Author Topic: A rant about lack of support on open source software, or anything else actually  (Read 42878 times)

0 Members and 2 Guests are viewing this topic.

Online peter-hTopic starter

  • Super Contributor
  • ***
  • Posts: 3823
  • Country: gb
  • Doing electronics since the 1960s...
Quote
I can count the entire donations with my two hands.

I happily donate, and have done so.

I think it is a good model. I run a forum (not electronics but techy) which runs entirely on donations. It can be done.
Z80 Z180 Z280 Z8 S8 8031 8051 H8/300 H8/500 80x86 90S1200 32F417
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 14932
  • Country: fr
In expensive, highly profitable commercial products, this boils my blood off. If you use OSS, show some damn gratitude and give'em a small part of you big income.
It's even simpler, in my opinion: if you eat from a pond, make sure the life in it is sustainable; don't count on draining it yourself and moving on before anyone else notices.

Just an example: Years of maintaining/developing the STM32 soldering firmware, helping in the forums because lazy people doesn't RTFM...
I can see the traffic in Github, about 500-900 visits a day and 60-100 unique visitors.
Hey please fix this. Please add this, please help me.
I can count the entire donations with my two hands.
Want support? Sure!
Exactly.  And conversely, if you get the right people the monetary support they need, they can dominate an entire niche.

The problem is that companies and organizations that throw that kind of money around, simply don't have the technical ability or knowledge to pick well.  It is either done at a personal level, or via social interactions.  You never get funded because you're the best around; you only get funded if you get a people-person to promote your work and convince the money-bags it is the right thing to do.

And this is how we end up with something else than a meritocracy in the support of free/open source projects.

If it matters, it is the exact same thing in scientific research.  It is much easier to get funding for research already being done elsewhere, because those are proven popular, and thus low-risk to those who decide who and what gets funded.  The science itself does not matter that much anymore.  Even in industry, a single eccentric billionaire can revolutionize an entire industrial field, if they simply break out from that mold.

Yes. Yes.

The root issue (I think I mentioned that before) is that we humans are an ultra-social species and, apart from exceptions, we need conformity to "fit in". Probably because conformity has helped the species as a whole to survive and build "consistent" cultures. Social animal species everywhere have shown that, beyond us, this appears to be a working model.

Great discoveries and breakthroughs have pretty much always come from people breaking the mold, even when obviously they have always built on previous knowledge. Still, to break through, at some point you need individuals that, at least momentarily, will get out of conformity and accept that.

And with the advent of new technologies, AI and so on, this conformity we still seem to need as much (if probably even more) as we did in the stone age is, IMHO, dragging us down and may eventually lead to our demise.
This "AI" we'll be increasingly relying on is essentially an hyper-extension of this conformity of our societies. It's going to build ultra-conformity.

Speaking of the state of scientific research, this has become a complete disaster. Not new. I started to notice the trend a good 15 years ago, but it's getting worse every year.

As to eccentric billionaires, yeah. Even if a small minority, I actually think there are many "eccentric" people out there, potentially able to break the mold.
But when they're not billionaires, they have to constantly take care of making just enough money to survive (the "rat race"), which, I think, is wasting their potential. Not that I have a solution to that.
 
The following users thanked this post: Nominal Animal

Offline DiTBho

  • Super Contributor
  • ***
  • Posts: 4042
  • Country: gb
Speaking of the state of scientific research, this has become a complete disaster. Not new. I started to notice the trend a good 15 years ago, but it's getting worse every year.

Useless youtubers get 8K euro / month, providing nothing but bullshit videos, whereas serious researchers get less than 3K: fSck!
The same applies to the movie industry: we spend more money making science fiction movies about space travel than we actually invest in space research.

Are you surprised that in 2024 open source still means the above? To me, frankly Agent Smith from the Matrix movie is always right  :o :o :o

It is certain that in 100,000 years "homo sapiens" will no longer be the "only dominant specie" and will have to deal with a different "hominid"(2), otherwise, being so idiotic as we are, it is mathematically certain that we will become all extinct like dinos.

