The Linux Kernel Has Been Forcing Different Behavior For Processes Starting With "X"

[oiaohm]

App developers have to develop applications that use APIs, otherwise application development time and cost would be significantly increased. App developers are expected to ship applications as executables, because that simplifies the process of installing applications for end-users. The executables require a stable userspace ABI. The app developers are currently unable to express the contradiction of applications requiring an ABI vs. applications requiring an API, especially due to the regulator having a respected position of authority. The app developers are in a situation known as a double-bind (see description on Wikipedia of a double bind).

This contains mistake. Windows for example does not technically have a stable ABI.

https://learn.microsoft.com/en-us/windows/win32/sbscs/application-manifests

Since year 2000 when you make application under windows it contains manifest data listing the runtimes the application in fact was built with and in fact needs.

App developers' preferences are generally in this order: stable API, API, stable ABI, ABI. Also note that shim libraries may exist that can convert an ABI into a shim API. This might change the order of preferences to some degree.

This is what application developers say they want. But it not what they are really using. Visual Studio(windows) and Xcode(mac os) both when they make software mark the software with what it needs based on what it was built with. More detail than you need to dynamically load libraries.

This incentivizes the ecosystem of Linux distributions ("the ecosystem") to providing tools for managing the situation of unstable APIs. Those tools have grown sufficiently sophisticated to solve practically all the problems arising from of such a situation, except for the need of app developers to ship executables.

Is this true I would say not on ship applications. Remember you have NIX and Flatpak and Snap these all provide application developers with way to ship their executables with include data like windows manifest will be processed to get a snapshot in time of the ABI so application works predictably. Yes snapshot in time of the ABI is what Windows, Mac OS and Android really use for what appears to be ABI stability.

Application developers on Linux are mostly in a bind because they are not willing to back a solution. You need to step back and look how windows and macos really work.

The features that make Windows and MacOs appear to have stable ABI when they really don't have been integrated into those platforms default IDEs(integrated development environments) that application developers don't need to think about it most of the time leading to the false belief that the ABI of Windows and MacOS is stable so they can now demand this of Linux and do nothing when the Linux world refused to give them this. Then look at how well integrated is nix and flatpak... into IDE solutions under Linux notice mostly its not.

https://jgrulich.cz/2018/09/03/flatp...t-in-kdevelop/ here KDE that packages most of there applications as flatpak and here is the KDE default IDE. Notice hell yes you have to manually write the flatpak manifest. Under Windows with visual studio the SXS manifest would be 90%+ automatically written by the IDE based on what you choose to use when you create new application same with xcode and macOS.

xfcemint does the problem have anything todo with Stable ABI? is a good question and the answer is most likely the problem has nothing todo with stable ABIs. Or is the problem simply the solutions under Linux that have manifest for compatibility have poor IDE integration so are too hard for most application developers to use? This is most likely the true answer.

Are application developers in double-bind is presume you need to remove. Maybe application developers should invest in fixing the IDE problems so that making portable applications on Linux is simpler.

[oiaohm]

I think that I have just found a good example of point 8: wlroots for Wayland. That's basically it, right?

- point 8: I find it to be likely that the possibility of incentivizing creation of smaller services and smaller libraries that can be used by multiple vendors (to simplify development of services) has not been sufficiently considered and discussed.

This here is a presume that it goes right.

https://github.com/Plagman/gamescope/issues/26

Gamescope at times has had issues with other wlroots users because it contains a fork of wlroots at it base.

Multivendors using smaller libraries seams like a solution until those developers start carelessly including their own copies resulting in either dependency hell or dll hell if you don't have some system to manage this.

There is a reason why windows has SXS " Side-by-side assembly" and NixOS does what it does and flatpak and snap and appimage... all exist.

Application developers need fences/rules so they don't ruin everyone day..

Your point 8 solves nothing. Point 8 has been done for decades. Always with the same failures. Big one being conflicting versions version requirements. Library like wlroots is not 100 percent ABI stable so version dependency is very critical.

Yes you found example. Not example that proves point 8 but example that proves point 8 idea is flawed.

