3 ms·
That is true, but our benchmarking showed that it was worth it, since we saved so much on hardware/rack space compared to what we'd have needed to deliver equiv
by gaius 13y ago
That is true, but our benchmarking showed that it was worth it, since we saved so much on hardware/rack space compared to what we'd have needed to deliver equivalent functionality with a forking CGI/Perl implementation ;-)
- tptacek 13y agoIt turns out that writing web apps in C++ to run on $40,000 Sun servers sitting behind a $35,000 Cisco router hooked up a set of bonded DS1s is not actually a great way to launch a new web product, so I'm not sure I'm ready to concede that this was a "totally reasonable" way to develop products. But, I was there too, and yes, that is how everyone did things. My point though isn't simply that things sucked, but that people should do some mental multiplication when they compare 1999 startups to 2013 startups; it's like comparing the Spruce Goose to a modern military UAV.
- gaius 13y agoSo what should we have used? Rails? Just that little matter of it not being so invented yet. Given that we knew C++, and were smart enough to build with reuse in mind, and needed far, far fewer $40,000 servers (and the REAL cost, co-lo space), it wasn't just reasonable, it was genius! Speaking of genius, that leads me neatly onto: are things really cheaper? I am thinking of this story, even in '99 I'd have balked at $20k/mo in hosting charges alone http://www.theregister.co.uk/2013/02/15/heroku_marketing_misstep_draws_customer_anger/ http://www.theregister.co.uk/2013/02/15/heroku_marketing_mis...
- 3pt14159 13y agoASP pre dot net is what you should have used. It was perfect for web apps and way faster to develop in than C. I am 100% serious. I worked at a startup back then (one of the few that survived) and that was what we used.
- dba7dba 13y agoNice to see someone who knows about 'Spruce Goose'.