3 ms·

No, because the field that in one system is called "cash" is called "settled" in another; system A uses BCD for money, system B uses 32 bit unsigned ints reckon

by dsr_ 3y ago

No, because the field that in one system is called "cash" is called "settled" in another; system A uses BCD for money, system B uses 32 bit unsigned ints reckoned in millicents, and system C uses a cash object which can define original currency, calculated currency value, guaranteed conversion preference, contract reference number, and precision. Understanding semantics is required, not just form.

raverbashing 3y ago
Yes, you got it. But I still think AI can at least make it easier
- rvba 3y ago
  IT guys are paid a lot not to be lazy. Changing data types in an interface is a boring job (didnt Office Space joke at that?) but doing this via some broken AI will probably introduce even more spaghetti. Reality is that those systems should be killed one by one and finally when there is one system, it should be upgraded to something better. What is very difficult.
  - shapefrog 3y ago
    20mil for the new system, 80mil + 5 years of implementing for the translation layers to communicate with the legacy systems. > Reality is that those systems should be killed one by one and finally when there is one system It take so long to replace each of the systems, by the time you get to the "one true solution" 40 years have passed and you are back where you started.
- simonh 3y ago
  If you mean deep learning type systems, you need to train them on vast training sets of similar data so they can learn very general patterns, similarities and connections. Like giving them 10,000 pictures of dogs, so they can learn to recognise pictures of dogs. These banking systems are all massively complex one off systems, often with very different sets of functions or even definitions of the problem, and the implementation details are different in almost every way possible. They’re the exact opposite of the sorts of problems modern trained AI systems are good at dealing with.
  - osclarto 3y ago
    well said
- ta988 3y ago
  If you can write one in COBOL and it can finish processing the whole set of transactions in less than 24h you may have a chance. /s

by dsr_ 3y ago

raverbashing 3y ago

Yes, you got it. But I still think AI can at least make it easier

rvba 3y ago
IT guys are paid a lot not to be lazy. Changing data types in an interface is a boring job (didnt Office Space joke at that?) but doing this via some broken AI will probably introduce even more spaghetti. Reality is that those systems should be killed one by one and finally when there is one system, it should be upgraded to something better. What is very difficult.
- shapefrog 3y ago
  20mil for the new system, 80mil + 5 years of implementing for the translation layers to communicate with the legacy systems. > Reality is that those systems should be killed one by one and finally when there is one system It take so long to replace each of the systems, by the time you get to the "one true solution" 40 years have passed and you are back where you started.
simonh 3y ago
If you mean deep learning type systems, you need to train them on vast training sets of similar data so they can learn very general patterns, similarities and connections. Like giving them 10,000 pictures of dogs, so they can learn to recognise pictures of dogs. These banking systems are all massively complex one off systems, often with very different sets of functions or even definitions of the problem, and the implementation details are different in almost every way possible. They’re the exact opposite of the sorts of problems modern trained AI systems are good at dealing with.
- osclarto 3y ago
  well said
ta988 3y ago
If you can write one in COBOL and it can finish processing the whole set of transactions in less than 24h you may have a chance. /s