[oiaohm]

We kind-of agreed on "psycho-therapist". It appears to me as less ambiguous than "therapist". In my native language, the word "therapist" is used for a wide variety of professions, and the word "psychiatrist" is commonly used instead of "psychotherapist".

So your native language has fused those words in UK, South Aficia, USA, and Australian medical(most English medical). Psychiatrist has the qualifications to prescribe medications and do studies and the like but are not allowed to perform sessions. Psychotherapist does all the person to person sessions. This is a very important split.

Yes in most if not all English language countries if you are on a couch talking about your life you don't have Psychiatrist in the room its Psychotherapist this is from the way the medical system is setup.

Psychiatrist might give person suger tablets instead of real drugs or lowered does and the Psychotherapist working with the person is clueless to this so gives them no tells. Also psychotherapist if session is not going right can legally look at the prescriptions to double check the Psychiatrist. Yes the system basically has the Psychotherapist like one script behind the Psychiatrist and the Psychotherapist is their to hopefully prevent someone being incorrectly drugged by the Psychiatrist.

Fun point here is with Psychology is both the Psychiatrist and Psychotherapist can have the same qualifications. The major difference is what they are registered to-do in the medical system. Yes the 4 countries I mentioned does not allow a person to be registered as Psychiatrist and Psychotherapist at the same time.

Base of the difference is a medical safety thing. Maybe the country you are in xfcemint does not have this clear and legal difference between the two.

Yes Psychotherapist giving out drugs commits criminal offense and Psychiatrist doing sessions commits criminal offense in those 4 countries.

[oiaohm]

I have heard that psychotherapists are reasonably effective at curing stubborness.

https://somaticpsychotherapytoday.com/the-power-of-stubborn/

This what happens when a person does not understand this topic. Psychotherapists don't cure stubbornness. Stubbornness is part of being human. The best they can do is guide it in a productive direction. To guide stubbornness in productive direction requires a logical argument that not flaw.

Pure Double bind/"Irreconcilable Differences" means the stubbornness cannot be guided.

You are misrepresenting my words. I said: a regulator, his psychotherapist, representatives of application developers and their psychotherapist(s) need to engage in a serious and reasonable discussion.

Something you miss psychotherapist try not to be stupid enough to put two parties with "pure double bind"/"Irreconcilable differences" in the same room. Why because the stubbornness can turn into anger and end in violence. Again this is a pure natural human nature outcome.

I said DOUBLE BIND. Double bind is not the same as "Irreconcilable Differences". "Irreconcilable Differences" implies an unsolvable problem.

"Double bind" has a clear explanation of what is happening, while "Irreconcilable Differences" doesn't provide any insight into the problem, it is just a statement of resignation.

Yes "double bind" explains what happening. The reason why psychology in training has two forms of "double bind" is that before you know that its "Irreconcilable Differences" these presents as a "double bind" these are the "pure double bind" no amount of psychotherapists work can change this they are being valid stubborn as per human nature attempt to push past that will bring out very bad parts of human nature. The pure double binds you have to get all parties to accept as "Irreconcilable Differences" or you are not going to move forwards.

The other form is correctable. Because there is some falsehood.

You are right that "Irreconcilable Differences" is resignation and acceptance that you have unsolvable double bind.

Having explain of what happening might seam like you can solve the problem but the problem is this is too little information. I guess you would have hear of "who what where when why and how" in writing. Psychology you are mostly interested in why and how with "who what where and when" being minor factors. Double bind only answers how. Pure/Irreconcilable Differences is valid why answer for being in double bind location and it the worst answer because if this is the answer its not solvable by using psychotherapist personal in a group setting because all you are going to do is get stuck in the same arguments that no party can prove as absolutely invalid at worst by putting them in a room end up with someone in hospital possibly the psychotherapist who sets up this foolishness.

None of your 10 points has in fact consider why we are in this location. I think you would think you were writing why but all you have been is writing how. Yes methods you have writing to how to fix problem only put in place a theory how and hope the theory how will work without considering why we are in the problem correctly in the first place.

