Dont get me wrong - I really like and appreciate your comment.

However, and adding to other replies, by SDL I assume you mean the Simple Directmedia Layer?

SDL looks rather strong from my perspective and still my typical goto when having fun making a game. You could argue SDL lost some customers in favour of other libraries like RayLib - or moving away from making things from scratch to Unreal, Unity, etc.

SDL still seems popular - as SDL version 3 was officially released less than a year ago (or it feels like it) However, I guess it depends what you need to do.

I do mean SDL, and the discussion is not about its popularity, but about something formal, but perhaps I was not clear. First, yes, SDL is a very good library, and I have referred to them in the past for references about issues related to software architecture.

However, the way I saw SDL 1.x, I am expecting a strong contract. Every now and then SDL drops support for one thing or another. Where you had to worry about different APIs, now if you want to retain your strong contracts, you have to worry about different SDL versions. I am aware of something like the "sdl12-compat" layer for example, somewhat similar to JQuery Migrate, but it does not change the fact that the underlying contract is not strong, is not trustworthy, because of both changing APIs and changing compatibility, similar to the JQuery situation.

I think STL was the intent.

I remember the first time I tried jQuery -- which was 2009.

Before jQuery, I had vanilla JS code that factored in Safari, Firefox, Opera... and... IE6, IE7, and then IE8 which was (from memory) recent at the time.

Trying to design a visual drag n' drop editing interface on the web was a chore at the time especially with the differences in IE browsers! It was suprising how many customers were still using IE6!

A lot of this is purely by memory, now. I even have shivering memories reminding myself I was using VB.NET with ASP.NET Web forms. I really HATED it!

I remember ASP.NET provided dynamic web pages with things like Placeholder tag, etc. Again, It felt bloated even back then but I made it work. It was a balance of using what was recommended by other developers, and trying to ensure performance is good.

By around end of 2009, I tried jQuery as an experimental branch and very impressed with its capabilities. Despite being a decent Javascript developer at the time I was inexperienced with AJAX (Technically I was a Junior dev in ways) but jQuery shows me the way. It was not long before I ditched certain features of .NET Web Forms for jQuery.

At the time, there may have been a little overhead replacing various javascript code I am written with jQuery but the rewards were high. It was cleaner frontend and backend code, thanks to simple AJAX calls.

Since then I've always had huge respect for jQuery! While I don't consider myself a decent javascript as I don't use it much. However, when I do come back to web development, I cannot be asked with the modern ways. To me it's bloat. I just stick with htmx, now. If I have do some fancy things visually then I will use jQuery.

Very sad news.

I don't have as much hatred towards Gnome 3 like everyone else does.

Don't get me wrong, I am certainly not defending it. I was a little heart broken as I really liked Gnome 2. However, I tried to be optimistic with their plans overall.

(I think the early days on Gnome 3 featured something call Gnome Legacy to keep that Gnome 2-ish feel. I likely stayed on that for a while)

I still use Gnome 3 today... but Xfce would certainly be my second choice.

I don't have "hatred" towards Gnome3. I use it for friends and families desktops, they seem to like it. I have also rolled out about ~20 Gnome3-based desktops for my employer.

That said, there are definitely areas were Gnome could be improved. Some of them are understandable and probably stem from a lack funding / devs. Others less so, like removing the options to scale / stretch / center the wallpaper w/o installing "Gnome tweaks".

Yeah - I find it a little frustrating that the first thing to do after installing Gnome3 is to install Tweaks.

It's not even the chore of installing it, Ansible will mostly do that for me. It's that I can't comprehend why something as basic as fundamental wallpaper config is not part of the normal GUI. The reasoning for that one is beyond me.

I have been a Linux user since 2005 or 06. It was one of the best decisions I ever made in the world of computers. I started with Ubuntu and went on a distro hunt such as Fedora, Mandriva, Mint, and others.

In the end, I have stayed with Debian for a good 10 years, now. Both for desktop/laptop and server. (Maybe one day I will dive deeper with Arch.. but I also like the direction of Guix)

While I use Windows at work, I did have one machine at home dual-boot with Windows. I used it at times for games or work. Now, I cannot find a good reason to have it. My recent Debian install for my laptop + has Steam and Heroic (Epic) without issues.

Despite all the above, I am not someone who forces Linux on others - though I speak my truth and make jokes about it.. but I have little issues with my Linux installs. While I have an ATI card which installed without issue, I did not have much problems with nVidia cards in the past.

I dont hate Windows as a pure Operating System. If you remove all the fluff on top of it - especially with what is going on with Windows 11, you likely have a stable, reliable kernel + core applications.

I do not like the direction MS is going with Windows 11. I am expecting the next version of Windows to simply be named "Windows AI" but don't worry -- they will tell you AI is optional.

Boiling Frog Syndrome - little by little more features will be mandatory. :-)

They keep reminding you so you end up buying a new PC with Windows 11. That's all it is. No extra code logic is needed. Just keep pestering the end user with a popup. They don't care for any variations.

The goal is to get everyone on Windows 11.

This is not 1998 or even 2008. Times have slowly and "progressively" moved on. Truth is you NEVER OWNED a copy of Windows. You always purchased the rights to USE it. Now that technology has improved especially the internet Microsoft have tried to gain more control over YOUR computer.

Look at Nintendo and their Switch 2. In their world you do not own it. If they think you are doing something "you shouldn't" they will brick it! Sure, I am not suggesting Microsoft does this with Windows but I am 100% certain this type of tactics has been discussed in high-end meetings. The key here is CONTROL.

