A rant about lack of support on open source software, or anything else actually

[peter-h]

So, how would one stay in contact with customers?

Fax?

Customer input is the best (arguably the only, nowadays) form of market research.

If you listen only to your few huge OEM customers, you end up making only weird chips with weird features and not much general applicability. You will sell development tools for $20k (instead of the free Cube IDE for example) - like e.g. Nokia did if you wanted to develop apps with the highest access privilege - because every OEM can pay 20k without a thought.

Forums are a powerful channel but need to be used correctly. Most vendors use them incorrectly and then complain that they get mostly damage. No surprise!

[brucehoult]

And Arduino, being as open and accessible as it is, has been ported to dozens of platforms.  You can have the full power (well, minus library cruft) of a Cortex M4+ (or more? I don't even know what the most powerful MCU is, supporting Arduino*..), or as little as an AVR.

Well, at least the Teensy 4.x, with 600 MHz Cortex M7 which is actually perfectly happy if you choose 960 MHz in the Arduino IDE. It's actually faster (single core) than a Raspberry Pi 2.

There is probably faster. See below.

*Well, probably something like rPi, though that'd be going through OS calls to set up peripherals, GPIOs, etc., not library drivers.  Hm, maybe it's not actually implemented on that, actually...

You could run bare-metal on a Pi ... except the boot process is highly not documented.

Would also seem like -- if you can get your product out at all -- now would be a great time to flood those bottom tiers with products.  This could be a great opportunity for RISC-V, say?  Since there are, as yet, maybe not so many production applications for them (I would guess..?!), maybe they could raise production and gain new adopters.  Is there Arduino on any yet?

The very first RISC-V chip and board ever offered for sale -- the HiFive1, in December 2016 -- came with support for the Arduino IDE.

A few days after mine finally arrived in Moscow I did this, using pure Arduino C++ code including the infamous digitalWrite() to implement a 1-bit DAC, toggling a GPIO (or not) once per µs.



That wasn't using all the SPI flash, so an hour later I did this:

SBCs to rival the rPi's?

Noting at the outset that other SBC manufacturers using ARM aren't able to match Pi prices...

RISC-V SBCs of approximately Pi 3 performance have been available for more than 4 years, though at higher (and steadily falling) prices.

I've tested some using a simple benchmark I originally wrote to compare Pi 3 and Odroid C2 and XU4 with x86 machines in 2016 (https://hoult.org/primes.txt)

An extract from the results:

Code: [Select]

11.190 sec Pi4 Cortex A72 @ 1.5 GHz T32          232 bytes  16.8 billion clocks
 11.445 sec Odroid XU4 A15 @ 2 GHz T32            204 bytes  22.9 billion clocks
 12.115 sec Pi4 Cortex A72 @ 1.5 GHz A64          300 bytes  18.2 billion clocks
 12.605 sec Pi4 Cortex A72 @ 1.5 GHz A32          300 bytes  18.9 billion clocks
 15.298 sec HiFive Unmatched RISC-V U74 @ 1.5 GHz 250 bytes  22.9 billion clocks
 19.500 sec Odroid C2 A53 @ 1.536 GHz A64         276 bytes  30.0 billion clocks
 23.940 sec Odroid C2 A53 @ 1.536 GHz T32         204 bytes  36.8 billion clocks
 27.196 sec Teensy 4.0 Cortex M7 @ 960 MHz        228 bytes  26.1 billion clocks
 27.480 sec HiFive Unleashed RISCV U54 @ 1.45 GHz 228 bytes  39.8 billion clocks
 30.420 sec Pi3 Cortex A53 @ 1.2 GHz T32          204 bytes  36.5 billion clocks
 36.652 sec Allwinner D1 C906 RV64 @ 1.008 GHz    224 bytes  36.9 billion clocks
 39.840 sec HiFive Unl RISCV U54 @ 1.0 GHz        228 bytes  39.8 billion clocks
 43.516 sec Teensy 4.0 Cortex M7 @ 600 MHz        228 bytes  26.1 billion clocks
 47.910 sec Pi2 Cortex A7 @ 900 MHz T32           204 bytes  42.1 billion clocks
112.163 sec HiFive1 RISCV E31 @ 320 MHz           178 bytes  35.9 billion clocks