You don't want to admit that "Irreconcilable Difference" is a real reason why for lot of the lack of progress. Yes you had incorrect idea that psychotherapists can cure stubbornness and this is something psychotherapists absolutely cannot do and it why the term "Irreconcilable Differences" has to be used by psychotherapists in making plans forwards.

Yes it can be a very hard path to get people to accept that the problem is Irreconcilable Difference due to human stubbornness as well.

[oiaohm]

I find it likely that the regulator has plenty of actions on his disposal, he is just too stubborn to consider them seriously. One source of the problem is irrational thinking in his head. The regulator needs to consult a psychotherapist and attempt to explain to him why is he keeping application developers in a double bind, while simultaneously not wanting to keep the application developers in a double bind.

No false logic. https://www.kernel.org/doc/Documenta...i-nonsense.rst This is the most known example. Here the developer in charge who is the regulator take the problem of stable kernel driver ABI very serous-ally. Problem here it is a true catch-22/ pure double bind.

The fundamental problems problems that happen with stable ABI keep on turning up. Psychotherapist cannot change a fundamental problem.

Yes having all the core Linux kernel developers sitting down with psychotherapist is not going to be able to change anything.

Yes the stable ABI nonsense is that people want a stable in kernel driver ABI without understand the major downsides. Of course third party driver developers want to keep their driver code out of the kernel tree. This has lead to decades of repeated issues with Nvidia binary drivers and other third parties because they would not accept this are fundamental limitations.

You can find other examples with python changes and so on. Where the comes a point where only way to fix issues is break Stable ABI.

Basic thing you keep on missing is that a regulator may have plenty of actions at his disposal but the problem is mutual exclusivity. You choose stable ABI this equals you now have limitations on what internal structures you can change. You cannot have freedom to change internal structures and Stable ABI that depends on those internal structures as they are.

Yes the base to most pure double bind/Irreconcilable Differences are choices that have mutual exclusivity. Lot of what you have been writing is like saying "you can have your cake and eat it as well" this is a physical impossibility because by the rules of nature is mutual exclusive. Software design and development does have fixed set of rules that does force mutual exclusive choices.

Can you show me where the term "double bind" is mentioned in that mailing list? After that, you have to prove that all 10 points from my analysis have already been seriously considered.

Irreconcilable Differences is what is commonly mentioned not the term pure double bind. Also take a close look at your 10 points does any of them work at all once you have true Irreconcilable Differences. No where in your 10 points is acceptance of that.

Yes the first part you stated your fundamental problem no point digging out the references across the LSB mailing list. Thinking the other side is just being stubborn means you don't consider their point of view.

Sometimes when you do consider the other point of view bugger they have had to make a call between X and Y. I would have liked Y but X and Y are mutual exclusive so they can only choose one. This is exactly what happen in the stable ABI nonsense. Now comes acceptance and working out how to live with that choice.

xfcemint you need to lose the idea that regulators have absolutely choice freedom the real world do does not work that way. Any solution to these problems has to have something in the solution to deal with the mutual exclusivity/Irreconcilable Differences/pure double bind. This also means any solution will have to be imperfect. Even windows so call stable ABI is imperfect that why Microsoft has shims to attempt to mask over the imperfections.

[hotaru]

I'm not an expert in this subject, but in my opinion the word "therapist" is too ambiguous. "Therapist" can also mean a phisiotherapist (Wikipedia: physical therapy).

The most precise and appropriate term to descibe the specialist (that I'm thinking of) is "psychiatrist". The source of the problem is in the minds of affected actors, in their thought processes. A psychiatrist is a qualification required to solve those kind of problems (in this situation, the source of the problem is a repetitious pattern of irrational thinking).

no, a psychiatrist is a doctor who specializes in prescribing medication as an alternative to actually solving mental health problems. "therapist" is the correct word to use here.

[oiaohm]

I'm not an expert in this subject, but in my opinion the word "therapist" is too ambiguous. "Therapist" can also mean a phisiotherapist (Wikipedia: physical therapy).

The most precise and appropriate term to descibe the specialist (that I'm thinking of) is "psychiatrist". The source of the problem is in the minds of affected actors, in their thought processes. A psychiatrist is a qualification required to solve those kind of problems (in this situation, the source of the problem is a repetitious pattern of irrational thinking).

psychotherapist is the global term.