(1) Youtube and TikTok are definitively full of idiots, starting from those who wear an Apple Vision Pro headset and then film themself while driving a vehicle weighing over a ton.

(2) probably with the "ultra-sapiens", provided that we do not have to adapt to Mars' or Titan's climate, in case... the genetic engineering will give a big boost to species diversification.

Hope this will "patch" the current "bugs" with humans  :-//
The opposite of courage is not cowardice, it is conformity. Even a dead fish can go with the flow
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2250
  • Country: br
    • CADT Homepage
In nature progress isn't so much through learning and development but by selection.
At least we should hear if somebody found a piece of open source software useful. I mean for a certain use case. Please post here.

I am currently implementing LwIP patches mentioned in the ST forum in order to improve reliability. It started with their STM32H743 FreeRTOS+LwIP http server example and was working. Except i had some crashes after 1 or 2 hours. Yesterday i thought maybe connection of lab power supply with alligator clips was bad and soldered a 2700uF buffer cap to the 5 V supply of my board. It can sustain 3.3 V supply for about 20 msec. Now during 20 hours with about 1 500 000 packets transferred HTTP timed out twice, but service continues.
Anyway, i looked at the FreeRTOS plus TCP proposals and they provide ready to use STM32H7 examples as well. How about that one?

Regards, Dieter
 

Offline coppice

  • Super Contributor
  • ***
  • Posts: 8966
  • Country: gb
Speaking of the state of scientific research, this has become a complete disaster. Not new. I started to notice the trend a good 15 years ago, but it's getting worse every year.
It has been going downhill a lot longer than that. The whole point of tenure was to allow smart driven people to be unencumbered to research, without being the traditional smart guy wealthy enough to go their own way. However, even 40 years ago I worked with more than person who had left their academic position because life had become depressingly political there .
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2250
  • Country: br
    • CADT Homepage
Please ignore that. coppice could have written his comment months ago..
I was asking about FreeRTOS plus TCP.

Regards, Dieter
 

Online Nominal Animal

  • Super Contributor
  • ***
  • Posts: 6558
  • Country: fi
    • My home page and email address
The root issue (I think I mentioned that before) is that we humans are an ultra-social species and, apart from exceptions, we need conformity to "fit in". Probably because conformity has helped the species as a whole to survive and build "consistent" cultures. Social animal species everywhere have shown that, beyond us, this appears to be a working model.
I believe a large part of humanity is evolving towards eusociality, possibly as a way to avoid the behavioural sink (also known as the fate of Universe 25).

Interestingly, this has direct implications to (and, to some degree, explanations as to how) open source development.

(Do not get bogged down by the eusocial insect biology, i.e. one "queen" (or "queen" and "king" pair as with termites) birthing every member of a hive.  That is just a biological detail of how insects arrived at eusociality.  Human biologists are well aware of that "queen" naming being inaccurate, because the proper name would be "god"; it's just that using the proper name would not fly in human societies.  Also, instead of a single hive, consider a multitude of co-operating and competing hives, with the "gods" forming a separate class, a pantheon.)

The relevant human castes here are toolmakers and engineers and scientists.  Technically speaking, they do not share tools per se, but the knowledge and plans of said tools.  (This is an important distinction to remember, as otherwise one may confuse open source tool development with communism.  The two are orthogonal, and not related in any way other than word games and bad analogies.)  Also note that humans have always been and always will be toolmaking species.  In fact, we evolved from a toolmaking species, so one could say we evolved into a toolmaking species even before we became humans.

We learned long ago that to become better caste members, those who know need to teach the young ones.  It looks like this too is as old as our species is, because at least some of even the oldest cave and rock paintings and carvings were used as a teaching tool (in hunts and ecology).

Open source software is the natural continuation of this in relation to software tools.  We seem to be hard-headed and not understand this, because we keep failing.  For example, some cryptographic knowledge has been re-discovered relatively recently, because the original discoveries were done during the second world war and were kept secret for political reasons.  The battlefield medicine of ancient Romans, especially amputations, was so advanced in terms of survivability rates of comparable wounds, that we managed to lose it for a millenium, and did not regain the same survivability rates until 1800s.
For science and engineering, it seems that Antikythera mechanism indicates the same has happened there too.