Prices for the HiFive boards:


Dec 2016: $ 59 HiFive1

Mar 2018: $999 HiFive Unleashed (quad core, 8 GB RAM, SD card, gigE, UART .. and that's *it*)

May 2021: $665 HiFive Unmatched (quad core, 16 GB RAM, Mini ITX, PCIe, M.2)

In December 2021 the "VisionFive v1" came out in China for $180 with 8 GB RAM. It has the same U74 cores as the HiFive Unmatched, but only two of them.

Pine64 have pre-announced the Star64, with quad 1.5 GHz U74 cores. They say it will be the same physical form factor, and about the same performance, and price as their ARM-based Quartz64 RK3566 quad A55 board i.e. $60 with 4 GB, $80 with 8 GB. They say it should be shipping in a few months.

Within a couple of months we can expect to see demos of Intel's "Horse Creek" platform with SiFive P550 cores which should come in I think similar (or maybe better) as the RK3588 which has recently started to appear in ARM SBCs around $250. SiFive will release a board using this as the replacement for the HiFive Unmatched.

So that leapfrogs the Pi 4 level of board.

[westfw]

Arduino for example, you can dismiss it if you like, as some unprofessional hack-job -- but the fact remains, its presence affects professionals.

For example, prior to Arduino, I would have rated PICs (8bit PICs) and AVRs as "approximately similar" in the level of interest they received from hobbyists (both beginning and advanced) and probably educational settings.
Social-media wise, there was the PICList Mailing list, and there was the AVRFreaks forum.  Both were pretty good, with a number of helpful experts, and a lot of interesting material.  Not always perfect - regular complaints about condescending attitudes and "meanness", and people who left when their behavior was criticized - but ... very useful.
You could buy the more popular chips and some development boards from a relatively large swath of hobby vendors, and the less popular chips from Digikey or other "mid-tier" distributors.  They were reasonably priced.
The base software (ie an assemler) was free, and there was more powerful SW that you could buy.

Post-Arduino, the AVR has pulled way ahead.  PICList is practically empty, AVRFreaks is 30% Arduino questions, the Arduino forums get 100+ message/day.  Plus ancillary support from other vendors with their own forums (Adafruit, Sparkfun, etc), plus stackexchange and etc.

Typical social media for product support? Are you serious?

That depends on what you consider "social media."  Instagram, SnapChat, and TikTok?  You've got to be kidding.  Twitter and Facebook?  Probably not.  (although the Z80MBC2 group on FB has been pretty good.)  Standalone forums like Arduino, Adafruit, Sparkfun, AVRFreaks, EEVBlog, individual vendor forums?  Those can work really well; *I* consider them "social media."  Email lists?  ditto.

[westfw]

That's what I think about two of the Olimex's Ethernet chips

Did you mean WizNet?  I didn't think Olimex DID "chips."

Given the number of their customers, 1 or 2 would be like pissing in the sea while making the life of those engineers miserable.

If I look at the support forums from NXP and NVidia, the number of support engineers is just that. 2 persons (maybe 3) that have some access to the software and hardware development team. ...   1 or 2 support engineers can make a lot of difference.

2 or 3 people dedicated to servicing the forum can make a big difference, if they have sufficient experience.  2 or 3 applications engineers TOTAL (or "per region") seems to be more typical, and having them spread their attention amongst phone calls, email, in-person visits, and "social media" is a recipe for disaster.
And the companies that DO have separate people for social media seem to think that it's either a "marketing duty" or a job for beginners (who then graduate to the hands-on AE positions if they learn enough.)  More disaster.

