This isn't going to work the way you think it will.
This is actually good thinking.
So you were on a right track earlier in the thread that you'd linked to, and which made you angry.
This guy is on the right track to the solution.

Basically, any one-dimensional solution isn't going to work in the social practice; it doesn't matter whether you call it merit, karma, reputation or something similar. It will always be vulnerable to the abuse, whether you call it "merit cycling", "karma farming", "circle jerking", "mutual admiration society" or something similar.

What could work is adding another dimension to the "merit value". It is hard for me to come up with a good name in English, but for sake of this thread lets use the word "funniness". The funniness can be both intentional and unintentional. It really won't matter in the grand scheme of things. Even now some people do give merit while posting a disclaimer that it made them laugh.

I've seen such systems working quite well on other forums that do support plugin architectures. I know this one theoretically does, but it wasn't kept up to date with the official code base, so they won't simplify implementation. I will not post links to those forums because of they aren't in English and to avoid doxing myself.

For the people like gmaxwell with mathematical mind the argument is simple: avoid attempting to order people linearly like the Real numbers. For each two real X & Y numbers exactly one is true X < Y , X = Y or X > Y. At least switch to the Complex numbers, where there are only two choices: either X = Y or it doesn't.

For the people preferring more humanistic approach I recommend familiarizing themselves with the philosophical thinking of Herbert Marcuse that he published in his https://en.wikipedia.org/wiki/One-Dimensional_Man . This book is from 1964 so it will not contain anything related to the modern social media. But it does contain lots of very useful thought about the things that are discussed in this thread. Even where Marcuse is wrong, the passage of time and social experience gained until now will make those places obvious.

Awesome resource. Thanks for posting the images with links on your website!

Edit: Oops, looks like my send-able merit had decayed, I'm sending you whatever I have left over.

👊

Apologize, please delete if you can.

Thank you very much for your very thoughtful and well reasoned reply. You don't have to explain yourself to me, I'm fully supporting the general gist of your business model. I've made a comment about your question related to financial side of the deal.

Way too many people in the cryptosphere are getting "some crazy guy walking in off the streets" treatment because they focus on the wrong things: how they dress to the meeting, what car they roll in with, etc. and most importantly: they seem to be proud of not knowing anything about https://en.wikipedia.org/wiki/Uniform_Commercial_Code . So I really doubt you when you say they asked about "cold hard cash" (do they asked for suitcases or would plain bags be OK?) This isn't some contraband deal, properly drafted letter of credit would be actually better than a suitcase of money, because it is not only a guarantee of payment but also very good promise of the future business.

I've been in this business, met some people both successful (like Xilinx) and failures (like Plus Logic, but search for "Plus Logic Inc", otherwise you'll get overwhelmed with irrelevant stuff). I don't really know what happened at your meeting, but I went through very detailed postmortems of the meetings (and lawsuits) I was then involved. So I have some actual knowledge what led to success and what caused the failures.

It may look like nitpicking, but it starts with really small things:

1) using correct physical units (e.g. kWh and not Kw per hour or similar abominations) and other technological details;
2) using correct semantics (e.g. subject-object distinction) even if the grammar is lacking;
3) not mixing up an agent with a principal in a transaction;

and ends up at really important stuff like:

4) know what you don't know and don't be afraid to admit to it.

Oftentimes the experienced businesspeople will intentionally say something disagreeable or outright stupid just to gauge the reaction of the interlocutor and to see firsthand how they will get corrected or even if they get corrected at all. Many salespeople are so drilled (or self-drilled) into ABC (Always Be Closing) that they will try to agree just about to anything if they think that brings them nearer to closing. Success with retail sales (e.g. fast talking and clever one-liners) doesn't always translate into a success in the high-level transactions.

I don't know who advised you on the financial and legal side of this deal. But if you hire a good lawyer, he will tell you mostly the same things I type in here for free. But the good commercial lawyer will know the details of the UCC and will not be trying to make general statements like I'm doing here for the benefit of readers of an Internet forum. In particular you may want to have an in-person review of what was said in your meetings and conversations with Xilinx & Paypal while your memory is still fresh. You guys are getting prepaid terms probably because you had somehow inadvertently behaved like common scammers despite actually trying to work a legitimate deal.

And that is my "angle" in this thread.

