Timeline for Would this be considered a secure password hash?
Current License: CC BY-SA 4.0
28 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 3, 2022 at 10:18 | history | edited | Peter Morris | CC BY-SA 4.0 |
Thanks
|
| Feb 3, 2022 at 10:17 | vote | accept | Peter Morris | ||
| Feb 2, 2022 at 7:16 | comment | added | Wernfried Domscheit | Apart from all the theory given in the various answers: At least some people think, developing a hash algorithm is easy. But look how many, or better how few password hash algorithms are actually available, it seems to be a tough job and only few experts around the world are able to develop them. It seems to be much easier to develop a programming language | |
| Feb 2, 2022 at 6:13 | comment | added | Wernfried Domscheit |
Why do you trim the == from your salt?
|
|
| Feb 1, 2022 at 22:20 | comment | added | marcelm | @PeterMorris No problem! I also spotted an interesting defect in your algorithm. All this is becoming way too much for the comments so I wrote an answer. | |
| Feb 1, 2022 at 22:19 | answer | added | marcelm | timeline score: 9 | |
| Feb 1, 2022 at 19:07 | comment | added | AndreKR | "and must contain one of each [Upper, Lower, Digit, Other]" -> I normally let my password manager generate a random password for each website but for websites that force me to use certain special characters I have a single password that basically just consists of special characters. Those are usually the websites that don't matter. | |
| Feb 1, 2022 at 18:02 | history | rollback | kelalaka |
Rollback to Revision 2
|
|
| Feb 1, 2022 at 18:00 | history | edited | Peter Morris | CC BY-SA 4.0 |
Added additional questions regarding Argon2id
|
| Feb 1, 2022 at 17:23 | comment | added | Peter Morris | @marcelm Ah I see, thanks for that information, I appreciate it! | |
| Feb 1, 2022 at 13:17 | answer | added | U. Windl | timeline score: 1 | |
| Feb 1, 2022 at 12:12 | comment | added | marcelm | @PeterMorris No, you created your own password hashing function. The fact that you used SHA512 as a primitive in your function is irrelevant. It's pretty easy to create an insecure algorithm using secure building blocks. AES-ECB would be a well-known example of that. Cryptography is very hard to get right, that's why people who know what they're doing only use well-vetted algorithms that have stood up to public scrutiny. Rolling your own crypto opens you up to much larger risks than using well-known algorithms. | |
| Feb 1, 2022 at 11:52 | comment | added | Peter Morris | I didn't create a crypto hashing function, I used existing ones. | |
| Feb 1, 2022 at 9:53 | comment | added | ScottishTapWater | @SethR - That's not necessarily true... I'm both an expert software and electronics engineer, to a postgraduate level. I still use Stack Overflow and Electronics SE because other experts use it too and nobody, not even experts, can know everything. | |
| Feb 1, 2022 at 9:00 | history | tweeted | twitter.com/StackCrypto/status/1488436976491634691 | ||
| Jan 31, 2022 at 20:27 | comment | added | Seth R | @Persistence experts in cryptography are not looking for advice on Stack Exchange. If you are reading this, you should not be creating hash functions. | |
| Jan 31, 2022 at 13:56 | comment | added | ScottishTapWater | You have a point... The Dunning Kreuger effect is strong | |
| Jan 31, 2022 at 12:40 | comment | added | marcelm | @Persistence No, because people tend to think they're an expert in something they're not. This comment is made in the context of countless SE questions about hobe-brew crypto, and the answer is always: don't. Actual cryptographers who do this professionally won't be stopped by a comment on crypto.SE anyway. | |
| Jan 31, 2022 at 11:50 | comment | added | ScottishTapWater | @marcelm - If everyone followed that advice we'd have no password hashing functions... Maybe couch that with "unless you're an expert in cryptography" | |
| Jan 31, 2022 at 9:35 | answer | added | Meir Maor | timeline score: 10 | |
| Jan 31, 2022 at 8:19 | comment | added | marcelm | Do not create your own password hashing function, full stop. | |
| Jan 31, 2022 at 6:51 | history | became hot network question | |||
| Jan 31, 2022 at 5:16 | comment | added | Royce Williams | Agreed that the salts are larger than needed - but salts are for not just resistance to parallel computation, but also resistance to pre-computation. So they need to be not just unique within the scope of the target hashlist / platform, but also practically unique globally - enough to make it infeasible to precompute and then store them. For example, bcrypt and scrypt use a 128-bit (16 byte) salt - still quite a bit smaller than 64 bytes, but enough to make building a list of all values largely infeasible. | |
| Jan 31, 2022 at 2:30 | comment | added | Gordon Davisson | You're putting way too much emphasis on the salt -- all it really needs is to be unique, so an attacker can't run parallel attacks against multiple accounts that have the same salt. Beyond that is overkill. | |
| Jan 30, 2022 at 23:52 | history | edited | kelalaka | CC BY-SA 4.0 |
polish and password tag
|
| Jan 30, 2022 at 23:17 | answer | added | kelalaka | timeline score: 33 | |
| S Jan 30, 2022 at 22:50 | review | First questions | |||
| Jan 31, 2022 at 2:27 | |||||
| S Jan 30, 2022 at 22:50 | history | asked | Peter Morris | CC BY-SA 4.0 |