Decred Journal – September 2023

Circuit by @Exitus

Image: Circuit by @Exitus

Highlights of September:

Contents:

Network Upgrade and Service Outage

On August 29, consensus changes Change PoW to BLAKE3 and ASERT and Change PoW/PoS Subsidy Split To 1/89 activated on Decred mainnet at block 794,368. The initial difficulty value was reset to target 1.45 TH/s with the expectation that GPU hashrate would start mining on the fork day. This assumption had basis in history, as at the launch of Decred in February 2016, GPU mining started without any announcement. Also, the upcoming switch to BLAKE3 was publicly known for around 5 months since April 2023. GPU mining did not start as expected, however, and the network was left with a difficulty that was extremely high for CPUs. As a result, the chain stopped producing blocks for a couple of days.

Any services depending on the chain getting new blocks got suspended, for example:

After 24 hours of no blocks, miners faced an issue where dcrd would refuse to mine on top of old blocks. The workaround for the issue was to set back the system clock and regenerate certificates if necessary.

Block 794,369 was eventually mined on September 1st and the chain got moving again. Some dcrd nodes had to be restarted due to hitting a path where they weren’t trying to fetch new blocks. After that last issue was resolved, the new ASERT algorithm adjusted the difficulty and restored normal block production speed within hours.

This 2.9 day outage has set a new record, beating the previous outage in April 2017 when it took 9.6 hours to mine block 127,072.

Decred GPU Mining Software

GPU mining kicked in one week after the hardfork and quickly displaced CPU mining. Two software options for GPU mining Decred were publicly released in September:

For any help, join the #pow-mining chat on Matrix or Discord.

BzMiner GUI running on HiveOS

Image: BzMiner GUI running on HiveOS

Bison Relay v0.1.9 Release

This release improves posts UX, key exchange maintenance, and fixes bugs reported in v0.1.8.

Notable changes in both GUI and text apps:

GUI app highlights:

There is a short video demonstrating what’s new in the GUI app.

Text app highlights:

Full release notes and downloads can be found on the release page. It is recommended to verify the files as described in the README.

Improved notifications in Bison Relay v0.1.9

Image: Improved notifications in Bison Relay v0.1.9

DCRDEX v0.6.3 Release

This release contains several important fixes and improvements. All users are advised to update. Highlights for the standalone app:

Read the full release notes here.

Note: As of writing, dexc binaries are also available as part of the core v1.8.1 release. These binaries offer a number of advantages: they are signed with Decred Release key, support more operating system and CPU architectures, and are built with a newer and more optimized Go toolchain.

Development

The work reported below has the “merged to master” status unless noted otherwise. It means that the work is completed, reviewed, and integrated into the source code that advanced users can build and run, but is not yet available in release binaries for regular users.

dcrd

dcrd is a full node implementation that powers Decred’s peer-to-peer network around the world.

Changes backported to the v1.8.1 release:

Changes merged in master towards future releases:

dcrwallet

dcrwallet is a wallet server used by command-line and graphical wallet apps.

Changes backported to the v1.8.1 release:

Changes merged in master towards future releases:

Decrediton

Decrediton is a full-featured desktop wallet app with integrated voting, StakeShuffle mixing, Lightning Network, DEX trading, and more. It runs with or without a full blockchain (SPV mode).

All changes below have been included in the v1.8.1 release (published in October).

Progress towards Ledger integration:

Other:

vspd

vspd is server software used by Voting Service Providers. A VSP votes on behalf of its users 24/7 and cannot steal funds.

September’s vspd work has made incremental quality of life improvements for users, admins and developers.

All changes listed below have been included in v1.3.0 and v1.3.1 releases, which came out in September.

User-facing changes:

Changes for VSP admins:

Internal and developer changes:

Missed tickets in vspd admin panel

Image: Missed tickets in vspd admin panel

gominer

gominer is a Decred proof of work miner for solo and pool mining with OpenCL and CUDA devices.

The main goal of September’s gominer development was to quickly react to the first GPU miners who did not share their code publicly and thus began to dominate the hashrate. The updates to gominer allow anyone to mine with their GPUs. Besides adding BLAKE3 GPU mining support, an effort has been made to automate or document the setup process to address multiple pain points reported by the miners.