This is just plain bullshit, based on misunderstanding between consumer finance (credit card) and business finance (letter of credit). My bet is that you didn't even click through to skim the Wikipedia page. Such strawman wouldn't even be worth further comment, but for the benefit of other readers I will comment:

The semiconductor manufacturing industry is capital intensive, therefore they are heavy users of credit (and other classical, non-crypto, financial instruments), but they use it very intelligently. This has nothing to do with consumer finance chargebacks or anything related to retail finance.

I had personal experience working with Actel, Xilinx, and some less well known names from Taiwan, Malaysia & Korea. Everything was financed with some form of credit, although as an engineer I wasn't really involved in the financial details beside general interest.

I think my former colleague summed it the best: "Rectoscopy is a less invasive procedure that what you will go through with drafting and acceptance of a letter of credit."

---

Thanks for the comment. I really appreciate it now. Back then a coworker showed me the "show me the money" scene from the "Jerry Maquire" movie and I couldn't find it either funny or educational. But now I understand how, while an artistic exaggeration, it is a perfect representation of certain type of business dealings. Some find it offensive, but it is just short and to the point.

Edit: The point being: people don't like when you burn the candle from both ends.

No well-designed FPGA board became paperweight ever. There's a quite thriving secondary market in them, they have so many different uses.

The only FPGA boards that had fallen badly in price either:

1) didn't have any memory attached and no way to attach memory
2) didn't have sensible communication channels to get the data in and out

Historically on this forum most FPGA talk was about ngzhang's Icarus & Lancelot and Enterpoint's Cairnsmore. Those are the quite perfect examples of not being well-designed. In particular Cairnsmore had problems with internal clock distribution between the 5 FPGAs on the board.

---

I have no angle here (or "I have no alpaca in this race", as they say on this board.)

I'm just observing and it is quite obvious that "I offered a letter of credit" and "Companies do not spin up $20M worth of product on letters of credit" are not even nearly synonymous and you somehow suddenly switched sides between yourself and Xilinx. This is a very definition of a word "shifty" as it is used in business negotiation. It is quite obvious that you hiding something material and they must have found this while negotiating the terms. If you habitually talk in a business meetings like you post here on this forum then I completely don't wonder why they get suspicious, require prepay for everything and count their fingers after shaking hand with you.

That statement is patently false. Letters of credit were, are, and will be written for various sums, both lower and higher. It is a mainstay of many industries, not just produce importers.

I don't have an inside knowledge of your venture so I can't make any definitive statements.

What I will comment about is one fact: you may think that you can freely mix true and false statements and they will average to "reasonable". No they average to "are you on drugs?" or "what film set did you came from?" or "did you forgot to take your medications?".

This vacillation between heady optimism and dark paranoia in a single person is not only off-putting. It is the reason to raise suspicion among both crypto-currency people and traditional finance people.

It does work just fine on Windows. The syntax is something like keyhunter.py \\.\PhysicalDisk2 . Of course one does have to use "elevated command prompt" a.k.a. "run as administrator".

Well, that simple matter is called https://en.wikipedia.org/wiki/Letter_of_credit . People routinely have been paying that way for importing the produce that haven't even grown yet. I feel weird explaining to the highly educated people what I learned as a kid from the bazaar fruit importer.

I did an edit of your post to get to the narrative.
Basically it has been tried already in the days when Bitcoin could still be mined by the GPU cards. Some guy did that in Germany or in Finland. It was linked on this forum, but either I can't find it or it is gone. He couldn't keep his setup working even a single full year.

The main technical point is that protection from the elements in a non-trivial undertaking. It is not only about keeping it cool, it is also about keeping it warm and dry when not operating or operating with constrained power.

The moisture condensation and icing will void any warranty for the electronics, unless you buy marine-rated equipment.

Additionally, the warranty on panels in the USA will probably be unenforceable. I've been told that those companies reselling and installing the solar equipment are intentionally created to go bankrupt, they may as well offer 100 years of warranty. That is apparently some political stuff more difficult to explain, which I don't fully understand. I didn't see the actual panels discussed in this thread, but I've seen some other sold and installed in the USA and they weren't designed to last.

I'm sorry for raining on your parade.

DC to DC inverters?

