Fun times, working for a charity group maintaining donation records, wrapping Enterprise Donor Relation Management System in a protective layer of spreadsheets, because that was the only tech I was permitted to use in my woefully mis-scoped, woefully underpaid role. (Couldn't even use VBA.) I had Excel formulas to detect which bank sort codes had been helpfully converted into dates and produce a normalised column of sort codes i.e. exactly what was in the source file
Another excellent talk from Kevlin, it covers a lot of the things I keep banging on about (and which the "smart kids" ignore until they get bitten in the ass).
it is a bug, because the programmer wrongly assumed that (a+b)/2 in C works _the same way_ as it did in the algorithm description (which is completely oblivious of hardware specifics). when kevlin says "bug in the binary search algorithm" he implies "bug in _this_ binary search algorithm _implementation_ " it works the same way _only_ if a plus b is less than or equal to the largest integer representable in the type of a and b
"the time function shall return the value of time in seconds since the epoch" um "the time function must return the count of seconds between its call time and the epoch" wtf is "the value of time" ?!
![Procedural Programming: It’s Back? It Never Went Away - Kevlin Henney [ACCU 2018]](http://i.ytimg.com/vi/mrY6xrWp3Gs/mqdefault.jpg)
I've heard most of these stories before!
... ... ... but Kevlin tells them so well, I keep coming back.
Intel updates for 1970 xD
That's a feature from their time machine, accidently leaked!
Kvelin is a great speaker!
Thanks for the upload.
Fun times, working for a charity group maintaining donation records, wrapping Enterprise Donor Relation Management System in a protective layer of spreadsheets, because that was the only tech I was permitted to use in my woefully mis-scoped, woefully underpaid role. (Couldn't even use VBA.) I had Excel formulas to detect which bank sort codes had been helpfully converted into dates and produce a normalised column of sort codes i.e. exactly what was in the source file
Another excellent talk from Kevlin, it covers a lot of the things I keep banging on about (and which the "smart kids" ignore until they get bitten in the ass).
5:54 That's not PHP. Great talk nonetheless.
Isn't it a bit of a stretch calling it a bug in that Binary Search algorithm? Every algorithm fails when the basic commands of the language misbehave
The command didn't misbehave, the programmer used the wrong type because his assumption of the maximum value of the expression was incorrect.
it is a bug, because the programmer wrongly assumed that (a+b)/2 in C works _the same way_ as it did in the algorithm description (which is completely oblivious of hardware specifics).
when kevlin says "bug in the binary search algorithm" he implies "bug in _this_ binary search algorithm _implementation_ "
it works the same way _only_ if a plus b is less than or equal to the largest integer representable in the type of a and b
"the time function shall return the value of time in seconds since the epoch"
um
"the time function must return the count of seconds between its call time and the epoch"
wtf is "the value of time" ?!
The value of time is $128.42
@@Keisuki 🤔 how much accuracy does that get me?
@@GeorgeTsiros For $128.42, you can buy 1.042 degrees of accuracy
😜it works
Don't touch it
🤚🏽