4 ms·

Last year, in a fit of nostalgia I finally bought an Apple IIe, a system I had wanted 30 years ago, at a garage sale. Somehow in moving it back to my house, I

Last year, in a fit of nostalgia I finally bought an Apple IIe, a system I had wanted 30 years ago, at a garage sale. Somehow in moving it back to my house, I dislodged the floppy controller card, leaving the system to boot from ROM. I turned it on, it beeped, and there was the prompt. I perceived it as instantaneous. I know that from that point I could have written a BASIC program and saved it to a cassette tape. The anecdote falls apart once you actually start trying to boot ProDOS from disk and have to rely on the something besides solid state. This experience got me thinking. What has changed in the meantime? Computers have become communication devices and communications devices have become computers (though even this ancient Apple came to me with a modem installed). Developer time has outstripped hardware costs, and somehow an hour of developer time is worth the same as wasting a million hours of the users' time (1 million users x 1 hour or whatever formula you wish). I think we all recognize this problem, and it's already too expensive to fix. The hardware and software space are too federated, too balkanized, too complicated to ever integrate a system into a cohesive whole in the way the Apple II was.

ryandrake 9y ago
> Developer time has outstripped hardware costs, and somehow an hour of developer time is worth the same as wasting a million hours of the users' time Key point here. It's hard to get the biz guys and product managers to agree with spending developer hours on performance improvements, because it's hard to measure the effect it has on business metrics. Does making your application take 750ms to launch rather than 8 seconds really bring in more sales? Who knows? Nobody tries to measure it either, so we'll never fix the problem. Feature cram, on the other hand, is easier to justify, which is why every application eventually ends up with a plug-in interface, theme-able, skin-able, able to read E-mail, and able to interact with Facebook and Twitter.
- michrassena 9y ago
  And I'm not even talking specifically about application performance, though what you're saying is still absolutely spot on. Think of all the time wasted because of poor, or incomplete documentation. An I sympathize with anyone writing documentation. We're definitely still in the era of move fast and break things. These costs which accrue to everyone because no one can afford to take it on themselves, I would call externalities in the same sense we do when talking about pollution.

ryandrake 9y ago

> Developer time has outstripped hardware costs, and somehow an hour of developer time is worth the same as wasting a million hours of the users' time Key point here. It's hard to get the biz guys and product managers to agree with spending developer hours on performance improvements, because it's hard to measure the effect it has on business metrics. Does making your application take 750ms to launch rather than 8 seconds really bring in more sales? Who knows? Nobody tries to measure it either, so we'll never fix the problem. Feature cram, on the other hand, is easier to justify, which is why every application eventually ends up with a plug-in interface, theme-able, skin-able, able to read E-mail, and able to interact with Facebook and Twitter.

michrassena 9y ago
And I'm not even talking specifically about application performance, though what you're saying is still absolutely spot on. Think of all the time wasted because of poor, or incomplete documentation. An I sympathize with anyone writing documentation. We're definitely still in the era of move fast and break things. These costs which accrue to everyone because no one can afford to take it on themselves, I would call externalities in the same sense we do when talking about pollution.