Frankly, what you wrote isn't even sounding sane.

Trying to come up with a workable explanation:

1) DC to AC inverters
2) haul power off the field to some shed in a shade
3) rectify AC and regulate DC
4) feed miners short distance in the shed

could probably work if you buy recycled/re-manufactured/B-stock parts and still somehow manage to get politically-motivated subsidies for the cost.

I understand that not mounting on the roof would greatly simplify inspection for fire and electrical safety. But you would still require some inspection if planning on connecting to the grid or use the power in the residence/business abutting the field where all that is installed.

On a second thought I even don't think that the inspection would be easier to pass. Fire safety definitely easier; but the electrical safety would be worse: the field mounted electrical stuff would need to be made child-safe. I heard second-hand stories about that type of safety inspections and various lawsuits related to leaving dangerous equipment unattended.

Edit: Additional posts preserved for further reference.

This doesn't even look like science-fiction right now, it is more of a fantasy genre.

The solar cell manufacturers can't even reliably build-in the temperature sensors or other environmental sensors. Various radio-equipment manufacturers still can't reliably integrate electronics with antennas, even though the radio equipment isn't required to be maximally exposed to the elements.

Quite a few roof-based solar installs don't survive single year without sprouting leaks or other problems. Having recently seen the quality of workmanship on some of the solar power installations I observe a market regression in quality. Lots of them will require major repairs or will be completely torn down in the coming years.

Currently the flat, thin magnetics that would be required for power regulators are very expensive and are used only in high-end high-power CPU chips.

I'm not the one that would try to prohibit daydreaming, but people need to be aware of the distinction between dreams and reality. Otherwise it gets into the realm of either psychiatry or fraud.

Well, I don't know about the vagaries of Intel's marketing. I'm somewhat more interested in how the rumor mills work on this forum or in somewhat broader cryptocurrency news/rumor mills. Nowadays whatever QuintLeo says, I'm immediately thinking that opposite is true. And you have started to repeat a lot of his proclamations.

Thanks for clarification.

Sorry, I don't know off-hand how to do that using modern hardware and software. I've done it in the past on the hardware from 2003/2004.

You should seriously consider just buying a better gateway. Either buy Asus DSL-AC68U to replace your BT jobbie, or:

3) reconfigure your BT jobbie to a bridge and and a plain Ethernet gateway like Asus RT-AC68U.

4) hire somebody locally to consult you.

Do you know this from some source at the foundries or do you just repeat somebody's baseless speculation? Edit: Wasn't that QuintLeo who convinced you that Intel isn't offering foundry services? End edit.

It would seem that Bitcoin mining IC is near perfect test product for a new process: it is repetitive, it tolerates bad yield, it is free of actual trade secrets or intellectual property in its circuitry. It would be beneficial for both foundry and designers to cooperate and profitably sell the wafers fabricated during process bring-up.

GMO being Japanese could have an additional benefit of really being able to sign binding non-disclosure agreements, unlike the mainland Chinese.

I would love to hear from somebody who has current experience. Mine is really dated, but even a very small fab-less company had been trusted with internal fabrication process metrics provided that they made actual technological and legal effort to protect the trade secrets of the fab.

It costs next to nothing, basically only traces, all the logic is synthesized in the FPGA. Gives the board other uses besides mining.

The reason why old ZTEX boards have much better resale value than the old mining-only boards from ngzhang. His Lancelot & Icarus boards were really crippled by not having any way to get signals to the IO pads of the FPGA. They had only 2 (out of 8 available) signals connected between USB USART chip, greatly reducing available bandwidth and protocols. Similar weak connectivity was between the two FPGA chips on the board. It makes the whole board unusable for anything else, not even toy projects could be reasonably built on them.

They look fine for the moment. You'll need to write a script to scrape those numbers every hour and see if they are stable. If they are stable you shouldn't have problem staying continuously online for months at a time or even years.

True, it isn't.

I'm guessing you have this: https://business.bt.com/smart-hub/ . If I'm not mistaken this is a French SagemCom re-branded for British Telecom with absolute minimum of internal memory. They are generally of shitty quality and with lots of functionality intentionally disabled. Instead you get pretty graphics and pointless animations.