BLAKE3 GPU mining:

GPU programming is not something the Decred developers do every day, but this OpenCL kernel has shown to be quite fast. For example, it achieves 10-12 GH/s on an RTX 3070, which is close to 12 GH/s reported by WhatToMine for IronFish, another mining algorithm based on BLAKE3. A better comparison would need to consider the details (size of the hashed data, number of hashing rounds, etc.), but the hashrates reported so far suggest that gominer is good enough to make GPU mining fairer and accessible to more people.

README updates:

Other changes:

dcrpool

dcrpool is server software for running a Decred mining pool.

Similar to gominer, dcrpool development has been reactivated after a period of low activity since summer 2021 when release v1.2.0 came out. The developers upgraded the codebase to the latest Go features, fixed multiple concurrency and shutdown edge cases, and paid off some technical debt, all in preparation for adding pooled GPU mining support.

Upgrades and changes:

Fixes:

Refactoring:

Documentation:

Test code:

Lightning Network

dcrlnd is Decred’s Lightning Network node software. LN enables instant low-cost transactions.

DCRDEX

DCRDEX is a non-custodial, privacy-respecting exchange for trustless trading, powered by atomic swaps.

September changes included in the v0.6.3 release:

Work listed below has been merged to master towards future releases.

Decred:

Firo:

Ethereum:

Polygon (MATIC):

Client GUI:

Client backend:

Client, trading bots:

Documentation:

Server:

Other:

Decred staking UI in DCRDEX

Image: Decred staking UI in DCRDEX

Polygon integration in DCRDEX

Image: Polygon integration in DCRDEX

Cryptopower

Cryptopower is a multi-coin desktop GUI wallet for DCR, BTC, and LTC. It runs in a privacy-preserving light SPV mode without needing full blockchains, supports Decred staking, mixing, voting, and other unique features.

All changes listed below have been merged to master towards future releases.

New Overview page:

DEX integration:

User-facing changes:

User-facing fixes:

Decred:

Litecoin:

Android:

Internal and developer changes:

Bug reports and feedback are welcome via GitHub Issues or the #cryptopower Matrix chat!

Proposal voting in Cryptopower

Image: Proposal voting in Cryptopower

Recent trades in Cryptopower use centralized instant exchange services

Image: Recent trades in Cryptopower use centralized instant exchange services

DCRDEX onboarding pages in Cryptopower

Image: DCRDEX onboarding pages in Cryptopower

Design for the upcoming Overview page in Cryptopower (actual implementation may differ)

Image: Design for the upcoming Overview page in Cryptopower (actual implementation may differ)

Documentation

dcrdocs is the source code for Decred user documentation.

decred.org

dcrweb is the source code for the decred.org website.

Bison Relay

Bison Relay is a new social media platform with strong protections against censorship, surveillance, and advertising, powered by Decred Lightning Network.

September changes in both GUI and text apps included in v0.1.9 release:

GUI app changes included in v0.1.9 release:

brclient changes included in v0.1.9 release:

In other news, @vctt is making a poker game proof of concept built on top of brclient. In the second iteration, the poker logic has been converted to a bot, which is easier to integrate in existing clients. The code can be found here.

First iteration of mobile design in Bison Relay v0.1.9

Image: First iteration of mobile design in Bison Relay v0.1.9

Pages in brclient v0.1.9

Image: Pages in brclient v0.1.9

Poker bot PoC game in Bison Relay GUI

Image: Poker bot PoC game in Bison Relay GUI

People

Welcome the new first-time contributors:

Community stats as of Oct 2 (compared to Sep 2):

Governance

In September the new treasury received 8,374 DCR worth $111K at September’s average rate of $13.23.

No treasury spend transactions were mined in September due to a slip in the admin approval timeframe, but a treasury spend tx spending 8,411 DCR was mined on Oct 1 after being approved with 99.9% Yes votes and 64% turnout, the highest so far. This TSpend covers July and the majority of August’s billing, in an effort to close the expanding gap between work being completed and being paid. It has 53 outputs ranging in size from 7 DCR to 1,719 DCR. The billing rate for July was $15.40/DCR and for August $13.89/DCR, at the average of $14.65/DCR the payment would cover about $123K of costs for both months combined.

As of Oct 5, the combined balance of legacy and new treasury is 866,862 DCR (11.4 million USD at $13.20).