I think the vendors ought to hand out "gifts" to those retired grumps who are on forums like here complaining "kids today don't know anything, but I can try to teach some of them, I guess."  I've gotten some occasional swag like that - it's very gratifying.  It'd be moreso if it were closer to the tax-related "gifting limit."  :-)  (yeah, I dunno what that would mean for the vendors' accounting.  "Gifts" are probably less attractive than "salaries", tax-wise, for the giver.)

The mistake I made, probably 15 years ago, was not going up the ETH learning curve then.  ... Which CPU? Not sure; was the ARM32 32F4-like stuff around then?

Let's see.  15 years ago.  2007. Pre-Arduino.
The first ARM Cortex M-3 was announced in 2006 - a Luminary Micro 20MHz CM3 with 8k flash and 2k RAM.
The Stellaris LM3S6432 microcontroller with built-in ethernet (including Phy!) (86k flash, 32k RAM) is probably more the sort of thing you would have wanted to use.  That looks like it came out in 2009.
You could have gotten the ENC28J60 Ethernet controller from Microchip - that looks like it came out in about 2006.

Those all a bit tiny for a modern TCP/IP stack, IMO.  Wishful thinking on the silicon vendors' part.  The whole "IoT" thing is relatively new, driven by the drops in silicon pricing for "reasonable sized" processors and network connectivity.  Cisco apparently claims that IoT was "born" in 2008-2009; I'd guess that was about the time that "things" passed "technical curiosity" and became "possibly profit-generating product."  It certainly wasn't the case in the early days of networking, even if CMU did connect a vending machine to ARPANet back in 1982 (probably costing $1000+ in 1982 dollars.)

[peter-h]

Let's see.  15 years ago.  2007. Pre-Arduino.
The first ARM Cortex M-3 was announced in 2006 - a Luminary Micro 20MHz CM3 with 8k flash and 2k RAM.
The Stellaris LM3S6432 microcontroller with built-in ethernet (including Phy!) (86k flash, 32k RAM) is probably more the sort of thing you would have wanted to use.  That looks like it came out in 2009.
You could have gotten the ENC28J60 Ethernet controller from Microchip - that looks like it came out in about 2006.

I think the only way back then was one of those modules. I don't recall the PNs but recall seeing some.

This stuff existed a long time ago. About 15 years ago I brought out a product based on a programmable protocol converter I was doing then (and still do) which did a sort of monitoring function, with capabilities to SMS a service engineer, different people at different times, etc. It used a Siemens TC35 modem. A stupid idea; should have been done with a self contained module. Siemens dropped the TC35, it became Cinterion, then they went pop and it became something else. It did SMS, GSM dial-up, GSM fax (yes that was implemented too, with fonts etc!) and email. The email was with SMTP so this thing must have had a partial TCP/IP stack. There was DNS. The interface to my system was via RS232, Hayes AT commands with nonstandard extensions. The CPU was probably an 8051 or similar. The TC35 is really old - 20 years? I recall looking at TCP/IP in the Z180 days, late 1980s, and it was about 64k of code, which made it hard (not impossible, with bank switching, especially in an RTOS). Almost nobody did it back then.

I am sure Olimex don't do "chips" - they might get some masked ROM CPUs "overprinted" I would never buy such a product. It is like the unusual LCDs, made by just one German company.

Anyway, digressing.

That depends on what you consider "social media."  Instagram, SnapChat, and TikTok?  You've got to be kidding.  Twitter and Facebook?  Probably not.  (although the Z80MBC2 group on FB has been pretty good.)  Standalone forums like Arduino, Adafruit, Sparkfun, AVRFreaks, EEVBlog, individual vendor forums?  Those can work really well; *I* consider them "social media."  Email lists?  ditto.

Yes, except mailing lists are really no good now. The ones I have been on seem totally dead. All of them. WWW forums is the way to go, because you get SEO, easy archive searches, a proper discussion capability. But it needs to be well set up and well run. Some forum software is utter garbage e.g. the ST forum https://community.st.com/ which is horrible. Only an idiot would have set up a tech forum which requires you to click on "view more" after a few posts. EEVBLOG (PHP-BB) is ok but needs a good setup and some paid admins doing constant patching because PHP-BB has been an easy hacking target.

