Not everything is the same. You are comparing apple to oranges.
No, I'm not. You stated that you have "used WinRar for many several years and it never disappoint me, not even once" as some sort of proof that you can trust it 100%... so I pointed out that this is exactly the type of comment people have made about various services/software over the years... which then turn out to be a scam or buggy and financial and/or data loss occurs. Just because something hasn't "failed" yet, doesn't mean it won't.
"Saying no procedure is 100% safe sounds like speculation to me" means I believe in a way that can give 100% security. And my method is one example.
I've asked several times now, how you can be 100% certain that WinRAR has no backdoors (or unpublicised bugs/flaws) that could allow an attacker to sidestep your encryption and access your private keys. The simple and very objective answer is that you can't be 100% certain of this, as it is closed source. You are putting your faith and trust in the WinRAR devs, just like you claim hardware wallet users are having to put their faith and trust in the hardware wallet manufacturers.
And just like you, we can't be certain that hardware wallets are 100% secure either... because there is no 100% secure system. There will always be attack vectors. Your continued assertion that your method is one example of 100% security without any real supporting proof seems to be showing a certain lack of objectivity.
I came upon that point too when writing my article. I would say you are an intelligent man if you can use my method to do similar dummy wallet. There is nothing to stop you from personal customization when you are 100% in control. Not being able to see that with my method means you are not.
"I wish when I debate with someone, that person would be objective"
and play the ball... not the man.
My article said it was only for cold storage, with an offline and formatted computer. But when it comes to spending, it isn't less convenient if you appreciate 100% control.
Example:
You have 100 btc.
You generate 20 addresses (with keys of course).
You transfer 5 btc to each address.
When you spend, you take out only 1 of the 20 addresses for use.
You spend only 5 btc and everyone knows you have at least 5 btc only (instead of 100 btc, because your addresses aren't a bunch of change addresses that reshuffle your 100 btc with every transaction).
If you get the $5 wrench attack, you can pretend you have only another 5 btc address (just as when you pretend you have only the dummy wallet with your hardware wallet).
If you want to cover the trace of your 5 btc, you use washers.
How can you do all the above with hardware wallet, satisfactorily?
Is that a serious question? Because it should be fairly obvious to "an intelligent man" how one could satisfactorily achieve that with a hardware wallet.
and I'm not quite sure what you mean by:
everyone knows you have at least 5 btc only (instead of 100 btc, because your addresses aren't a bunch of change addresses that reshuffle your 100 btc with every transaction)
Why would your total of 100 BTC be shuffled with every transaction?
Why could you not also have 20 addresses with 5 btc each when using a hardware wallet?
Assuming you're spending less than 5 btc as per your example, the wallet would only need one UTXO so "you spend only 5 btc and everyone knows you have at least 5 btc only (instead of 100 btc, because your wallet only needs to use one input from one address)".
For added defense against the wrench attack, you simply create multiple different wallets with one address each... hand over the passphrase to the "exposed" 5 BTC and your dummy wallet.
The bonus is that with a hardware wallet there is exactly ZERO evidence that these multiple wallets even exist... whereas, with your system of 20 different encrypted private keys and an attacker has hacked your email or found your USB/CD with the encrypted files... they can actually see all the different encrypted items in your inbox or on your CD etc. So, I'm not quite sure how you deny the existence of other keys/addresses?
You are not being objective. Besides, using those seeds allows easier brute force hacking because they are all dictionary words with all small caps (do you realize this?).
Do you realise that a 24 word seed... is effectively like having a 24 character password from an "alphabet" that has 2048 possible characters in it... whereas your proposed password of 20+ characters (we'll even be generous and say 24 character to compare apples to apples) using alphanumerics + symbols gives you a total of 26 upper + 26 lower + 10 numbers + say ~30 symbols... for a total "alphabet" size of ~92 total characters to choose from.
204824 combinations vs.
~9224 combinations... Tell me again which one is going to be easier to brute force?