Treasury monthly balance in USD; note that it heavily depends on the exchange rate

Image: Treasury monthly balance in USD; note that it heavily depends on the exchange rate

Voting was completed for two proposals in September, but the voting period was disrupted by the activation of DCP-0011 and a period of around 3 days when the blockchain was stalled waiting for a new block to be found with the new hashing function, followed by a period of rapid block production as GPU miners began operating on the network. The delay of greater than 24 hours without a block caused certain safeguards to activate in the wallet software, which became convinced that the client must be offline and would have prevented normal voting behavior.

There were two proposals published this month, although one of them is a repeat of the Cake Wallet integration proposal.

More events and details surrounding the above proposals have been covered in Politeia Digest issue 64.

Mining

Decred mining was rebooted on August 29th when ASICs mined their last block and the chain switched to BLAKE3 proof of work.

Since the switch would remove all ASIC hashrate from the network, the mining difficulty had to be reset to a low enough value to allow commodity hardware to mine new BLAKE3 blocks. For comparison, ASICs mined their last block at difficulty of 3.5 billion while the next block had the difficulty value of 101 thousand, or 35,068 times lower.

The initial difficulty value was manually set to target about 1.45 TH/s hashrate in anticipation of GPU hashrate, but it didn’t immediately join the network and only CPU miners were active upon BLAKE3 activation. A single CPU achieves between ~5-75 megahashes per second (from old laptops to high-end CPUs). Assuming an “average CPU” as 20 MH/s, getting 1.45 terahashes would require some 72,500 CPUs. The first BLAKE3 block (794,368) was extremely lucky to be found in only 50 minutes, assuming it was found with CPUs, but afterwards the chain stalled.

In a positive development, this incident activated many old and new community members to come together and push the chain forward. The #pow-mining chat became very busy as people converted any available hardware to Decred miners and endured the generated excess heat. Devices ranged from tiny Raspberry Pi’s (1 MH/s), to old laptops (2-5 MH/s), to new laptops and desktops (10-20 MH/s), to monstrous 120 Watt+ processors with 16 and 32 cores (60-75 MH/s).

I have to say, I love how this is bringing everyone together. I’ve missed this collaborative spirit. [@jazzah]

By the end of August 31st around 25 individuals reported a total of 550 MH/s, but that (and any unreported hashrate) was still not enough. On early September 1st, @grumlin and @davecgh turned to cloud hashrate and within ~3 hours block 794,369 was found. It was 324 KB in size and included 240 transactions accumulated over the days. Since no blocks had been mined for so long, the ASERT algorithm only needed one block to reduce the mining difficulty from ~97,000 to ~1,800 and from there CPUs had no problem finding new blocks and the target block time was restored in just a few hours.

September 1-5th was the short period when CPU mining was feasible as the difficulty settled around 500 and the corresponding network hashrate stayed between 6-10 GH/s, the equivalent of ~300 CPUs or just 1 mid-range GPU. Mining address distribution became significantly more decentralized compared to the ASIC era.

Mining difficulty during the CPU days

Image: Mining difficulty during the CPU days

Block distribution by mining address as of September 6th

Image: Block distribution by mining address as of September 6th

However, the CPU mining days did not last long. Developers anticipated that GPU miners would show up soon given that even a single GPU mines as fast as roughly 500 CPUs, and GPU farms are quite common. First signs of GPU mining got spotted around September 2nd but it wasn’t definitive because they seemed to be intentionally keeping it in the range of plausible deniability. With a more reliable confirmation on the 5th, the developers rushed to add GPU mining support to gominer.

On September 6th, a significant increase in block production rate was a clear indicator of GPU mining happening on the network. The hashrate increased to ~32 GH/s by 15:00, from ~9 GH/s just one day before. In about one day, rewards for ~150 blocks were collected by a single address (normally ~288 blocks are mined per day). First GPU mining patch was merged in gominer around 14:30 and by 15:30 first community members got their GPUs up and running. After GPU mining was made available to everyone the hashrate went straight up, reaching ~52 GH/s by 18:30 and ~700 GH/s by 20:00. New blocks were being mined as fast as 100 per hour (the normal rate is 12 blocks/hour).