[westfw]

I think the only way back then was one of those modules.

Yeah.  My EMail Archives say that the Lantronix X-Port, which was the earliest "single port serial server" that I can remember, dates back to at least 2003.  It looks like the WizNet W5100 existed in 2007 (but I don't know if normal people could actually GET one.)

Of course, there were a lot of vendors doing multi-port "terminal servers" that could be operated in reverse, or "milking machine" mode, dating back to mid 1980s.  Cost per port was a bit excessive, and by the mid 2000s they tended to be focused on many-port dial-in Internet Service (SLIP/PPP) and built-in 56k modems, rather than driving the price down on the "simpler" serial services.  Sigh.

[peter-h]

I was reminded of this thread when writing this
https://www.eevblog.com/forum/programming/best-thread-safe-printf-and-why-does-printf-need-the-heap-for-f-etc/msg4357471/#msg4357471
OK, ST didn't supply the source code so you can't call it "open source" but it was still junk. And 99% of users will never find out. The stuff will just occassionally crash

[tellurium]

i was telling my wife that furniture is next. Want to sit on your chair to eat lunch? Pay. A big screw will come out the middle so you can't sit on it. pay per hour . if you forget the subscription the chair will literally screw you. Next toilets : they will weigh your morning core dump and charge you per gram , notifying the doctor and pharmacist what pill to send because the color and consistency is off. Al wirelessly toyour phone. The application will be called BlueStool (instead of bluetooth)

IMO that is a reality long time now, with all these hidden taxes like VAT. Every time you pay, you get shaved off slightly (or not so slightly). And people will be creative in making you pay more often.

But hey, wasn't that the case centuries ago? When people paid taxes for every window (want sunlight in your hourse, pay!) ? For crossing a bridge? Or even on entering a neighbor town?

[5U4GB]

And those who buy 10-20 chips often require the most attention.

We've run into that, a customer that wanted support for something where they were essentially writing their code via our free-if-not-abused-too-much support system.  These guys were probably generating more support traffic than all other customers combined until the licensing guy told them they'd used up their free quota and would have to pay from now on.  Suddenly their developers became much more competent and didn't need a lot of support any more.

[westfw]

a customer that wanted support for something where they were essentially writing their code via our free-if-not-abused-too-much support system.

That's why vendors should be (and frequently are) so supportive of things like Arduino. Or the hobbyist market and/or forums with little company participation in general.   If you can get the "users" to support each other (for free!), everybody wins!

[SiliconWizard]

And those who buy 10-20 chips often require the most attention.

We've run into that, a customer that wanted support for something where they were essentially writing their code via our free-if-not-abused-too-much support system.  These guys were probably generating more support traffic than all other customers combined until the licensing guy told them they'd used up their free quota and would have to pay from now on.  Suddenly their developers became much more competent and didn't need a lot of support any more.

That's a well-known, and general issue in business in general.
Clients that generate the less revenue usually take the most time and effort.

But that's true outside of a business relationship as well to some extent. The less effort (not necessarily money) someone has to make to get something from you and the more they are likely to abuse.

[Nominal Animal]

And those who buy 10-20 chips often require the most attention.

We've run into that, a customer that wanted support for something where they were essentially writing their code via our free-if-not-abused-too-much support system.  These guys were probably generating more support traffic than all other customers combined until the licensing guy told them they'd used up their free quota and would have to pay from now on.  Suddenly their developers became much more competent and didn't need a lot of support any more.

That's a well-known, and general issue in business in general.

Or more broadly, just one facet of the Pareto principle.

[5U4GB]

It's really quite variable, most of our clients are pretty good, ranging from frighteningly competent (a large well-known test equipment manufacturer) through to reasonably OK but maybe a bit out of their depth and in need of some advice, so these guys were really the exception.  Having said that our stuff is highly tuned towards making it easy to use/minimising our tech support load, so most of what comes in is intelligent questions where it's not a chore to answer them.

