I think Jacques' fundamental point is completely without basis.
I ran the technical side of the Compliance group for a company managing over 850 billion dollars in assets. Ask me if I had a parallel technical organization to the trading, reporting, and analytics development teams sufficient to comprehend, test and validate every program and every process to the satisfaction of the union of all of the regulatory authorities.
Now ask me if, with my team of two, including myself, it ever arose that a certain very important three-letter financial regulatory authority sent a letter explaining that in two weeks, they would be performing an audit, which among other things, required the holdings records during a certain period of a british subsidiary which our company had purchased a few years ago. Ask me if that company kept their holdings records on a Wang minicomputer which had been decommissioned in the purchase. Ask me if, when we asked for that computer to be urgently shipped to the States so that we could attempt to get the holdings records off the hard drives, if it emerged that (a) the company had skimped on buying the proprietary SCSI rails for the 5.25" SCSI hard drives they had added to the system, instead balancing the SCSI drives inside the case on little blocks of wood on top of the other SCSI hard drives; and (b) when shipping the computer, they shipped it inside a box that snugly fit the case, without padding, resulting in the SCSI hard drives jostling off the wooden blocks and, in shipping, rubbing up and down against the motherboard during international transit to such an extent that when the case was opened, all of the chips, resistors, and capacitors had been cleanly scraped off the gleaming silver-and-green motherboard and formed a fine pile of gravel-consistency metal and plastic at the bottom. Ask me if we then contacted a Wang expert to determine if the SCSI hard drives were still viable, only to be told that the price for consultation would be six times the yearly budget for the entire team. Ask me if we then had giant reams of paper overnighted from England and then spent the next seven days, 20 hours a day, the two of us, typing in numbers from printouts of old trading records in a desperate attempt to rebuild the position records for the requested accounts, cross-checking each others' work, calling the traders to make sure everything seemed right, and finally finishing the rebuild two hours before the auditors walked in the door, drank our coffee, looked at the directory listing (and just the directory listing) of the thumb drive of data we provided to them, checked off a box on their sheet, and walked out.
Now, I further need you to ask me, was this sort of thing routine? Were we, routinely, put in a position, of, e.g., discovering that a senior lead trader’s excel spreadsheet upon which he managed an entire line of business no longer worked because he had exceeded Excel’s nested IF limit, which was two hundred and fifty five, and that as a result, an entire month’s worth of critical data didn’t even exist any more in any form at all, not even in a reverse engineerable way?
The world of actual compliance is not nearly as crisp and obvious as Jacques seems to believe. I do understand the desire to deflect blame from the engineers, but at the end of the day, the engineers did type the code.
I don’t think that Jacques intends to absolve the engineers of all responsibility. However, I think he’s generally on point with the argument that bad engineering is usually the fault of management, either because unrealistic deadlines are set, unethical requests are made, or compliance support is offered but underdelivered. How one should allocate blame can be debated, but I do think that the blame of engineers is more reflective of a “shit rolls downhill” corporate system than of where fault actually lies.
I find it strange that all the numbers attached to a vehicle (mileage, emissions, horsepower) are all taken on the company’s word and the (US) government never checks any of it. Would it be that expensive for the EPA or NHTSA to test each company’s new car models once a year? This year it was VW cheating emissions, last year it was Hyundai faking mileage numbers (and was fined), the list goes on and on…
I think Jacques' fundamental point is completely without basis.
I ran the technical side of the Compliance group for a company managing over 850 billion dollars in assets. Ask me if I had a parallel technical organization to the trading, reporting, and analytics development teams sufficient to comprehend, test and validate every program and every process to the satisfaction of the union of all of the regulatory authorities.
Now ask me if, with my team of two, including myself, it ever arose that a certain very important three-letter financial regulatory authority sent a letter explaining that in two weeks, they would be performing an audit, which among other things, required the holdings records during a certain period of a british subsidiary which our company had purchased a few years ago. Ask me if that company kept their holdings records on a Wang minicomputer which had been decommissioned in the purchase. Ask me if, when we asked for that computer to be urgently shipped to the States so that we could attempt to get the holdings records off the hard drives, if it emerged that (a) the company had skimped on buying the proprietary SCSI rails for the 5.25" SCSI hard drives they had added to the system, instead balancing the SCSI drives inside the case on little blocks of wood on top of the other SCSI hard drives; and (b) when shipping the computer, they shipped it inside a box that snugly fit the case, without padding, resulting in the SCSI hard drives jostling off the wooden blocks and, in shipping, rubbing up and down against the motherboard during international transit to such an extent that when the case was opened, all of the chips, resistors, and capacitors had been cleanly scraped off the gleaming silver-and-green motherboard and formed a fine pile of gravel-consistency metal and plastic at the bottom. Ask me if we then contacted a Wang expert to determine if the SCSI hard drives were still viable, only to be told that the price for consultation would be six times the yearly budget for the entire team. Ask me if we then had giant reams of paper overnighted from England and then spent the next seven days, 20 hours a day, the two of us, typing in numbers from printouts of old trading records in a desperate attempt to rebuild the position records for the requested accounts, cross-checking each others' work, calling the traders to make sure everything seemed right, and finally finishing the rebuild two hours before the auditors walked in the door, drank our coffee, looked at the directory listing (and just the directory listing) of the thumb drive of data we provided to them, checked off a box on their sheet, and walked out.
Now, I further need you to ask me, was this sort of thing routine? Were we, routinely, put in a position, of, e.g., discovering that a senior lead trader’s excel spreadsheet upon which he managed an entire line of business no longer worked because he had exceeded Excel’s nested IF limit, which was two hundred and fifty five, and that as a result, an entire month’s worth of critical data didn’t even exist any more in any form at all, not even in a reverse engineerable way?
The world of actual compliance is not nearly as crisp and obvious as Jacques seems to believe. I do understand the desire to deflect blame from the engineers, but at the end of the day, the engineers did type the code.
I don’t think that Jacques intends to absolve the engineers of all responsibility. However, I think he’s generally on point with the argument that bad engineering is usually the fault of management, either because unrealistic deadlines are set, unethical requests are made, or compliance support is offered but underdelivered. How one should allocate blame can be debated, but I do think that the blame of engineers is more reflective of a “shit rolls downhill” corporate system than of where fault actually lies.
I find it strange that all the numbers attached to a vehicle (mileage, emissions, horsepower) are all taken on the company’s word and the (US) government never checks any of it. Would it be that expensive for the EPA or NHTSA to test each company’s new car models once a year? This year it was VW cheating emissions, last year it was Hyundai faking mileage numbers (and was fined), the list goes on and on…