By late September 7th the hashrate reached the level expected on the fork day, around 1,500 GH/s or 150 GPUs (assuming a 10 GH/s GPU like RTX 3070). This was a good test of the new difficulty algorithm, which was designed in anticipation of such a scenario. It worked well, ASERT managed to return block times to normal by September 10th, despite the hashrate increasing to ~3,000 GH/s in the same period. With the old algorithm it would take 2 weeks in the best case.

Soon after gominer was in good shape enough that the developers switched to refreshing the codebase and adding BLAKE3 support to dcrpool. BzMiner developer @iedoc released BzMiner v17.0.0 with Decred mining support, capable of squeezing out a few extra gigahashes from the GPU compared to gominer.

BzMiner v17 managing 8 GPUs mining Decred

Image: BzMiner v17 managing 8 GPUs mining Decred

The rest of September was not as dramatic, but nevertheless the hashrate continued to grow and reached ~8,000 GH/s by the end of the month.

Mining difficulty increased x200 in just 2 days, and did another x5 through the rest of September

Image: Mining difficulty increased x200 in just 2 days, and did another x5 through the rest of September

Average block time was wild for a few days, but stabilized quickly

Image: Average block time was wild for a few days, but stabilized quickly

This chart shows just how centralized the mining was before the fork, only ~5 unique mining addresses

Image: This chart shows just how centralized the mining was before the fork, only ~5 unique mining addresses

Another view at mining (de)centralization, each color represents a unique mining address

Image: Another view at mining (de)centralization, each color represents a unique mining address

Network

Decred’s network upgrade and the corresponding technical issues have created volatility in several metrics as shown on the charts below, but most metrics have normalized by the end of September or early October.

Hashrate: September’s hashrate opened at ~0 TH/s and closed at ~8.8 TH/s, bottoming at 0 TH/s and peaking at 8.8 TH/s throughout the month.

No information about DCR mining pools operating in September is available.

Staking: Ticket price varied between 141-278 DCR.

The locked amount was 8.52-9.82 million DCR, meaning that 54.9-63.5% of the circulating supply participated in proof of stake.

Low ticket buying activity caused a drop in locked DCR and stake participation

Image: Low ticket buying activity caused a drop in locked DCR and stake participation

Ticket price went down to stimulate ticket buying again

Image: Ticket price went down to stimulate ticket buying again

VSP: The 15 listed VSPs collectively managed ~7,550 (+650) live tickets, which was 18.0% of the ticket pool (+3.5%) as of Oct 1.

The biggest gainers of September are bass.cf (+495 tickets or +54%), stakeminer.com (+357 tickets or +107%), and stakey.net (+336 tickets or +128%).

Distribution of tickets managed by VSPs

Image: Distribution of tickets managed by VSPs

Nodes: Decred Mapper observed between 147 and 159 dcrd nodes throughout the month. Versions of 153 nodes seen on Oct 1: v1.8.0 - 89%, v1.7.x - 4%, v1.9.0 dev builds - 2.6%, v1.8.0 dev builds - 0.7%, other - 4%.

The majority of nodes are running dcrd v1.8.0. The red area before Jan 2023 indicates incomplete data we had at that time.

Image: The majority of nodes are running dcrd v1.8.0. The red area before Jan 2023 indicates incomplete data we had at that time.

The share of mixed coins varied between 60.7-62.6%. Daily mix volume varied between 40-906K DCR and made a new all-time high.

Daily mixed DCR dropped during the chain stall, but came back with a vengeance marking a new ATH

Image: Daily mixed DCR dropped during the chain stall, but came back with a vengeance marking a new ATH

Total mixed and unspent DCR dipped, but quickly recovered

Image: Total mixed and unspent DCR dipped, but quickly recovered

Decred’s Lightning Network explorer saw 211 nodes (-2), 429 channels (-2) with a total capacity of 190 DCR (+4), as of Oct 2. These stats are different for each node. For example, @karamble’s node reported 213 nodes (-4), 446 channels (-7) and 191 DCR (-1) capacity on same day Oct 2.

Decred's Lightning Network capacity remains stable

Image: Decred’s Lightning Network capacity remains stable

Uptick in missed tickets in September was likely due to technical issues after the hard fork

Image: Uptick in missed tickets in September was likely due to technical issues after the hard fork

Monthly DCR emission was slightly higher than usual due to a period of fast blocks

Image: Monthly DCR emission was slightly higher than usual due to a period of fast blocks

Thanks to @bochinchero for providing and improving these charts. About 40 other charts not used in this Decred Journal issue are available in the dcrsnapshots repository; everyone is welcome to share them on social media.