[Nominal Animal]

It's really quite variable, most of our clients are pretty good, ranging from frighteningly competent (a large well-known test equipment manufacturer) through to reasonably OK but maybe a bit out of their depth and in need of some advice, so these guys were really the exception.  Having said that our stuff is highly tuned towards making it easy to use/minimising our tech support load, so most of what comes in is intelligent questions where it's not a chore to answer them.

Yeah, the 80%/20%, or one fifth, isn't precise at all, it's the idea how effort and results seem to correlate whenever humans are involved.

In your case, it just means that it is normal to have a small fraction of clients that consume a relatively large fraction of support resources.  It sounds like you monitored the issue and dealt with it effectively, and I do applaud that –– and I would even if I myself were that client.

My point is, such cases are to be expected because of humans, regardless of how good your documentation and non-human-interactive support is.
Those who are not aware of the Pareto effect, may think that because they have a few such users, their documentation etc. is insufficient; but it may just be human nature instead.  Similarly, those who only listen to clients who contact support, should remember that just because the clients ask for something, it may not be what they need or even want.

With regards to open source software, most people do not grok that the free/open refers to libre, not zero cost.  (I am not referring to OP here, but to humans using open source tools in general.)
Free/open source software definitely evolves in a market of sorts, and you don't get support –– or anything, really, except for a copy of the software itself –– for no cost.  It is just that typically, the cost is paid in time and effort by knowledgeable people, instead of money.  This means that to get support with LwIP, MbedTLS, etc., you need to have contacts with such people, and have some way of convincing them to help you.  Smallish sums of cash money are often considered more a nuisance than help, because the knowledgeable ones tend to already have well-paying jobs, and an exchange of money only makes things more complicated (especially the direct relationship, expectations, timetables –– contracts, really).  So, even if you do have a reasonable budget, it is definitely not straightforward at all.

The end result is that people like OP find it very difficult to obtain support, because we haven't yet discovered the business models that can bridge things here in a mutually beneficial manner.

In fact, based on what I know about the business side (used to run an IT company myself for a few years around the turn of the century; did well, but it burned me out mentally) and about the FOSS world, I suspect such business models have to be based on nonprofit model.  I do not like the direction companies like RedHat (owned by IBM) or Canonical (Ubuntu) are taking their development and support, but Linux Foundation, Apache Software Foundation, Free Software Foundation, and other nonprofits in this area seem to work.

I'm guessing what OP and others' really need, is a credible foundation-type nonprofit, that provides access to knowledgeable developers for support on basically any FOSS project; roving problem-solvers and bug-hunters, if you will.  You don't actually pay the developers directly, but support the foundation, which in turn supports the developers.  Since the developers do not work for the people asking for support, but for the foundation, there are ways to resolve the "not very competent client demanding excessive support" problem, which a direct relationship does not allow (except by dropping the client).
Problem is, the developers and learners who'd love to be part of that even part-time, and the business/admin people (with enough spine to not actually exploit the structure for their own personal financial benefit) who could run such an organization, just do not meet in the current world.

Or perhaps I am utterly wrong, as I sometimes am. 

[5U4GB]

The end result is that people like OP find it very difficult to obtain support, because we haven't yet discovered the business models that can bridge things here in a mutually beneficial manner.

The freemium business model is good for this, free but minimal support vs. non-free but commercial support.  There are even specific licenses like Sleepycat for this, the user gets to choose either GPL and no cost or commercial but with support service.

[madires]

Several small and large companies are making their money by providing professional support for open source software, sometimes also offering an enhanced version (more features, etc). In many cases those companies were founded by the original authors of the open source software. It's not a new business model.

[coppice]

And those who buy 10-20 chips often require the most attention.

We've run into that, a customer that wanted support for something where they were essentially writing their code via our free-if-not-abused-too-much support system.  These guys were probably generating more support traffic than all other customers combined until the licensing guy told them they'd used up their free quota and would have to pay from now on.  Suddenly their developers became much more competent and didn't need a lot of support any more.