Because software tools suddenly became ubiquitous instead of just specialist devices, they have become trade goods for the general population.

The current situation is a result of the conflict between the trader caste and the toolmaker caste, with everybody else caught in the crossfire, often seeing the situation from a single viewpoint and not understanding the other at all.  Even the copyright/copyleft licenses and their differences are easily described in terms of this conflict: GNU GPL shields toolmakers and tool-users against the trader caste by making it impossible for a pure trader to make a profit off such tools, while letting toolmakers sell their wares and designs, and co-operating traders to recoup their reasonable costs.  MIT, BSD, and Apache licenses avoid the conflict by capitulating to the trader caste by noting that they don't really "take" anything, as they just "copy"; and thus whatever the traders do is not directly off from the toolmakers.  Essentially, that the cost of the trader caste making a profit without contributing anything is not worth the conflict.

End users only interacting with traders demand the same service from toolmakers and engineers, and this causes a severe conflict between them.
For example, I've tried a couple of times to explain why the number of end users is nonpositive (i.e., either zero value or negative value) for most open source software developers.  Telling Linux developers that "until you do X, you won't be as popular as Windows" is as sensible as telling a heavy arms manufacturer that unless they convert their long range heavy rifle into a small hand weapon suitable for target practice they won't become as popular as cheese is.  It is pure nonsense.  For open source development, for the toolmaker caste, contributions is what matters, not the number of users.

The lack of support for open source is a direct result of this conflict between traders and toolmakers, with end users desiring support being the civilians caught in the crossfire.

If we look at history, for a hundred years back into mains voltages, it really took legislative standardization effort to rein in the traders.

Here, we are not talking about standardization into a single operating system or any single software suite –– that would be dictatorship, and lose all the benefits we get from open competition; and competition between projects is really what keeps software projects evolving instead of stagnating ––, but about interoperability (and thus also interfaces and format standards).  Currently, the trader caste is exploiting walled gardens and vendor lock-in to capture their clients.  While that benefits them, it really hinders competition and end-user choice; this is what we need legislation to quash.  Perhaps then we might be able to educate the end users and let an ecosystem of support services for open source software projects sprout into life.  As it is right now, the trader caste is not willing to interact with such a new niche at all – and not just with respect to open source software, but also hardware repairs (see right to repair for details).

Please ignore that. coppice could have written his comment months ago..
I was asking about FreeRTOS plus TCP.
No.

Even though it looked like out of the scope of the ongoing thread, it was actually all very much on point: about trying to understand how and why the current situation came to be, and how to change it.

Once again, the most useful information is not just simple answers to the stated questions, but in understanding the situation and reasons underlying the situation and the reasons why things are the way they are, because only then can we escape the inertia of human culture/history and do better.
 
The following users thanked this post: pcprogrammer

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2250
  • Country: br
    • CADT Homepage
Do you feel better now? Have another beer and go to sleep. I'm still waiting for a contribution concerning open source software.
« Last Edit: July 21, 2024, 08:18:06 pm by dietert1 »
 

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 4008
  • Country: nl
Do you feel better now? Have another beer and go to sleep. I'm still waiting for a contribution concerning open source software.

To be honest, you are the one who is out of place in this thread.

If you want answers on your open source problem start a new thread on asking the specific question you have, instead of riding along on a rant on open source software lacking support.

Nominal Animal's post provides a lot of insight into the question why the lack of support is there. As a provider myself I too do not provide real support in what I put out. I write what I write for my own joy and if someone finds it useful that is just a bonus, but when they start bothering me with I need this or that added, it becomes about their joy and there is no benefit in it for me. I'm not talking about financial benefit here.

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8337
  • Country: fi
To be honest, you are the one who is out of place in this thread.

If you want answers on your open source problem start a new thread on asking the specific question you have

Agreed. For a very specific technical question, just open a new thread instead of bumping an old generic discussion one.
 