Ecosystem

Voting Service Providers:

Exchanges:

Other news:

Join our #ecosystem chat to get more news about Decred services.

Warning: the authors of the Decred Journal have no idea about the trustworthiness of any of the services above. Please do your own research before trusting your personal information or assets to any entity.

Outreach

Decred Vanguard

Decred Vanguard is a community-based marketing effort with the goal of increasing Decred’s outreach and social media presence.

Updates: We ran a successful meme competition and gave out $200 in DCR for prizes. The other recipients of the monthly Decred Vanguard reward of $100 in DCR were @tothemoon for his Decred Discord server contributions, and @PubPete for his high activity.

Are you a meme creator, artist, strategist, or just someone passionate about the Decred Project? We’re expanding our community-driven marketing program, and we want YOU!

What’s in it for you?

Interested? Contact @Exitus on Twitter/Matrix/Discord.

Cypherpunk Times

Engagement stats for September:

Other

Events

Attended:

@arij has signed a partnership agreement with National Institute of Innovation and Advanced Technology (NIIAT) based in Casablanca, Morocco. This 1-year agreement will facilitate cooperative efforts like workshops, study sessions, and various interactive events. One of the key topics of the event was the introduction of Islah.city, a platform where citizens of Casablanca can identify and collaboratively resolve urban challenges. The potential to integrate blockchain technology and timestamping has been discussed. The event garnered attention on the national TV news channel Al Aoula, the video clip is available on Twitter and YouTube. See more in the full report.

Media

Selected articles:

Decred:

General crypto:

Guides:

Videos:

Audio versions of some videos are published on the Cypherpunk Times Spotify podcast.

Check out the new Shorts section on @DecredTV!

Translations:

Non-English content:

Discussions:

A reality check from @Tivra:

This is not “crypto” this is some trashy reality show some of you follow on X. Crypto happens on GitHub and on chain not on video. You control where your attention goes, it’s ultimately your responsibility [tweet]

Other:

Art and fun:

Merch design by @lewildbeast for the hero miner who solved block 794,369 (DCR reward needs to be updated

Image: Merch design by @lewildbeast for the hero miner who solved block 794,369 (DCR reward needs to be updated)

Interpretation of the UP ONYL paradigm by @Void

Image: Interpretation of the UP ONYL paradigm by @Void

Markets

In September DCR was trading between USDT 12.33-15.25 and BTC 0.00048-0.00057. The average daily rate was $13.23.

Historical DCR/USD accumulation zones by @saender

Image: Historical DCR/USD accumulation zones by @saender

Recent DCR/BTC, data from Coin Metrics

Image: Recent DCR/BTC, data from Coin Metrics

Recent DCR/USD, data from Coin Metrics

Image: Recent DCR/USD, data from Coin Metrics

DCRDEX monthly volume in USD

Image: DCRDEX monthly volume in USD

Relevant External

The debate around Drivechains for Bitcoin and BIP-300 was reignited after Luke Dashjr wrote an implementation. There are discussions about whether this kind of sidechain which would allow for flexible creation and use of other assets would be a positive for Bitcoin, whether there are technical downsides or if allowing people to create scam-like assets on Bitcoin is the worst aspect. There are also discussions about how to activate the soft fork, with already some miners signaling support. A new player entered the mix in September with claims of being able to operate a sidechain with a two-way peg on Bitcoin without any soft fork changes, but they as yet had no explanation which people could make sense of.

Paxos accidentally paid 19.8 BTC for a Bitcoin transaction, but F2Pool apparently returned the BTC.

The Brave browser announced a partnership with Electric Coin Company to add Zcash support to its wallet, and to develop a “private and decentralized messaging feature” with the support of Filecoin and the InterPlanetary File System (IPFS).