It goes like this: Some application developers consult a psychiatrist. The psychatrist then explains the situation and provides some remarks on appropriate language and actions in similar situations. In this specific case, if the regulator refuses to consult a psychiartist and organize discussions, then the application developers can attack the regulator in public by ridiculing him.

Having a psychiatrist sit down and go though the group mental problems and lay the finding of his research what they are in front of the application developers would be helpful. The problem you just wrote attack the regulator. If the regulator happens to be in a catch-22 and has to pick a side between like stable ABI and performance or stable ABI and security and the regulator has a userbase they have to support happens to value security or performance more they are not in the wrong for their user base and should not be attacked.

Application developers are really good at asking for a stable ABI and complaining when stuff does not without accepting that the party not keeping it stable majority of cases are in catch-22. Valve developers maintaining the stable ABI for the steam runtimes learnt this. Do note I said runtimes valve attempted to make a single runtime then found out this caused horrible performance problems as the catch-22 of keeping ABI stable hit. Even Windows if you look closely is multi stable ABI in user space. This is fully not 1 size fits all.

I don't claim to be all-knowing. You might be mostly right, but I think that you are not completely right, and that at least some small part of the stuff that I have been rambling about is actually novel and useful.

Remember the LSB mailing list was at this for 15 years. Sorry you have not come up with anything novel that was not already covered in that mailing list. The hard reality is what was worked out in the LSB mailing list never really got out there.

xfcemint you do demo exactly the problem we are up against with application developers vs distributions. Lack of acceptance that double bind/Irreconcilable Differences exist. At times the distribution stated objectives are counter to-doing stable ABI and for their end users those objectives can be important.

Same problem exists between monolithic kernel design and microkernel designs there are Irreconcilable Differences in places if you stick to pure microkernel or pure monolithic design. The best all rounder is likely to be something developed middle ground taking advantage of monolithic design and microkernel design where it suites and possible managed OS design and maybe a few others.

Managed OS design also promises things that microkernel and monolithic designs don't that is single driver file working cross CPU types without needing the user to have a development suite set up.

Not being willing to accept Irreconcilable Differences is why people do many things wrong.

Xfcemint think you have still been stuck that we have to be able sit down and talk this out. Think of a person in a failed marriage they can throw away decades of their life not really living their life because they have not accepted that they have Irreconcilable Differences. Yes if they had gone their own ways they would be happier and in fact made progress instead of being stuck at stand still.

How to get the application developers wanting to release on Linux to be aware that particular Linux distributions are Irreconcilably Difference to their objectives and the end users of these distributions at times need these differences and other Linux distributions like nixOS or other solutions like flatpak are not as Irreconcilably Difference and this might result in them having to ask their users to jump though a few extra hoops to use their application.

Xfcemint this is the case the stable user space ABI problem with Linux solutions are sitting there. So its not unsolved problem. The problem is acceptance of the solutions and what that mandates how things have to be done. Yes this acceptance includes accepting that different groups will never maintain stable user space ABI because it does not suite their use case and users.

[oiaohm]

I find it likely that you are not sufficiently qualified to put forward such a statement. To get a correct answer, an expert in double bind situations (i.e. a psychiatrist) should be consulted.

I find it likely that you are not sufficiently qualified to put forward such a statement. To get a correct answer, an expert in double bind situations (i.e. a psychiatrist) should be consulted.

I used the word Therapist and Psychologist for reasons. Psychologist study human behavior. Therapist treat human behavior. There is a big difference between the two.

I would also make a recommendation for application developers, that they should also consult a psychiatrist in order to help them in solving the double bind from their side.

This would be Therapist.

Psychologist would be used to study the group behavior. Yes may over look Therapist collected information to confirm if the case is truly pure double blind.

Linus Torvalds with Linux foundation assistance got psychologist involved to study the behavior to work out what was going on. There is a Linux foundation published white paper on this stuff.

Here is another reccommendation:

When the problem of a useful userspace ABI is being discussed, the psychiatrist from both sides should be present, i.e. a regulator's psychiatrist and application developers' psychiatrist should be present. This is to avoid a possible issue of psychiatrists being biased towards one side (the side that approached and hired the psychiatrist).

