By Andy Valencia
I built a nice device for a huge company (well, I did most of the system design and all of the software). It made lots of money, and – yay! – I was a rock star at my company. Sic transit gloria mundi.
I measured its performance, and determined that it could, easily, handle three times as many users. The hardware team kicked off building such a thing, and I helped the software team that had taken over my code. It was going to be great!
One of my peers ran into me in the hallway, and we talked about our current projects. With mine, he expressed concern that it would be underpowered. I explained that the current product – full out – used 25% of the CPU, so tripling the size should be 75%. That left 25% as a buffer! I assured him that it was going to be fine. He left me with a skeptical face and a shrug.
The device came out and…it was underpowered. Actually, it was barely OK when used like its smaller predecessor. But all of the code had bloated, and that ate up most of the headroom. And as a hit product, the sales teams sold it as far and as wide as they could. It was suddenly being asked to do fancy things that demanded a lot more processing. It wasn’t pretty when it ran out of gas.
The funny thing is, I was right. It’s just that my peer was right, too. I knew all the details, and could provide a mathematical proof that I was right. My friend, not mired in the details, could do a dead reckoning of bloat and the enthusiastic over-promising of sales people. I was right in the details; he got it right in the big picture.
We limped along until the hardware folks – almost a year later – finally turned out the same product with a large CPU upgrade. Just like that, everything settled down. A lesson for me at my company’s expense.
Last month’s column was on various search engines. Some of them let you jump outside the corporate echo chamber, but that still leaves your own personal echo chamber. When you read search results, or Wikipedia articles, or social media posts, keep in mind that you can be correct – and the surprising or contrary statement that your finger is automatically left-swiping can be correct, too. You can learn from somebody, even though you’re sure they are wrong.
Elon Musk wants to make Twitter federated. This means he dreams of a world where @firstname.lastname@example.org can post a message which @email@example.com can see, comment upon, and boost. And @firstname.lastname@example.org can join in, too. Rather than one corporation, the discussion could span people using the services of large corporations, or small ones, and even people participating by way of a computer their local club owns. Or one they own for themselves. It’s all decentralized.
This isn’t imaginary. Parts of it already exist.
You can already dip your toe into these waters, because the protocols to implement this have been designed, and are being used. You can read about it by way of vashonloop.com/goto/masto and get yourself an account via joinmastodon.org. Did I mention no ads?
When you go out into a Twitter alternative without coporate filters, you’ll find people who are definitely wrong – and rude, sometimes. And you’ll find like-minded peers. The real treasure is finding new ideas that refine or even replace what you’ve “known” previously. They’ll be wrong and then…right, also. If you try out the Fediverse, I hope you’ll send me a note: @email@example.com.