> All of the current histrionics that we hear from DeFi advocates regarding escaping the evil centralized Banks and Regulators and what not
There are several different people that seem inclined to be attracted to crypto. This is just one and there's a huge crossover with gold bugs. There are also people who have made a lot of money or missed out on making a lot of money and have bought into the narrative that crypto is the future and/or they're desperately seeking to be in on day one of the next Bitcoin.
I agree about rewrites in general. Almost always, in fact. But I also believe in software entropy and it can reach a point where the current requirements are so far removed from the original requirements that subsequent changes can become increasingly expensive and risky to the point where a partial or total rewrite might make sense. But people also pull the trigger way too often.
Considerations for a rewrite:
1. Timeline. Will the current system stagnate for a year? If so, it's a problem;
2. Can a rewrite be partial and coexist with the current system? If not, huge problem. You reduce timelines and risk by planning for partial rewrites; and
3. Is a rewrite or migration reversible? If not, it's a red flag.
It's also why it's so important to build in the capability to upgrade the system cleanly in part. A good example (of what not to do) is Git's utter reliance on SHA1 hashes. At the time it came out I'm absolutely shocked there was no allowance for updating the hashing algorithm given that MD5 obsolescence was recent history at that point.
> And I do think that there are equal parts incompetence and laziness at play.
I think naivete plays a big part too. That and hubris ("this time will be different").
There are several different people that seem inclined to be attracted to crypto. This is just one and there's a huge crossover with gold bugs. There are also people who have made a lot of money or missed out on making a lot of money and have bought into the narrative that crypto is the future and/or they're desperately seeking to be in on day one of the next Bitcoin.
I agree about rewrites in general. Almost always, in fact. But I also believe in software entropy and it can reach a point where the current requirements are so far removed from the original requirements that subsequent changes can become increasingly expensive and risky to the point where a partial or total rewrite might make sense. But people also pull the trigger way too often.
Considerations for a rewrite:
1. Timeline. Will the current system stagnate for a year? If so, it's a problem;
2. Can a rewrite be partial and coexist with the current system? If not, huge problem. You reduce timelines and risk by planning for partial rewrites; and
3. Is a rewrite or migration reversible? If not, it's a red flag.
It's also why it's so important to build in the capability to upgrade the system cleanly in part. A good example (of what not to do) is Git's utter reliance on SHA1 hashes. At the time it came out I'm absolutely shocked there was no allowance for updating the hashing algorithm given that MD5 obsolescence was recent history at that point.
> And I do think that there are equal parts incompetence and laziness at play.
I think naivete plays a big part too. That and hubris ("this time will be different").