Today - if Microsoft want to push a program and "encourage" you to use it.. they will install it without any form of consent. Sorry, but if I OWN a computer then I want control with the software installed, including an Operating System. Microsoft has always been a huge '??' in this field and, to me, it is getting worse. I am not even talking about government involvement with the big techs, either.

Copilot -- I dont care.

This rumour (is it a rumour) that Windows takes photos of your screen every so oftern... NO THANK YOU!

I might have to continue using Windows (11) in my job, being given a work laptop, etc. At the end of the day I do not care as its not my laptop and decisions are taken outside my control by specific IT departments. Whatever. At home I am 100% GNU/Linux. It is sooo much faster and programs loads in <1.5 seconds compared to 5-20 seconds on Windows 11.

My biggest concern is the future of GNU and Linux. Well, the Linux kernel more so especially when Linus hangs up his keyboard. Hopefully the next guy in charge cares about our Freedoms. Honestly I imagine an alternative world where a corporations takes control. The beauty, thanks to the GPL, is people can branch off an continue their own. Sadly... GNU/Linux MIGHT get infiltrated one day and most of that will be people NOT caring about our FREEDOM. This, in my opinion, is all dependent on the future generations.

Being a teenager, I honestly viewed the N64 (Ultra 64) as being an unstoppable force during the early news. We even had a 486 PC running Doom since the early-to-mid 90s. LOL. I couldn't wait to see what Doom would be like on the N64.

In typical fanboyism, I viewed the main SGI systems are the superior systems to the N64.. but they were for the office.. not the home. The other was Panasonic M2.

Of course.. either the N64 was released in the UK or not far away, I remember walking into PC World (a cool computer shop at the time) with a demo of Tomb Raider. I believe it was running a Voodoo1 card and the realisation kicked in. The N64 is already surpassed.

Wasn't long before we had a Voodoo2 card and the first game we played was a demo of Turok:Dinosaur Hunter. It was much better than the N64 version (which I owned)

Once I started playing GLQuake it was a PC master race.

It was at this point I snapped out being a Nintendo 'fanboy' and accept that hardware gets you so far... its the games that make a console.

I still think the N64 was a great console. Goldeneye, Ocarina of Time, etc. Great memories. Also got a Gamecube, Wii, and Switch.

As I got older and snapped out of the fanboyism, I realised the Playstation was a good console. I am on the fence with the Cartridge vs CD argument. There are cases for both. If we look at Mario 64, many aspects worked well on Cartridge. I guess faster load times and transition of music. With CD.. you have CD quality music and more variation of textures and level design (generally speaking)

Putting all the aside, Playstation is just as much capable of doing a good Super Mario 64 port despite the (many like) PS1 jaggy polygons.

I was doing some pretty decent rendering in HTML4 back in 2008.. supporting various browers (let alone IE6, 7 and maybe 8 at the time)

Around 2010, I did experiment with things like Silverlight and SVGs. SVGs was OK, but the performance quality was not there. It might be a lot better now.

My only gripe is with Vec3_new() function in "Memory Ownership" section.

It assumes you want a single malloc of Vec3. It tries to behave as if you are doing a 'new' in an OOP language.

Let the programmer decide the size of it.

Mock example (not tested)

  struct Vec3* Vec3_new(size_t size)
  {
    if(size <= 0) {
      // todo: handle properly
      return NULL;
    }
  
    struct Vec3 *v = malloc(sizeof(struct Vec3) * size);
  
    size_t i;
    for(i = 0; i < size; i++) {
      v[i].x = 0.0F;
      v[i].y = 0.0F;
      v[i].z = 0.0F;
    }
  
    return v;
  }

I am a big fan of 'do one thing and one thing well' -- that UNIX philosophy.

Whether it's a program that does something well... or simply a function/procedure --- it all depends on the problem I/we are trying to solve.

I never liked using the word "Microservices" but my aim has always to build SIMPLE solutions. I learn new words in this world. For the most part I am building "Miniservices" but there are a few that are considered "Microservices" but again are not complicated!

I just like to refer it as "Distributed Computing" because the solution can be anywhere between Monolithic or Microservices. Truth is you are building a combination of them that communicate in one form or another.

I will always remember a Till system (past job) that was sending data to the server poorly and slowly with a Monolithic solution and a Database. Was it becoming a pain to handle with new shops being added in Europe? Yes. However, this is NOT the fault of Monolithic. It's just the solution that was used for "good" originally but is struggling now.

The solution I replaced it with allowed data being sent to the server using ZeroMQ. It worked out well.. was fast and reliable. Each section was broken down on the server. Again - is it is perfect solution or does it prove that "Monolithic is worse that Micrsoservices" (or Distributed Computing) -- NO! Truth is our software is a mix of them all!

I personally think ZeroMQ is severely underutilized. Fixes so many problems.

But it can be hard to encourage adoption. It’s not HTTP, or a conventional queue system.

Requires lots of explanations, thinking, and eventually meetings.

I made a lot of uses for the Push-Pull pattern in ZeroMQ.. but there was others used as well here and there (Pub-Sub, etc)

However -- I cannot praise the use of ROUTER-DEALER !! What a GREAT pattern for sending large chunks of data without waiting for a reply for each.

HTTP is not fit for such a task.

However - I totally get you! Trying to explain certain decisions really does take up time and effort. Before you know it, I have lost 4 hours one day, a few another, etc.

In the end you ask if I should have done something mediocre... but everyone understands.

("mediocre" is not the correct word to use. I mean I could have gone with Kafka or RabbitMQ. It's just an extra layer which would have involved infrastructure and further delays (at that time))