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2250
  • Country: br
    • CADT Homepage
If there appears a completely ignorant answer three hours after my post while nobody felt the necessity to post for 4 or 5 months, we know exactly what is going on. It's no insight at all. Just destructive, like most of the answers before my post.
The difficulty with getting constructive interaction on open source software is people like you, nothing else. This is my insight.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22019
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
If there appears a completely ignorant answer three hours after my post while nobody felt the necessity to post for 4 or 5 months, we know exactly what is going on. It's no insight at all. Just destructive, like most of the answers before my post.
The difficulty with getting constructive interaction on open source software is people like you, nothing else. This is my insight.

You barge into a dormant thread, which you haven't been involved in; you make some vague statements, then start insulting members who are reacting to this behavior?

I gather you're under some manner of stress, perhaps you're working on a project, it's not going anywhere, it's frustrating, whatever.  I sympathize with that, but we can't exactly help you if you don't even ask a question.

And we certainly won't help you if you're just going to lash out at us.

Perhaps a vacation is needed, clear the mind?  Get some distance from it, mull it over, come back and write a clear and concise statement of your concerns and questions?

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: pcprogrammer

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 8337
  • Country: fi
three hours after my post while nobody felt the necessity to post for 4 or 5 months, we know exactly what is going on.

I don't know if you do, but I do know what exactly is going on, as it happens here on a daily basis:

  • Someone bumps an old, inactive thread for no good reason, when they probably should open a new thread
  • Only this person sees the warning message of the thread being old
  • Others see the thread on the top of the list, and fail to notice the old date
  • They start reading it as if it was a new thread
  • They comment to any random part of the discussion, well, why not?
 
The following users thanked this post: magic, pcprogrammer

Offline coppice

  • Super Contributor
  • ***
  • Posts: 8966
  • Country: gb
If there appears a completely ignorant answer three hours after my post while nobody felt the necessity to post for 4 or 5 months, we know exactly what is going on. It's no insight at all. Just destructive, like most of the answers before my post.
The difficulty with getting constructive interaction on open source software is people like you, nothing else. This is my insight.
The ignorance is entirely on your part. This is a public forum. Nobody is required to help you. If they do, that's a pure bonus. The vast majority of questions asked in forums are either never answered, or answered incorrectly.
 
The following users thanked this post: pcprogrammer

Online Nominal Animal

  • Super Contributor
  • ***
  • Posts: 6558
  • Country: fi
    • My home page and email address
At least we should hear if somebody found a piece of open source software useful. I mean for a certain use case. Please post here.
Basically all of the software I use every day is open source.  You're actually using several right now, because this forum runs on Simple Machines which itself runs in PHP.

I've also contributed to quite a few commonly used open source projects, including the GNU C library and the Linux kernel, over the last three decades or so.  (Not that much, just suggesting fixes to bugs I've drive-by troubleshooted due to one of my users or friends having hit by it, or because I do occasionally fix bugs in open source projects just for fun.)  You can look at my contributions here and elsewhere (I'm the only Nominal Animal as far as I can determine) to make up your mind as to what my experience and reasoning is worth; my earlier stuff I contributed under my real name, which I don't want to openly publish right now (but is easy to find out if you really want).

My opinions are irrelevant, as are the opinions of everybody else; it is the reasons underlying those opinions that can be examined and learned from.

Anyway, i looked at the FreeRTOS plus TCP proposals and they provide ready to use STM32H7 examples as well.
I don't use ST microcontrollers myself; I prefer NXP ones.

For the Teensy 4.1 I do have (based on NXP i.MX RT1062, with DP83825 10/100 ethernet PHY, software developed in the Arduino environment using the Teensyduino add-on), there are two TCP/IP libraries immediately available: NativeEthernet based on FNET, and QNEthernet based on lwIP (but does not require an RTOS).  The former is included as part of Teensyduino, and the latter is available in the Arduino library manager.

I don't really like any of the ones I have looked at, but thus far haven't had a reason to do the huge amount of work to develop and test and verify my own implementation, either.

How about that one?
Is that a question?  How about that one what?

Have another beer and go to sleep. I'm still waiting for a contribution concerning open source software.
I don't drink or consume anything stronger than sugar and caffeine right now.  Psychological projection, perhaps?

