6 ms·

> This is just an example of systemic problems that can and should be solved by increased automation and removing humans from the process I've seen this patter

> This is just an example of systemic problems that can and should be solved by increased automation and removing humans from the process I've seen this pattern fail too many times to take it seriously anymore. Programmers (myself included) tend to rather drastically underestimate the likelihood that data can end up in an error state. Humans make more mistakes, to be sure, but they can also sensibly correct mistakes when they discover them (or have them reported). Automated systems make fewer errors but once they do, the problem can be nigh irresolvable. Just look at the endless trouble we read about every few weeks from people who get their Google accounts suspended. Those people are suspended incorrectly by automated systems, and the widely-responded best recourse is to be well-enough connected to be able to get a human who works for Google to bat for you. Machines are great for increasing scale, but they are terrible for handling edge cases. If edge cases are important to your system, you need humans deeply involved.

qppo 6y ago
More specifically here the humans in the process are the credit reporting agencies, which shouldn't exist at all and neither should systems to keep them up to date. I've also seen automation fail spectacularly because of inconceivable edge cases during design. Almost universally it's because the process being automated was originally human, and the conceptual model of an automated process is as a human that doesn't sleep. If there's friction to structurally altering the process so it can be automated - including restricting the input data and output conditions - then you're going to see spectacular failures. Automation isn't perfect but it really falls over when the stakeholders don't understand that it isn't perfect and the process needs to be altered for the project to succeed.
- jimmaswell 6y ago
  Why should there be no credit reporting agencies? Should everyone get the same mediocre credit limits/interest rates/etc instead of responsible borrowers being rewarded?
  - qppo 6y ago
    credit reporting agencies != credit reporting itself. An unaccountable private enterprise should not have access to my financial information, let alone three of them. I feel that's the domain of a governmental agency with far less access to information and is or far more regulated than the existing bureaus. > Should everyone get the same mediocre credit limits/interest rates/etc instead of responsible borrowers being rewarded No, but non-borrowers or low borrowers shouldn't be penalized when they do need credit. I'm not against credit reporting, I'm against the usage of my data against me by people that don't have an incentive to make sure its accurate and keep it secure.
    - syshum 6y ago
      Well you can accomplish you goal with out having the government service as a credit reporting agency. Which I absolutely do not trust the government to do. What you really need is data ownership rights, this would allow you to tell a credit reporting agency to forget you, the consequence of that of course would be that you likely would have a hard time getting credit at all but it would be your choice. Even today, for most data, you have given permission for the data to be shared with the credit reporting agencies, it is in the contract or agreement you sign when you open an account, take out a loan or consume other financial services. >>I'm against the usage of my data against me by people that don't have an incentive to make sure its accurate and keep it secure. I find it odd that you believe government would have incentive to make sure it is accurate and secure.
      - hyperman1 6y ago
        As most countries have a working person registry and identity theft is one of these problems rampant in the US and rare in other country's : Why not trust the governement? While they are far from perfect, this article discusses the worst case behaviour of your governement, except it is unfixable. In another country, you go to the central registry to fix the root of the problem, and every other entity reads the fix. The question is not if they are secure etc... The question is if they are worse than a cluster of unaccountable megacorps.
m463 6y ago
> Programmers (myself included) tend to rather drastically underestimate the likelihood that data can end up in an error state. Have you priced ECC DDR4 recently? That stuff is expensive. What I find interesting is that the chances of a bit error is minute... but the number of bits in a common system is large. (Actually to a human mind I would say it is astronomically inconceivably large) I first became aware that bit errors <=> memory size were colliding when I heard about large sun systems were doing something called memory scrubbing. https://en.wikipedia.org/wiki/Memory_scrubbing https://en.wikipedia.org/wiki/Memory_scrubbing
- GauntletWizard 6y ago
  Yeah. ZFS on modern systems schedules a complete read and “resilver” once a month. This is an aggressive schedule - But only “Probably”. I’m of the opinion that if you don’t have the capability of rereading 100% and rewriting 1% of your entire “cold” dataset every week, your data is already lost.
  - m463 6y ago
    So is this because flash is not as permanent as spinning disks? or is it memory related?
    - GauntletWizard 6y ago
      Flash’s failure mode is generally better than spinning disk, in that complete drive failure (from head crash) is rarer and bad sectors, while technically “more common” on flash, are immensely predictable (X writes, +/- a few %) whereas on hard drives they’re fairly random. The real trick is that the forced exercise catches data errors (1 copy is missing/bad) long before they become corruptions (multiple copies are missing/bad). The bandwidth to read all of your flash disk in the space of a week is also much easier to come by, because flash has a much “better” ratio of bandwidth:storage size. The advice is very much rooted in being a “one size fits all” for both spinning rush and solid state - Spinning rust needs the checksums, flash can simply afford the waste.
syshum 6y ago
I have for years tried to get programmers to understand why active controls in software are necessary For to many times in LOB applications do I see error produced by users doing things "they were not suppose to do" or in an out of order manner or something like that to many programmers go with the "well the user should not have done that" instead of understanding that their program should have never allowed the error in the first place. Machine do as they are instructed by humans. Machines do not make mistakes, programmers do.
- lmm 6y ago
  You can't possibly anticipate everything that the user might want to do. Sometimes it's better for the tool to empower the user than to be all "I'm sorry, I can't let you do that Dave".