Nouns DAO, which auctions one NFT a day to fund a treasury and had amassed a sum of around $50 million in this fashion, underwent a fork which saw over half of this value exiting through a newly introduced quit mechanism. The Nouns DAO is controlled by voting of the holders of Noun NFTs, and this community has long been split between a camp which embodied the original spirit of using the fund to promote the Nouns meme, and a camp which was disturbed by frivolous spending amounts as high as $90,000 on naming a frog species, and which saw themselves as entitled to some share of the fund proportional to the number of Nouns NFTs and voting power they had. As the product of a long-running discussion, v3 of the Nouns DAO’s contracts incorporated a mechanism that allowed 20% of the NFTs to trigger a week-long forking discussion and voting period. The majority of Nouns holders chose to exit and their share of the funds is now moved to a new DAO that has a ragequit mechanism, in addition to the familiar mechanisms of Nouns DAO, which will allow them to further exit and claim their individual share of that balance, if they so desire.

Terra Classic is to cease minting and reminting USTC in hopes of returning it to parity with the US dollar, after a vote with 59% in favour of the proposal. Since the the decline of the price of LUNC, there are concerns about spam and a proposal to increase the minimum deposit amount to 5 million LUNC was approved with 93% support.

Coinbase, in its role as an Ethereum validator, received a 570 ETH MEV payment from the Curve hackers in July to prioritize the transactions which allowed them to profit from the attack. That’s about $1 million profit for Coinbase at the moment, and they seem reluctant to return it to the affected party, Alchemix, citing a legal opinion that they are under no obligation to do so.

The SEC has reached a settlement in which the makers of the Stoner Cats NFT collection and associated web series were fined $1 million and had to kill all the Cats in its possession. The case attracted attention because it featured celebrities Ashton Kutcher and Mila Kunis promoting the project and participating in the creation of the comedy series, which the $800 NFT would allow holders to access. The token sale generated 3,650 ETH ($8.2 million at the time) for the company, and they received a further $20 million in royalties in transactions where holders sold their NFTs on the secondary market. After the news broke, the price for the NFTs rose substantially, possibly because people expect the holders to be able to access part of the settlement, but $1 million seems quite small in the scheme of things. The people who lost out in this situation are those who bought the NFTs at their all-time high prices of around 0.3 ETH, they are unlikely to benefit from any financial penalties, and the recent rise in price to 0.05 ETH is little consolation.

Claimants in the ongoing bankruptcy processes of FTX, BlockFi and Genesis have had their details leaked by Kroll, a provider of bankruptcy services. A hacker performed a SIM swap attack on a Kroll employee and was able to access files with details of all the claimants. This further raises concerns that participants in the claims process for bankrupt crypto companies are being targeted by scammers who attempt to gain access to whatever crypto the claimants might hold off the exchange. The leaked details allow the scammers to produce more convincing materials which reference compromised information, one reported example is offering marks a way to speed the claims process or get a better outcome.

The Commodity Futures Trading Commission (CFTC) has announced settlements with three DeFi operations (Opyn, ZeroEx, and Deridex), who it charged with offering illegal derivatives trading. The companies agreed to shutting down these services and paying fines of $100,000 to $250,000 after cooperating with the CFTC’s investigations.

The CFTC is hampering the development of prediction markets by objecting to any markets relating to political outcomes. These are explicitly banned in over 10 states but the CFTC has a broader objection in that it does not want to become an “election cop” adjudicating disputes about how the contracts have been operated.

Senator Elizabeth Warren has gathered support of 9 other senators for her Digital Asset Anti-Money Laundering Act, which would crack down on non-custodial wallets, extend Bank Secrecy Act responsibilities, and other measures to tackle illicit use of crypto.

Chase bank in the UK, owned by JP Morgan, has announced a ban on crypto-related transactions for its customers, who from mid-October will no longer be able to make debit card or bank transfers to crypto companies like exchanges. The reason given was that this is a response to a growing number of scams targeting UK users with a crypto component.

The UK’s Online Safety Bill has been passed by Parliament and is on track to become law. The legislation will require the UK’s regulator Ofcom to produce a code of conduct detailing how providers of online messaging platforms must pro-actively police illegal content on those platforms. One particular aspect which has been criticised would undermine end to end encryption by requiring service providers to sift through encrypted private content for evidence of illegal activity. Both Signal and WhatsApp have indicated that they would leave the UK market rather than comply with the new rules which would require them to undermine their encryption.

That’s all for September. Suggest news for the next issue in our #journal chat room.

About

This is issue 63 of Decred Journal. Index of all issues, mirrors, and translations is available here.

Most information from third parties is relayed directly from the source after a minimal sanity check. The authors of the Decred Journal cannot verify all claims. Please beware of scams and do your own research.

Credits (alphabetical order):