Re:The numbers that explain Intel’s decline
by jiggawatts on Thursday November 24, 2022 @06:10PM
(#33737260)
When I was a dev working on small teams, say, 2-3 people, we all had a collective "ownership" of the product. If some page was slow, taking multiple seconds to load, we would simply fix it.
I just noticed that a due to a small error (misconfigured caching), a popular government's website is taking 4+ seconds to load its most used page.
There are three teams responsible for this thing. The team that "owns" it are scientists, not programmers, so they literally don't understand anything I'm saying. Their manager is 60 years old an probably still has his emails printed out for him by his secretary.
The database was done by a "information" team, and they don't care about the poor performance of downstream apps.
The downstream app was written by a "web" team, and they're (somewhat correctly) blaming the database team and their schema.
Everyone points fingers at each other. Years go past, the site is rewritten at great expense, but that doesn't help. Meanwhile the contractor that wrote the database has left to go back to India. The scientists upgrade the performance of their bit... because that's what they understand, but that has no bearing on the performance of a system three steps removed from their systems.
All attempts at fixing any of this in a meaningful way get rejected because "they're waiting on Big Name Consultancy to provide a report", or because "the next release 2-3 years from now will fix it", or... whatever.
It's just incredible to watch how fantastically inept these big orgs can get.
The fix for this issue is one line of code, but after years of trying they've failed to apply it, because despite dozens of people across multiple teams being involved, nobody actually writes code. The DB coder quit, remember? The website development was outsourced. And so on.
I suspect much of the same thing is happening at Twitter, and also at Intel.
These orgs become fantastically disfunctional, because it's just meetings and paperwork, but no action. The action people get outnumbered 10-to-1, then 100-to-1, and even worse. They just get told no, they get "priorities" assigned to them, asked to justify everything they do, and so on.
> One of the people fired Musk’s latest purge was Ikuhiro Ihara, a highly respected senior software engineer who helped lead the push to expand tweets to 280 characters.
> Ihara fought relentlessly for 280 in the face of strong skepticism from top executives, I’m told. He co-authored the blog post explaining the value of 280 characters when it launched
I mean... holy cow. Some poor guy wasted a decent chunk of his adult life fighting internal politics for something that ought to have been an obvious upgrade. Crazy.
I just noticed that a due to a small error (misconfigured caching), a popular government's website is taking 4+ seconds to load its most used page.
There are three teams responsible for this thing. The team that "owns" it are scientists, not programmers, so they literally don't understand anything I'm saying. Their manager is 60 years old an probably still has his emails printed out for him by his secretary.
The database was done by a "information" team, and they don't care about the poor performance of downstream apps.
The downstream app was written by a "web" team, and they're (somewhat correctly) blaming the database team and their schema.
Everyone points fingers at each other. Years go past, the site is rewritten at great expense, but that doesn't help. Meanwhile the contractor that wrote the database has left to go back to India. The scientists upgrade the performance of their bit... because that's what they understand, but that has no bearing on the performance of a system three steps removed from their systems.
All attempts at fixing any of this in a meaningful way get rejected because "they're waiting on Big Name Consultancy to provide a report", or because "the next release 2-3 years from now will fix it", or... whatever.
It's just incredible to watch how fantastically inept these big orgs can get.
The fix for this issue is one line of code, but after years of trying they've failed to apply it, because despite dozens of people across multiple teams being involved, nobody actually writes code. The DB coder quit, remember? The website development was outsourced. And so on.
I suspect much of the same thing is happening at Twitter, and also at Intel.
These orgs become fantastically disfunctional, because it's just meetings and paperwork, but no action. The action people get outnumbered 10-to-1, then 100-to-1, and even worse. They just get told no, they get "priorities" assigned to them, asked to justify everything they do, and so on.
PS: Here's a fun example: https://nitter.it/CaseyNewton/status/1595818243620380672#m
> One of the people fired Musk’s latest purge was Ikuhiro Ihara, a highly respected senior software engineer who helped lead the push to expand tweets to 280 characters.
> Ihara fought relentlessly for 280 in the face of strong skepticism from top executives, I’m told. He co-authored the blog post explaining the value of 280 characters when it launched
I mean... holy cow. Some poor guy wasted a decent chunk of his adult life fighting internal politics for something that ought to have been an obvious upgrade. Crazy.