This is a complex issue. Most of those high support/low volume people are a major drag on resources. However, among them are often gems that help you sort out your own systems, so that support runs much more smoothly for your big money making customers. Especially since those low volume people are often the ones using the latest bells and whistles, where your own support processes have yet to mature.

[westfw]

those low volume people are often the ones using the latest bells and whistles

The people playing with the latest and greatest chips are pretty indistinguishable from hobbyists, unless the chip was targeted at a specific company/project.

[Nominal Animal]

Several small and large companies are making their money by providing professional support for open source software, sometimes also offering an enhanced version (more features, etc). In many cases those companies were founded by the original authors of the open source software. It's not a new business model.

True; but I was talking about the case where the original authors do not want or cannot provide professional support.  It is actually quite rare for FOSS libraries' authors to do that.  When they can, it is obviously a good, working model.

We do have a lot of "roving" know-how that could do that kind of stuff, but do not want or cannot do (because of their contracts) such work as freelance, nor set up a company to do that.  Myself included.  I tend to do that sort of stuff just to help out a select few friends.  "Fixing" lwIP or MbedTLS is a long-term project, and would need a dedicated team; it's outside the scale of a few days' to fix a bug or missing feature...

those low volume people are often the ones using the latest bells and whistles

The people playing with the latest and greatest chips are pretty indistinguishable from hobbyists, unless the chip was targeted at a specific company/project.

And some of those will end up using their favourite chips in later mass-produced designs.  Early market capture via hobbyists and education, et cetera.

Of course, some != all.  Perhaps 80% of those are just hobbyists all the way, and only 20% serious designers somewhat later?  (Those numbers being just examples!) It makes calculating the actual cost/return ratio on hobbyist-type support for new chips very, very difficult.

[std]

I quickly learnt that you end up working alone.
No support from ST (32F417). They have a forum which is full of desperate people, no ST presence (they are only a €14BN company, so they can't afford to support their products) and virtually no useful replies from anybody. Just one guy occassionally posts good stuff but only after he's told you that you are a complete moron. They do loads of videos which are barely legible (very poor English).

Hello Peter,
I remember one of your posts on the "ST community" forum, where you were rightly surprised by the ridiculous UART Receive functions
It was probably you, I remember Z80 well in the signature (with which I have a lot of connections, I wrote assembler for several years).

So, practical question - is their forum is right to ask any questions?  I mean did you ever get any help at all? Is there anything useful there or is their forum completely pointless, should I ignore it?

No support on LWIP. There is a mailing list which is dead. Vast amounts of bogus info posted everywhere.
No support on MbedTLS. There is a mailing list which is dead.
No support on FreeRTOS. There is a mailing list which is dead. Fortunately FR runs well.
No support on FatFS. There is a mailing list which is dead. Fortunately FF runs well.

How did you determine this? Did you ask questions after they didn’t answer you, or did you determine this by observation?

How do other people manage?

The main question is -
Did you manage to identify the fishing spots? Sites and people who are really useful in developing the things listed above.

And it was never like this in the past, when chips were simpler. I developed literally hundreds of products over 40 years, all working alone. Just did it from data books. No internet. The problem is the new functionality. It is so massively more complex.

When we worked earlier, we did not imagine that there was help and information out there somewhere. Therefore, we had to do everything mostly ourselves. The Internet is tempting and therefore you waste time trying to find some useful information. All the time is spent searching. But the catch is that there is 95-97% garbage. I noticed that I greatly reduced the loss of time when instead of Googling solutions to problems I began to sort them out myself. And I knew this and had always worked this way before, but the Internet corrupts, even after advising this to myself I periodically forget and again start resorting to a cheap way to solve problems...

[DiTBho]

When we worked earlier, we did not imagine that there was help and information out there somewhere.
Therefore, we had to do everything mostly ourselves. The Internet is tempting and therefore you waste time trying to find some useful information.

