nightcraw1er.488: Look, I didn’t say it was fair or nice, but that’s the world. If they close down it’s up to them if they keep servers going. Oh, and md5 is fine for general use, mostly to check I haven’t deleted something or changed it accidentally. I know it has issues, and maybe one day I will get round to redoing it.
B1tF1ghter: NO. That's the point.
It not only allows SAME CHECKSUM for 2 different stacks of data * but it ALSO is UTTERLY WORTHLESS for detecting changes for the VERY SAME REASON.
* Md5 is fundamentally FLAWED on a DESIGN level. It imo should have never be let out into the wild and used.
It's a worthless trash. It always has been but people kept denying it for many years.
There is this expcetionally dumb attitude thing in IT industry (and I say that as a person involved in said industry) - something that I hate and never really understood - for a lack of a better term let's call this "lenient underestimation".
People worldwide choose to use "the least difficulty path" when designing software, as in:
Using the kind of algorithms or their complexity levels (say in conversion matrices for hashing), least derivation rounds, shortest possible string for filesystem timestamps, and so on.
It's like a self fulfilling prophecy "free for all".
I always found it beyond dumb.
You can LITERALLY calculate *when* things will INEVITABLY go to utter SH*T just because someone decided to prioritize ease of coding and / or SLIGHTLY, marginally, perhaps *undetectably*, faster code speed over than longterm reliability and longetivity.
So for example in 2038 things will majorly go to sh*t, it will get FAR WORSE than "year 2k", JUST BECAUSE some people decided to use attrociously "the shortest possible" lowered number space for timestamps almost EVERYWHERE in Linux code.
We are ALREADY IN 2K21 and BARELY ANYTHING is already fixed to fix the problem!
Like what the F!
Humans really do seem to be LAZY and LOVE waiting until the very last moment! >:(
In which case it may already be too late!
How is this related to MD5?
It is directly.
It:
1.Was badly designed therefore TECHNICALLY really WORTHLESS *since day one*
2.Was used for WAY TOO LONG by the industry and SOMEHOW it's STILL USED, even tho it was MULTIPLE TIMES already PROVEN it's WORTHLESS to the point it's DANGEROUS to use it.
I'm sorry dude but I don't feel like giving you a lecture on HOW and WHY.
It may take you YEARS to dig through enough docs to understand this.
So just note this:
MD5? WORTHLESS. Should not be used for ANYTHING. Not even for absurdly unimportant stuff. That is unless you are a masochist that loves pain and loosing time.
SHA1 - some people debate it, but it is BROKEN already, generally speaking use this only if you really NOT CARE about what you are hashing with it
sha256 - TECHNICALLY "still" "usable" tho this is about the time to stop using it if you SERIOUSLY care about data integrity in the longterm and security, it's ALREADY broken and it's imo pretty funny how most people in IT industry seem to not notice that at all
KNOWING how this all works and simultaneously using anything less than sha256 is INSANITY.
It's not really my problem that there are double standards EVERYWHERE and people delude themselves into thinking "it's 'still' fine for now".
It's not.
Things are different than most people in the industry worldwide report.
Don't repeat peoples mistakes.
Don't use double standards.
Don't cheap out.
Do it right or don't do it at all.
PLEASE, don't repeat others' mistakes just because majority is IDIOTS who don't know any better yet work in the industry!
If you care abour data integrity and security read this (as a starter, there is MUCH MUCH MUCH more to read on this):
https://en.wikipedia.org/wiki/SHA-2#Comparison_of_SHA_functions Tldr:
My personal advice is to use sha3-512 and something like Blake & ChaCha20 derivatives, as well as for example shake512 for data integrity. (use MORE THAN ONE per file)
You care if you CHANGED something in a file?
Well then MD5 should IMMEDIATELLY get out through a window for you.
Because you can LITERALLY get a bitflip in a file and the MD5 of it may not reflect the change with VERY high probability as that algorithm is badly DESIGNED.
nightcraw1er.488: I doubt most TOS, or Eulas, or anything like that has any real standing. You would need a lot of money and time to find out. Sorry, I didn’t understand your point on galaxy (it’s hardly been optional for years now).
B1tF1ghter: Yes well since these services are public (more or less) a lot of people can crawl through their TOS.
I can tell you that for example HumbleBundle TOS is literally illegal in EU and it's only a matter of time until they get sued by somebody.