually one of the seams ripped and the cover slid off every day. So I left the cover off. It was translucent. A unique material, I thought. So I kkept it in case I wanted to build something with it some day. The case I used it for used to house a generic 286. I put a 586 in it, I think. In the early 2000s that was just an old, very slow computer, not a #RetroComputing statement. I glued the sheet to the steel case from the inside after removing the burr and abraded the edges with a used corner of some sand paper. With bright cold lights inside you could sort of look inside the case. But it was mainly for style. The rest of the case was covered in some tape that I found at some building site once. That way the cut edges didn't look too rough. It certainly was a unique look. I still like the style of that case.

Plastic hose IDE and floppy cables

Ribbon cables, such as they were used for IDE, floppy drives and SCSI, used to be impossible to tuck away nicely. Round cables, such as pretty much all cables that are connected externally, can be clipped almost anywere. Wide ribbon cables need to be folded to wire them cleanly. That doesn't even work well if the case and all parts are designed for it, which they never are. In most PCs those cables used to be just left hanging around, blocking airflow and view. Some PC manufacturers used to cut the ribbons into five or six parts and fixed them with a stacked position with cable ties. That looked much cleaner. And suddenly round IDE and floppy cables became a thing. The connectors were the same. But inbetween they weren't ribbon cables anymore because the individual wires were split up and shrouded in a plastic tube. Those cables were more flexible and usually more colourful than conventional IDE cables. Again, you could buy them. But until the day I got a set with a motherboard that I bought, I didn't want "factory made" round IDE and floppy cables. I made my own by splitting up ribbon cables and stuffing them through an old shower hose or a piece of a garden hose. Not as flexible, but just as practical as bought ones.

Aluminium tape wrap

I don't know where I got it. But I had enough wide aluminium tape to cover an entire mini tower on the sides and the top. To cover up the imperfect edges I used red electric tape. So almost the entire case was striped red and silver at an unusual. Simple, no cost for me in that case, and quickly giving a nice, retrofuturistic look to a before boring, grey mini tower.

IC exterior

Another idea that I once had was to use all the ICs (at least those that were at least 1 cm wide) from all those defective motherboards and extension cards that accumulated over the years. People would bring and I would pick up so much old PC hardware that others didn't need anymore. Often the reason was that at least some part had a hardware defect, in which case I usually gave up my hopes to get it working again. I had an entire wall covered in old motherboards at one point. Much of this stuff was from the mid-90s or older and therefore not worth keeping intact even if it was in a great condition when I got it. Eventually there were just too many unused and defective cards and other boards and I decided to recycle their ICs before I got rid of what I thought I'd never want to even look at again. I cut off all those ICs with a knife (only SMD chips), covered the right side of a big tower in double-sided adhesive tape and neatly placed one next to another. Sometimes I used chips smaller than a cm to fill gaps. Not even half of the side got covered. I tried to get more from other people who wanted to throw stuff away. But what I got didn't bring me close to covering even the one side. My idea was to cover all the sides. I realised I had to pay money to get enough even broken electronics to finish the case. That wasn't beside the idea. Also it didn't feel right to use just any old ICs. It was supposed to be a PC chip case. I never finished the case. Unfortunately I didn't have the idea to make the IC field transition into something else, like a solid colour, even just the gray the case was before. Or maybe I didn't want to dso that. Nowadays I have some ideas where I could possibly get trunkloads full of ICs that I assume aren't recyclable otherwise. Maybe I wouldn't use a big tower if I'd ever started this again.

Sofa PC

I had a sofa in my room. I don't remember where I got it. I probably picket it up from the street after somebody got rid of it. At the time I was thinking about andd experimenting with getting a PC very quiet, if not silent, without compromising on performance. This was a much more prevalent topic at the time, because CPUs used more and more power with every new model (peaking in the Pentium 4, which is famous for needing much more than 100 W at the clock speeds it was marketed for being able to run at). CPUs power consumption wasn't throttled in as many ways as they are today. Coltage was usually fix, so it needed to be as high as you needed it to be at peak performance moments. Clock usually couldn't be changed dynamically. The CPU couldn't switch parts of itself off when it didn't need them. And the power it consumed it needed for its single core; so there was no core to shut down, either, in idle moments. Other components usualy also used more power than they do today. Motherboards were built by more individual ICs, chipsets didn't idle well either, voltage regulation wasn't as efficient for performance CPUs, hard disks needed more power than even today's spinniung hard disks. Automatic, temperature-based fan control wasn't as advanced either when you wanted to regulate it for a chip's temeperature, and you had to place the sensor not only outside of the chip, but usually outside of the cooler. And I wanted to have a surver run 24/7. I was actually running a few services on the internet from my bedroom at the time, of which one was used and relied on at a daily basis. So, what I came up with was to build a PC into my Sofa. It had thick foam padding and a cotton filling, which made it sound proof at least to lover frequencies. I removed enough of the cotton to give it a large room with wich to interchange air. And it sort of worked. It got hotter in there than I hoped. But it was pretty quiet. Now that I'm thining of it again, I could've done some things to improve aitflow to the outside withou opening the sofa up too much to leak sound. There were no frequencies, apparently, at a multipe of the resonance frequencies of any of the wood panels of the sofa. Sitting on it wasn't affected in any way. And the whole thing was easily accessable from the front. There was a cut-put right above the floor with a handle, with which you coulde pull out a board, that slid out on small wheel. On the board all the components were mounted. I used parts from a relatively cheap ATX case to make the motherboard and drives mount easily. I never got it down to the temperature I set out to get. And I couldn't use high-rpm hard disks because then that was the only noise in the room and it was very annoying. But it worked, and the sliding mechanism was fancy, even though it was so simple to build. The cutout and handle were actually easy to overlook in the pattern of the fabric that the sofa already had. I got the ideaa for the sofa mod after I ran the server inside my wooden desk for a while. That is another story. The sofa mod was kind of the enhanced version of the desk PC.

I wonder why I don't make things like these anymore. Because making something in a unique style that you like feels extra good on top of the feeling of making something yourself.

Comment via email
Mastodon