4 ms·

> There is just no way to securely create a secret key. That is wrong. You can generate one relatively easily with dice (google for it). > Because the algorit

by pg314 9y ago

> There is just no way to securely create a secret key. That is wrong. You can generate one relatively easily with dice (google for it). > Because the algorithm is so complex, you have to trust somebody else's software to create the key. You're conflating generating the key, which is just generating a random number in a certain range, with calculating the public address, which involves some operations on elliptic curves. The algorithm is not that complex. Using a library like openssl/ec it is not that hard to generate the public address. It takes about 30 lines of C code using that library. > How do you know that software is neither malicious nor buggy? You can read the source code.

TekMol 9y ago
You can read the source code. Which code? Any link to that 30 lines of C code?
- pg314 9y ago
  The source code of whatever software you're using. Here's some C code I used when I played around with it 5 years ago. No guarantees that it is bug-free or does anything useful (you still need a double sha256 and base58 encoding to get a string representation), but it gives you an idea of the complexity/simplicity. void sha256_ripe160(uint8_t *bytes, size_t len, uint8_t result[RIPEMD160_DIGEST_LENGTH]) { uint8_t buf[32]; CC_SHA256(bytes, len, buf); RIPEMD160(buf, 32, result); } void privkey_to_pubsig_hash(uint8_t privkey[32], uint8_t pubsig[20]) { static EC_GROUP *curve = 0; if (!curve) { curve = EC_GROUP_new_by_curve_name(NID_secp256k1); } assert(curve); BIGNUM *priv_bn = BN_new(); priv_bn = BN_bin2bn(privkey, 32, priv_bn); assert(priv_bn); EC_POINT *pub_pt = EC_POINT_new(curve); BN_CTX *ctx = BN_CTX_new(); int result = EC_POINT_mul(curve, pub_pt, priv_bn, NULL, NULL, ctx); assert(result); BN_free(priv_bn); BIGNUM *pub_bn = BN_new(); EC_POINT_point2bn(curve, pub_pt, POINT_CONVERSION_UNCOMPRESSED, pub_bn, ctx); EC_POINT_free(pub_pt); BN_CTX_free(ctx); assert(BN_num_bytes(pub_bn) == 65); uint8_t pub_bytes[65]; BN_bn2bin(pub_bn, pub_bytes); BN_free(pub_bn); sha256_ripe160(pub_bytes, 65, pubsig); } One of the myriad Python libraries to deal with bitcoin is probably a better starting point if you want to explore more.

by pg314 9y ago

TekMol 9y ago

You can read the source code. Which code? Any link to that 30 lines of C code?

pg314 9y ago
The source code of whatever software you're using. Here's some C code I used when I played around with it 5 years ago. No guarantees that it is bug-free or does anything useful (you still need a double sha256 and base58 encoding to get a string representation), but it gives you an idea of the complexity/simplicity. void sha256_ripe160(uint8_t *bytes, size_t len, uint8_t result[RIPEMD160_DIGEST_LENGTH]) { uint8_t buf[32]; CC_SHA256(bytes, len, buf); RIPEMD160(buf, 32, result); } void privkey_to_pubsig_hash(uint8_t privkey[32], uint8_t pubsig[20]) { static EC_GROUP *curve = 0; if (!curve) { curve = EC_GROUP_new_by_curve_name(NID_secp256k1); } assert(curve); BIGNUM *priv_bn = BN_new(); priv_bn = BN_bin2bn(privkey, 32, priv_bn); assert(priv_bn); EC_POINT *pub_pt = EC_POINT_new(curve); BN_CTX *ctx = BN_CTX_new(); int result = EC_POINT_mul(curve, pub_pt, priv_bn, NULL, NULL, ctx); assert(result); BN_free(priv_bn); BIGNUM *pub_bn = BN_new(); EC_POINT_point2bn(curve, pub_pt, POINT_CONVERSION_UNCOMPRESSED, pub_bn, ctx); EC_POINT_free(pub_pt); BN_CTX_free(ctx); assert(BN_num_bytes(pub_bn) == 65); uint8_t pub_bytes[65]; BN_bn2bin(pub_bn, pub_bytes); BN_free(pub_bn); sha256_ripe160(pub_bytes, 65, pubsig); } One of the myriad Python libraries to deal with bitcoin is probably a better starting point if you want to explore more.