qppo 6y ago

More specifically here the humans in the process are the credit reporting agencies, which shouldn't exist at all and neither should systems to keep them up to date. I've also seen automation fail spectacularly because of inconceivable edge cases during design. Almost universally it's because the process being automated was originally human, and the conceptual model of an automated process is as a human that doesn't sleep. If there's friction to structurally altering the process so it can be automated - including restricting the input data and output conditions - then you're going to see spectacular failures. Automation isn't perfect but it really falls over when the stakeholders don't understand that it isn't perfect and the process needs to be altered for the project to succeed.

jimmaswell 6y ago
Why should there be no credit reporting agencies? Should everyone get the same mediocre credit limits/interest rates/etc instead of responsible borrowers being rewarded?
- qppo 6y ago
  credit reporting agencies != credit reporting itself. An unaccountable private enterprise should not have access to my financial information, let alone three of them. I feel that's the domain of a governmental agency with far less access to information and is or far more regulated than the existing bureaus. > Should everyone get the same mediocre credit limits/interest rates/etc instead of responsible borrowers being rewarded No, but non-borrowers or low borrowers shouldn't be penalized when they do need credit. I'm not against credit reporting, I'm against the usage of my data against me by people that don't have an incentive to make sure its accurate and keep it secure.
  - syshum 6y ago
    Well you can accomplish you goal with out having the government service as a credit reporting agency. Which I absolutely do not trust the government to do. What you really need is data ownership rights, this would allow you to tell a credit reporting agency to forget you, the consequence of that of course would be that you likely would have a hard time getting credit at all but it would be your choice. Even today, for most data, you have given permission for the data to be shared with the credit reporting agencies, it is in the contract or agreement you sign when you open an account, take out a loan or consume other financial services. >>I'm against the usage of my data against me by people that don't have an incentive to make sure its accurate and keep it secure. I find it odd that you believe government would have incentive to make sure it is accurate and secure.
    - hyperman1 6y ago
      As most countries have a working person registry and identity theft is one of these problems rampant in the US and rare in other country's : Why not trust the governement? While they are far from perfect, this article discusses the worst case behaviour of your governement, except it is unfixable. In another country, you go to the central registry to fix the root of the problem, and every other entity reads the fix. The question is not if they are secure etc... The question is if they are worse than a cluster of unaccountable megacorps.

m463 6y ago

> Programmers (myself included) tend to rather drastically underestimate the likelihood that data can end up in an error state. Have you priced ECC DDR4 recently? That stuff is expensive. What I find interesting is that the chances of a bit error is minute... but the number of bits in a common system is large. (Actually to a human mind I would say it is astronomically inconceivably large) I first became aware that bit errors <=> memory size were colliding when I heard about large sun systems were doing something called memory scrubbing. https://en.wikipedia.org/wiki/Memory_scrubbing https://en.wikipedia.org/wiki/Memory_scrubbing

GauntletWizard 6y ago
Yeah. ZFS on modern systems schedules a complete read and “resilver” once a month. This is an aggressive schedule - But only “Probably”. I’m of the opinion that if you don’t have the capability of rereading 100% and rewriting 1% of your entire “cold” dataset every week, your data is already lost.
- m463 6y ago
  So is this because flash is not as permanent as spinning disks? or is it memory related?
  - GauntletWizard 6y ago
    Flash’s failure mode is generally better than spinning disk, in that complete drive failure (from head crash) is rarer and bad sectors, while technically “more common” on flash, are immensely predictable (X writes, +/- a few %) whereas on hard drives they’re fairly random. The real trick is that the forced exercise catches data errors (1 copy is missing/bad) long before they become corruptions (multiple copies are missing/bad). The bandwidth to read all of your flash disk in the space of a week is also much easier to come by, because flash has a much “better” ratio of bandwidth:storage size. The advice is very much rooted in being a “one size fits all” for both spinning rush and solid state - Spinning rust needs the checksums, flash can simply afford the waste.

syshum 6y ago

I have for years tried to get programmers to understand why active controls in software are necessary For to many times in LOB applications do I see error produced by users doing things "they were not suppose to do" or in an out of order manner or something like that to many programmers go with the "well the user should not have done that" instead of understanding that their program should have never allowed the error in the first place. Machine do as they are instructed by humans. Machines do not make mistakes, programmers do.

lmm 6y ago
You can't possibly anticipate everything that the user might want to do. Sometimes it's better for the tool to empower the user than to be all "I'm sorry, I can't let you do that Dave".