Are you new to free/open source development?  Your attitude is prevalent among those having difficulty understanding how the open source ecosystem works, and makes it very difficult to interact with you in any mutually beneficial manner.  In fact, you seem to be demanding performance from others, and trying to stop others from interacting in the thread you have now claimed as yours, exactly like a noncontributing end user demanding support services from developers having no contractual or even implicit agreement with that user.  (It is most prevalent among Windows users who are forced to switch to using Linux, or switch to Linux because they want the zero cost option but feel they should still be treated as a paying customer.)

If you understood how open source and mutually beneficial social interactions work in practice, you'd know that you need to contribute something useful first, and show that interacting with you has at least potential benefits; and staking something as "yours" and demand others to go elsewhere will simply lead to a project fork –– or in forums like this one, you being added to Ignore Lists.

You can find your own Ignore list by clicking on Profile or Profile > Summary, then hovering on top of Modify profile, Buddies/Ignore lists..., Edit Ignore List.  I use it to hide posts by members I feel I cannot mutually beneficially interact with, pruning it every month or two.

The vast majority of questions asked in forums are either never answered, or answered incorrectly.
Considering that the vast majority of questions asked in forums are incorrect too (that is, do not address the actual underlying problem to be solved, and instead ask for help with a pre-chosen sub-optimal solution approach), the most useful interactions tend to require delving deeper into the context and actual problem than is possible to provide in a single post or message.

In other words, the threads that examine the problem deeper and have some "meandering sub-threads" in them, tend to have some very useful and important nuggets of experience and knowledge in them, although they are rarely (approximately never) exactly about what was asked about in the initial post.  It usually takes iterating over a few messages before we get into the actual interesting meat of the matter at hand.
 
The following users thanked this post: MK14, pcprogrammer

Offline coppice

  • Super Contributor
  • ***
  • Posts: 8966
  • Country: gb
The vast majority of questions asked in forums are either never answered, or answered incorrectly.
Considering that the vast majority of questions asked in forums are incorrect too (that is, do not address the actual underlying problem to be solved, and instead ask for help with a pre-chosen sub-optimal solution approach), the most useful interactions tend to require delving deeper into the context and actual problem than is possible to provide in a single post or message.
You mean like someone asking a very specific question about a TCP stack in a very general thread, who then starts complaining about people posting general messages in the general thread?
 
The following users thanked this post: Siwastaja

Offline DiTBho

  • Super Contributor
  • ***
  • Posts: 4042
  • Country: gb
openH264 cannot be used for any serious professional video editing.
Ovs has plug-ins, and you need to install a library with propietary codec.

I wonder if things will ever change...  :-//

The opposite of courage is not cowardice, it is conformity. Even a dead fish can go with the flow
 

Online Nominal Animal

  • Super Contributor
  • ***
  • Posts: 6558
  • Country: fi
    • My home page and email address
The vast majority of questions asked in forums are either never answered, or answered incorrectly.
Considering that the vast majority of questions asked in forums are incorrect too (that is, do not address the actual underlying problem to be solved, and instead ask for help with a pre-chosen sub-optimal solution approach), the most useful interactions tend to require delving deeper into the context and actual problem than is possible to provide in a single post or message.
You mean like someone asking a very specific question about a TCP stack in a very general thread, who then starts complaining about people posting general messages in the general thread?
(;), but for the sake of discussion and others reading this thread later on, I'll act as if that was not a rhetorical question.)

No, only that keeping to the exact narrow definition of the initial post is not useful either, because the questions just about always needs clarifying or mutation before a proper solution or alternate approach is discussed.

Of course, threads tend to wander, and sometimes completely derail.  That is unfortunate, and is likely not very useful or informative, but it is what happens with humans and social interaction.  :-//

:-/O To expand a bit, not keeping to the exact narrow definition of the initial post means that when discussing lack of support in open source software projects, we shouldn't just gawk at the situation and describe to each other how horrible and annoying it is; that discussing why it is so, and what we may need to do to change the situation is more useful and informative.  It may not help with the initially stated problem – lack of support – but it can open up alternative, perhaps better approaches to break out of the problem.