Precisely my opinion 

[SiliconWizard]

True that the search for information online may waste a lot of time, but I think a secondary problem is that people's expectations have also changed due to this seemingly infinite resource: instead of just looking for some precise information to help them move forward, many people are now expecting others (and now "maybe", increasingly "AI", which is yet another potential issue of its own) to solve their problems entirely, instead of just getting some specific information and use it to solve the problem themselves. It's a very significant shift, and we're only beginning to see the consequences.

[magic]

That's the pessimistic interpretation: people got lazier than they used to be, or maybe kids these days are lazier than we used to be.

But I think there is a more optimistic possibility: enough questions have been answered already, enough tutorials written and videos made, that those with any smarts and searching skill simply don't ask that many questions anymore, unless about something new or obscure. This increases the relative fraction of questions asked by the reminder of the population

[Nominal Animal]

But I think there is a more optimistic possibility: enough questions have been answered already, enough tutorials written and videos made, that those with any smarts and searching skill simply don't ask that many questions anymore

That isn't good, either, because nothing truly excellent has been discovered or invented in absolute isolation.  A single viewpoint just doesn't cut it, and hasn't cut it for a hundred years or more in any scientific field.  The flood of low-quality questions means it is harder and harder to have the kinds of discussions and ask the kind of questions that lead to truly better implementations and innovations, at least in public.  And that shrinks the circles that interact that way.

In programming, anything dealing with the reasons for choosing a particular solution mechanism –– algorithm or approach –– is a high-level high-usefulness question, well outside the general low-quality babble of "which library should I use" or "which button should I press".  Of course, anything giving new ideas, including how others set up their workflow and what kinds of tools they use, is useful, but of more personal than technical nature.

What I see happening is the loss of collaboration based on critical, honest, direct arguments and different viewpoints.  People actively avoid discussing things they are not comfortable with, or that they feel is outside the immediate problem at hand.  When told the solution to their current problem is to solve the original underlying problem in a different way, they get annoyed because that is undermining already done work and not useful to them.  Short-sightedness is king, as is chatting and immediate question-response chains, not thoughtful responses and critical discussion.  Many dislike email, because it is asynchronous, and leaves a record of the communication; that is not immediate nor social enough today.

:'(

[peter-h]

is their forum is right to ask any questions?  I mean did you ever get any help at all? Is there anything useful there or is their forum completely pointless, should I ignore it?

There are 2 or 3 people on the ST forum who know stuff, and the best of them is also here as "wek".

How did you determine this? Did you ask questions after they didn’t answer you, or did you determine this by observation?

It is obvious by observation, and by asking questions. What doesn't help is that LWIP etc are old projects and the people who were working on them have moved on. Also I suspect that anybody building "real" products with these libs is re-using code as much as they can. I will certainly re-use 99% of my code on any new product. Even if I don't use ETH I will leave the LWIP stuff in there and just not start that RTOS task.

Did you manage to identify the fishing spots? Sites and people who are really useful in developing the things listed above.

EEVBLOG is ok, the ST forum if you are desperate, and there are others e.g. stackexchange and this is a fairly rare example where someone took the trouble to post his solution
https://electronics.stackexchange.com/questions/350088/stm32f407-lan8720a-lwip-freertos-no-received-ethernet-frames

The Internet is tempting and therefore you waste time trying to find some useful information. All the time is spent searching. But the catch is that there is 95-97% garbage. I noticed that I greatly reduced the loss of time when instead of Googling solutions to problems I began to sort them out myself. And I knew this and had always worked this way before, but the Internet corrupts, even after advising this to myself I periodically forget and again start resorting to a cheap way to solve problems...

Maybe. I would not say 95-97% is garbage IF you use google in a smart way.

I could not have developed the product I worked on just by reading data sheets and such.

that those with any smarts and searching skill simply don't ask that many questions anymore, unless about something new or obscure

Yes I think that's right. The old Usenet still contains a load of good stuff. But I see that a huge factor is the commercial one. Companies are developing social media policies which seek to prevent their employees posting anything useful. So leeching dominates. People here have complained about me posting questions but I have also been posting answers (when I got something working) which is not that common.