See if you can disable that "Smart" shit or just reconfigure it so it never finds "a quiet moment". To me it looks like some sort of temporary workaround for the problem on the provider's side. Other ADSL modems are capable of retraining to the line quality without rebooting, it is observable as just a short hiccup in the flow of the traffic.

You really need to learn how they provision your IP addresses, both IPv4 and IPv6. There are just to many variants of dual-stack provisioning to give you a reliable advice. Linux on a regular PC is certainly much better router than whatever is inside that box. But you may need to invest in a wireless network card or separate wireless access point (without router) to replace the WiFi functionality of your BT box.

Just don't make a mistake of trying to configure Linux to bridge. It is doable, but it wont help you. Linux on the PC has to be a router and a NAT box (sometimes called masquerading,) the BT box needs to be reconfigured for bridging.

Edit: At least it looks like they didn't disable bridging:
https://btbusiness.custhelp.com/app/answers/detail/a_id/35226/~/how-do-i-enable-bridge-mode-on-my-bt-business-hub%3F/c/5098/

Edit2: For my own reference I've found the closest SagemCom model: F@St 5360
http://www.sagemcom.com/broadband/gateways/dsl-gateways/fst-5360/
Those things are known to hold the xDSL sessions really well, half-a-year at a time, but are rather weak NAT boxes. WiFi is fast, but low range.

In the past I've used TP-Link TD-W8970 in similar situation (not miners, but a farm of computers for tests). The manufacturer has marked it end-of-life, I had hard time even finding the information. Other people used ASUS xDSL gateways, e.g. ASUS DSL-AC68U, but I don't know what was changed in the recent firmware.

Your best bet is probably:

1) convert your B.T. device to bridge mode and use a regular PC or Mac to share your connection.

2) buy an end-of-life Cisco IOS router (e.g. Cisco 1801 (ADSL over POTS), Cisco 1802 (ADSL over ISDN), Cisco 1841 + HWIC-1ADSL + HWIC-AP WLAN). They are now super cheap, but they aren't easy to set up, because they use Cisco's crown jewel IOS (Internetworking Operating System). On the flip side, there are plenty of sources to learn and get help. This is rock-solid industrial-grade solution for businesses and it was quite popular before VDSL or optical fiber pushed it out of the market.

Does B.T. charge you rental for your gateway or do you own it outright?

Also, are you sure that it is your router crashing&rebooting and not some line problems (real or artificially created). Where I used to live ISPs would remotely force customer's devices to reboot frequently to force them to upgrade to VDSL or fiber-optics. Do you know the noise margins and bit error rates on your line? I don't know the situation in the UK, but in many other countries ADSL was somewhat mired in political/financial problems on the provider side that made it hard to support for them.

Edit: Also, tell us how BT provides you with an IP address, is it via PPPoE, PPPoA, DHCP, L2TP, MER or something proprietary? This is very important!

Edit2: It was TD-W8970 or TD-W8980, just from the look of it. I recall returning TD-W8968 as it wasn't crashing, but it was just slow under load. Or it was TD-8961ND that I returned, again just from the look of it.

Edit3: I'll just cut&paste this from my gateway for my own reference, although we no longer have any xDSL where I live:
DSL line status
2.1 DSL link status
2.2 DSL synchronization mode
2.3 DSL last synchronization
2.4 DSL synchronization uptime 00 d 00 h 00 m 00 s
Rate and noise margin
2.5 DSL synchronization up Kb/s
2.6 DSL synchronization down Kb/s
2.7 noise margin down dB
Line quality (errors)
2.8 Errored seconds (ES) of downstream (since last synchronisation)
2.9 Severely errored seconds (SES) of downstream (since last synchronisation)

Not really. It is a Visual Basic for FPGAs. Only good enough for quick&dirty or fast prototyping.

Bitstream is both chip- and board-dependent.

Chip-dependent quite obviously because FPGA is a complex two-dimensional (or even 3D) device, not an array of bytes, or 72-bit words, etc. according to some JEDEC standard.

Board-dependent for the same reason: which signals go to which pin of the chip isn't standardized either. Some dependencies could be resolved, but it isn't clear that the boards are compatible enough. It was discussed at the beginning of this thread.

Additionally some boards may not have enough power to continuously supply the required current for full performance. You would then be forced to severely under-clock them to avoid tripping the on-board over-current protection.