Exactly.

There is a major difference between running a vm on your desktop and orchestrating a fleet of highly available virtual machines. Just one example might be vmotion. You can move a virtual machine from one physical host to another in real time with 0 interruption to services running on that host.

That’s some incredible stuff. Now days you can use things like XCP-ng to do the same but VMware was ahead of the pack for a decade.

They started dying when they were squeezed between cloud hyper scalars and the cheaper alternative hypervisors that finally had caught up.

Then the corpse was bought by Broadcom who is currently trying to milk it before the body completely rots.

Not saying we shouldn’t do both, but in reality waiting to destroy capitalism before fixing the grid just means you have too much theory and not enough praxis.

I do the same. I still get a ton of posts like this don’t include any relevant key words and still are about musk or trump. It’s impossible to escape.

Small note. Opensuse Leap is EXTREMELY stable. Just as stable as RHEL and more stable than non-LTS Ubuntu. It’s just less well marketed in the English speaking world.

In terms of algorithms, nothing. But you were the one who mentioned algorithms. I am speaking of code in general. I do want for persons to contribute back to the community if they use community sourced code. I don’t think we can trust corporations to be altruistic.

This all being said in your earlier message you were implying it’s all about ego. I was just saying it is not about ego.

For me it’s all about community resources and societal enrichment.

CC0 vs CC-SA is actually a really good (rough) analogy.

Sure. Very briefly. These are all open source licenses which (roughly) means the source is freely viewable and changeable. But the specific differences are:

• MIT/BSD - Anyone can take the code and do whatever they want, if they start with your code, improve it then make it proprietary there is nothing you can do.

• GPL - If someone makes changes to your code and improves it they have to make it available for use by the community too IF and only if they distribute the binary.

• AGPL - Like GPL except that even if they are running the code on their server and not sharing it they still have to give back improvements.

• MPL 2.0 - Like GPL but limited to specific files. This is useful for things like statically linked code. I don’t often recommend this but it can be needed for static only code bases like rust. Proprietary software can link with this and not be covered by the copyleft share alike stuff.

• LGPL - Like the GPL but for dynamically linked libraries. Proprietary software can link with this and not be covered by the copyleft share alike stuff.

• SSPL - Like AGPL but technically even more intense. If you use SSPL you must open source all the tooling you use to manage that hosted SSPL license. Any tools to make sure the SSPL software is running well or to set it up must also be open sourced.

The OSI technically does not say the SSPL is “open source” but given that they recently admitted that they regret defining the AGPL as open source I think the OSI might be showing a bit of corporate bias.

Pull requests have nothing to do with any of this. Also algorithms can’t be copyrighted nor patterned in the first place so it would not matter.

You could implant an algorithm in a proprietary code base and some gal could reverse engineer it and publish it as GPL or MIT or whatever and all would be a-ok.

Wait, so because a few execs violated the GPL and threw their employees under the bus, we should abandon copyleft entirely? That’s like ditching locks just because burglars exist. Companies that want to exploit software will do so, BSD or not. The GPL didn’t land those four guys in prison; their higher-ups did. Giving up and saying “ok big corp I’ll just do what you want“ just makes it even easier for corporations to profit at societies expense.

MIT and BSD software licenses might as well be renamed to “I love big daddy companies and trust them 100% uwu”

There is no reason not to choose GPL/AGPL/MPL 2.0/LGPL/SSPL if you are writing open source code.

MIT and BSD just let companies enrich themselves at societies expense.

Small warning about workman. It has issues with lateral movements and single finger n-grams. “ly” and “ct” being notable examples.

A piece of advice I heard that served me well was to look mostly at post covid designs. A lot of work was done on layout optimization around that time and the results show.

My recommendations in no particular order are:

Colemak-DH if you want to focus on a well supported layout.

Graphite or Engram or one of the hands down layouts are modern well optimized layouts I would consider if I was to learn something today.

Some people like MTGAP but in my book it was designed with too much of an emphasis on minimizing key spacing without a strong enough emphasis on how human hands work.

I personally use engram but it only works for me because I have strong pinkies. If you don’t it’s probably a bad choice.

Engram. It’s a great layout that focuses on pinky in rolls.

It’s a steep layout to learn even compared to thing like Colemak but I find it quite satisfying.

https://engram.dev/

This article is from 2024. Hopefully it worked out. 😆

Yah. Like I said. Lack of prudence with plenty of unpredictability on top.

That was more than a day ago though. Trump’s lack of prudence ensures that companies can’t plan ahead. That includes the Taiwan based framework.

I would recommend „Paper : paging through history“ by Mark Kurlansky or his other books on Salt or Cod.

They may sound boring at first glance but they’re very interesting and informative.

It matters though. Like in Germany telegram is associated with hard right wings groups. Telling someone you use telegram makes them assume that you are a part of hard right ideologies.

It’s a shame as the telegram app is really snappy. You always have to say that you are on telegram but are not right wing. Even then people can be suspicious.

Think of it like the flooded house analogy. They could paint the dry wall while tearing it out at the same time. But why would they?

They’re not refusing. They’re actually doing the opposite. But they needed to get their house in order first.

The 3.0 upgrade was the result of the getting their house in order and modernizing. Doing cosmetic changed before hand would have made no sense because those changes would have been thrown away when they would have to modernize things anyways.

I think I have an analogy.

Gimp was like an old American style wooden house that was flooded. After the water recedes you could try to make things look nicer by plastering and painting the walls etc. But as goes with flooded houses if you do this the mold will rot everything out.

In order to save a flooded house you need to remove all the dry wall and use fans to dry out the internals. Once things are dry then you can plaster and repaint things.

Gimp 3.0 was them ripping out dry wall and air drying the internals. Now that that is done it now makes sense to clean up the UI.

If you clean up the UI before you dry the walls out it’s just a waste of time because those improvements would need to be ripped out with the dry walls always.

It’s not perfect as far as an analogy goes but it’s close. Gimp should have never let the house flood in the first place. (Analogy breaks down here a bit). But since they did. They needed to fix the fundamental before it would be worth fixing the UI.

This all being said they could at this point genuinely refuse to change things UI wise. I hope they choose to pull a Blender or Krita but they don’t have to.