As a practical example, consider a programming question where someone is asking for help in speeding up their bubble sort implementation, because it turns out their application is taking too long to start up and is annoying the users.  Strictly keeping to the question is not useful, because other sort algorithms are superior; suggesting a different sort algorithm, perhaps quicksort, would be more useful than helping with inconsequential speedups with the wrong choice of algorithm.  In some cases, the sort itself can be eliminated: instead of reading all data and then sorting it, one can read the data into a self-sorting structure (tree, heap, et cetera) and use more CPU time but less real world time to have the sorted data available (because transfer of data from mass storage is still slower than sorting it, doing slightly more work done while reading the data instead of waiting for the data to arrive before working on it).  Thus, answering the stated question directly may actually hinder the original asker, but delving deeper into the topic at hand can be extremely useful.

If Dave suddenly decided that threads should avoid discussions and only answer the stated questions, like StackExchange network does, I'd be out of here in a second.  It is the discussions and interaction and experiences and reasons for opinions, not simple answers, I personally am here for.  And to try and help, in my own somewhat crude and unintentionally uncouth fashion.
 

Online Nominal Animal

  • Super Contributor
  • ***
  • Posts: 6558
  • Country: fi
    • My home page and email address
openH264 cannot be used for any serious professional video editing.
Could you use say x264 (with/via say libavcodec) instead?  I do believe it is well suited for professional video editing, at least quality and speed-wise, if supported by your editing suite or tools.

OpenH264 is a Cisco project, and Cisco has had difficulty understanding open source licensing (FSF sued Cisco for GPL violations in 2008).

Here, the problem of OpenH264 project support is most likely solved by using a different project, that implements the same H.264 video encoding and decoding standard.  This is exactly why I mentioned above that competition keeps software projects evolving instead of stagnating: if OpenH264 was the only implementation, we'd all be suffering.  (And that instead of standardizing on specific projects, we need to standardize the formats and interfaces those projects implement.)
 
The following users thanked this post: DiTBho

Online dietert1

  • Super Contributor
  • ***
  • Posts: 2250
  • Country: br
    • CADT Homepage
As i wrote above: Nature advances more by selection than by learning or development. There are certainly better places than this forum if it is about open source software. There are other places to learn about FreeRTOS plus TCP and i am going there. It's not about this thread, it's about this forum and bullying.

Regards, Dieter
« Last Edit: July 23, 2024, 07:26:04 am by dietert1 »
 

Online Nominal Animal

  • Super Contributor
  • ***
  • Posts: 6558
  • Country: fi
    • My home page and email address
As i wrote above: Nature advances more by selection than by learning or development.
That is an oversimplification.  Evolution can be modeled as genomes learning about the effectiveness of specific genes via experimentation – and some evolutionary biologists do so.

For example, it makes no sense for K-strategist individuals to care for their offspring, because it takes resources that individual could use to prolong and enhance their own life.  However, if we look at that individual being driven by a need to propagate their own particular genome, then the models work.  It also explains why e.g. squirrels often adopt the offspring of close relatives (not just "siblings", but "cousins" and "second cousins" too) when they lose their own parents.

In information-theoretical sense, there isn't enough difference between "learning" and "selection", really; it is like examining the same waveform in the time domain, and think the waveform is a different one just because you examine it in the frequency domain.

There are other places to learn about FreeRTOS plus TCP and i am going there. It's not about this thread, it's about this forum and bullying.
As in, your bullying to get your needs fulfilled isn't working effectively enough here, so you'll be going elsewhere with victims that are easier to bully?

I do believe adjusting your communication style (see e.g. here) would yield better overall results in the long run.  I do not say this to annoy you, and I am not being snide here: it provably really works, that's all.

It is utterly unrealistic to demand everybody else adjust the way they communicate, no matter how unjust/unfair/uncomfortable the way others communicate might be to you.  The only practically proven way to change the outcomes from communications is to adjust your own way of communicating, until you get better results.  I do so constantly myself.  Even though it does take a lot of effort, and can feel unfair/unnecessary/waste of effort, reality beats theory and even the most fervent wishes and desires: we have to do the work to get the results, and do so in the real world.
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 14932
  • Country: fr
As i wrote above: Nature advances more by selection than by learning or development.