This is just were you keep on getting terms wrong. These would be Therapists not psychiatrist. Therapists are trained in managing meetings of people to avoid stuff going really wrong. Pure psychiatrists are not. There are reasons why psychiatrist to perform their experiments need ethic boards.

Psychology you learn about motivations. A Pure double bind. Therapist is going to look at the problem Stable ABI and say this is case of "Irreconcilable differences" and not put both sides together instead have both sides enter acceptance to the case.

Yes acceptance that they cannot expect the other party they are in a pure double bind to work with them in particular areas because other parties interests and need don't align and never will. The python developer case keeping the stable ABI end up preventing upgrading the cpython core so hinding performance improvement this is a real world issue. There are many cases with stable ABI where this comes up.

Valve starting the steam runtime came out of acceptance. Flatpak freedesktop runtime comes out of acceptance. NixOS way of doing things come out of acceptance.

Stable ABI for userspace applications on Linux is likely always to be competing solutions. Application developers need to pick what one they think will service their userbase the best and suite the way they want to do development.

Yes there was a study into why Linux Standard Base failed as well. The old Linux Standard Base was like you ignored the existence of the pure double binds and though that getting people to talk enough and comprise enough solution could be found. Yes the Psychology report into why that fail mentions pure double binds.

Linux foundation is quite well funded. When ever Linux foundation believe there was a chance for Therapist to solve these problem they have provided them.

xfcemint; everything you think is a new idea was started being tried 2 decades ago and was attempted for 15 years.

https://en.wikipedia.org/wiki/Linux_Standard_Base

[oiaohm]

I find it amazing that both of the forum's biggest whackos agree on this one.

Note I wrote "pure double bind". Pure being keyword. A pure double bind is a double bind locked in place by "Irreconcilable differences".

a psychological predicament in which a person receives from a single source conflicting messages that allow no appropriate response to be made.

That is a double bind. A pure double bind there is no way to by discussion to solve the conflicting messages and the conflicting messages are in fact all true the problem is the double bind is coming from real "irreconcilable differences".

Solution to "pure double bind" is not more debate but acceptance. ""experts in solving similar problems" go talk to expert and ask them about the "pure double bind" case.

The normal double bind case you can find some flawed statement so breaking the double bind.

Pure double bind cannot be resolved by Therapy or anything else. They are catch-22 that are like the force of gravity that you just have to accept.

Pure double bind are items Therapist and Psychologist cannot alter or resolve or change no matter how much they attempt to assist.

Pure double binds normally are points that you have to agree to disagree and accept that one party is going to go one route and another party is going to go another and never cooperate on this point.

Some ways xfcemint it would be useful to have a list of all the "pure double binds". The points we cannot expect agreement so just have to work around.

When it comes to Psychology double bind comes in two forms. Psychological constructs double binds these are cases that people believe they are in a double bind but after closer inspection their is falsehoods in the ideas behind the double bind so these can be undone. The other is Pure/Real double bind these cannot be undone and are true no win only thing Therapist can do with pure/real double bind is get the party to acceptance. If there is pure/real double bind between two or more individual parties this is Irreconcilable Differences.

Yes solution to Irreconcilable difference is accepting that agreement is not possible and possible going own way.

xfcemint I most likely should have explained the importance of pure. You wrote that getting expert in these field involved could help problem here Linus Torvalds using the Linux foundation already tried that.

The problem here is Stable ABI is a catch-22/pure double bind yes this is why it is a source of Irreconcilable differences between the different parties.

These catch-22 points attempting to get mass agreement mostly end up wasting time.

[oiaohm]

I find it likely that the hidden actor is Linus Torvalds,

Linus Torvalds is not that hidden. Linux Torvalds power does not extend into user-space ABI other than the syscalls the Linux kernel provides.

There are different userspace parties who will not sit down with Linus Torvalds even if he asked. He tried and stated it was a herding cats problem and pointed out that there are a lot of Irreconcilable differences between gtk/qt and so on in libraries then irreconcilable differences between distributions ideas of what is right and wrong.