By James Kwak
Last summer, Lawrence Baxter wrote these two posts about the toxic combination of bad software—actually, software in general, since no software system is perfect—and too-big-to-fail banks. Baxter knows whereof he speaks, as he was previously a technology executive at a very large bank. Here’s what he has to say about it:
I don’t care what a CIO or even a CEO might say: if they claim that they can eliminate the real risk of such missteps, they just don’t know what they are talking about no matter how good they are. And if such missteps are inevitable, then we simply cannot avoid the question whether the dangers posed by large, complex financial institutions and systems could outweigh their benefits.
Think about that the next time you hear some CEO talking about his company’s state-of-the-art technology.
10 thoughts on “Another Perspective on Bad Software”
Oh that’s just great.
We’re going to litigate the economic destruction of millions of USA citizens from the perspective of Lady GIGO being stupid.
Moral high ground? I think not.
James, do keep in mind that the rise in bogus software is isomorphically related to the demise of a proper double-entry Book-of-Accounts. When half baked bookkeeping-software — 94% of today’s market — began storing its data in relational databases, which is retrieved by a System-Query-Language (SQL), rather than to store that bookkeeper’s history in a journal set isomorphic to its ledger, typical of a Lisp Stream, all hope for a proper data-control that would proves the validity of historical data was lost. SQL never has and never will provide a proof of balanced accounts, and their audit-trail. Only the seven centuries old Book-of-Accounts provides that power to audit for a compete accountancy.
Check out “Normal Accidents: Living with High-Risk Technologoies” by Perrow. http://amzn.to/10S1XBG
Catastrophic “accidents” are inevitable in tightly-coupled systems that have little slack built into them. We will be visiting 2007-08 again, soon.
Two months ago, we got a home mortgage that got sold to Chase. Chase has been remarkably unable to keep track of information associated with that mortgage. They’ve lost our mortgage payments, our homeowner insurance information, our flood insurance information, and our temporary mailing address. We do get automated phone calls from them complaining that something is missing. Chase, too big to fail, may also be too big to work for retail banking.
The thing that makes software and system-maintenance issues particularly problematic is that the bad events they lead to are generally entirely unrelated, aside from being software-related. There really is no history of that particular thing going wrong, no way to predict that it, or something similar to it would go wrong, and it probably will never happen again. It’s like watching a new Terry Gilliam animation: we can expect that something unusual will happen, but we can’t say anything about what it will be, except that it’s certainly nothing our experience has led us to prepare for, because he doesn’t repeat himself like that. These events really are “black swans”, in that there’s no way to prepare for them in advance and any particular one is extremely unlikely, but they’re collectively quite common, which is the real problem. You wear seat belts in cars, hard hats in construction sites, eye protection in machine shops, but there’s no safety equipment you can wear to prepare for completely generic danger, which is the situation they’re in.
@SamB – If they don’t have any of that information, then how are they claiming to own the mortgage? Based on what transaction? What are the DETAILS of that transaction? What is the LAW they wrote for themselves that basically allows them to claim virtual ownership over everything material in the material world?
Does Simon write here anymore?
You wouldn’t believe how toxic a combination of badly managed software development process/strategy and bankers is.
Hardest projects and tightliest deadlines I worked on as a s.developer were for the banks. Detail-are-not-important youre-expensive-and-slow zombies kicking one’s butt all the time.
Even boring retail banks give me hard time when I think where I should put my money only to sit still and be available through an ATM.
There are no real enforced standards, no real chain of responsibility for technical issues. You see people’s transactions get messed up with other people’s and only crosschecking forces a rush for those irresponsible and overworked coders to fix things.
Retail banking software is a joke, trading software…has a risk-reward ratio when deploying production stuff.
Good programmers don’t make good execs and execs don’t make good programmers.
oney and Speed: Inside the Black Box is a thriller based on actual events that takes you to the heart of our automated world.
“Based on interviews with those directly involved and data visualizations up to the millisecond, it reconstructs the flash crash of May 6th 2010: the fastest and deepest U.S. stock market plunge ever. Money and Speed: Inside the Black Box is developed by filmmaker Marije Meerman in close collaboration with design studio Catalogtree.
This exploratory documentary is a marriage of strong storytelling and meticulous visual analysis. A rare opportunity to experience what is happening inside the black boxes of our rapidly evolving financial markets.”
Money and Speed: Inside the Black Box
Posted: 22 Dec 2012 04:20 AM PST
@Woych – Thanks for the link to the flash crash documentary.
Now I’m wondering how many times I can slap someone in a milli-second…
The FEAR of upsetting the gamesters who had parasite’d themselves on to a legitimate trade done by one of their pre-selected victims is impressive, isn’t it? No one wants that 360 head spin foaming mouth fat cat plutocrat to come after them for revealing too much *math* in the black box.
Comments are closed.