Nominal answered better than I will, but I just don't see what makes you think that selection is not learning. And what you mean by development exactly.

Actually, selection is a form of learning - also one of the earliest forms of machine learning that we used, incidentally.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22019
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
On a long enough time scale, random generation and selection is indistinguishable from conscious design.

This principle is understandable even in fairly casual circles: randomizers are video game mods that shuffle around progression goals (items, upgrades, keys, etc.; among other things, up to and including characters, enemies, the entire world map..!), to bring a fresh, sometimes frustrating, often surprising or amusing, experience to an otherwise well-trodden old and enjoyed game.  (Placements are driven by a logical restriction algorithm, so that, barring bugs in generation, every random seed is completeable.  Unless chosen otherwise, of course; many games support different placement rules.)  One of the memes in the community concerns plandos ("planned rando"), a manually-placed version, in part or in full.  As you might imagine, plandos often have troll placements: paradoxes like key items locked behind themselves ("key" being relative: perhaps completion isn't gated by that particular item, perhaps its requirement can be bypassed in certain places with glitches), annoyances like major items placed in distant, circuitous locations; narrow quest lines (get item to get item to get...), or quest lines requiring many items/actions to unlock (defeat N bosses to progress), etc..

Well, it's understandable that such trolly quest lines are generated naturally, randomly, from time to time -- very rarely of course.  Typical statistics might be, say a few hundred item locations to shuffle, most of which are junk (money, ammo, health..), and a couple dozen of which are unique progression items (or N-repeated upgrades, where M < N level might be required for various locks).  And maybe there's one or two dozen particularly distant locations those key items could be placed in.  So it's, you know, one in a thousand, ten thousand, that sort of thing to get a random seed that's particularly annoying.  Without a selection bias, that's fine, indeed that's the fun of it.  But here's the kicker: when various people are generating and playing these seeds, and sharing the exceptional ones among themselves, bingo, selection occurs.  If people share only the most memorable ones, it's probably for some troll reason.

And if you wanted to sit there and grind through thousands, millions of seeds yourself, however many you want to try, you will eventually find your perfect troll of choice to share around.  And no one can complain that your seed is a "plando" because you can honestly and legitimately swear you've never touched it -- it might be shared as nothing more than a hash code from a generator website for example, where all the generation settings are visible to the player.

Or, for a simpler and more abstract case: bogosort could perfectly sort an array in a single operation.  It's extraordinarily unlikely to, for any practical length input -- but it's never impossible (given sufficiently true randomization).  A sort operation is an extremely particular selection process, but, again, given enough time, anything is possible, and it will, eventually, even if it takes the heat death of the universe -- find that solution.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Online Nominal Animal

  • Super Contributor
  • ***
  • Posts: 6558
  • Country: fi
    • My home page and email address
For open source software, it is competition that "implements" evolution, selection, learning, et cetera.

However, it is not competition for end users, because as I exaplained above, end users are actually nonpositive factor –– they create more work without contributing anything back to the project itself; a burden, not a resource –– for typical free open source projects.  It is competition in development time and effort: in contributions.  Fairness, exchanges (trading time and development effort) is at the core of this.

The stagnation/ossification among GCC developers prior to LLVM/Clang, and how currently both compiler projects are thriving, similar to competition between Apache and Nginx web servers, shows how this works in practice.  It works even when many of the developers are backed by companies with strict focus requirements, because they operate under the same competitive environment, just different focus.  (I personally highly appreciate the fact that Linux is not the only free open source Unix-like OS out there, that FreeBSD, OpenBSD and others are still in development, too; and happily work on mixed open source and proprietary projects while abiding by all of the licenses to both the letter and intent.  It is not nearly as restrictive environment to do business in as many believe it is; it is just very different to the standard off-the-shelf product development and sales stuff.)

The lack of support service businesses for open source software is a socio-cultural issue, not a technical one per se; and I assert that understanding this (situation and how it came to be, and what drives software projects forwards and helps them thrive and become better) is a key to solving or at least dealing with all this.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf