It's a lot nicer, but I see switching technologies as being a formula: if the benefit minus the cost is greater than zero, switch. The benefits can be gigantic and it can still be better not to switch if the costs are high enough. Your bank still runs COBOL, after all.
Most of the people writing these articles are in the best places to switch: noodling on the side, between gigs, or they're students. They're not running an enterprise supporting millions of litigious customers on hundreds of servers running millions of lines of proprietary software that all depend on MySQL. It would be absolute lunacy to switch in that case; even if Postgres were twice as fast, used half the space and cut your costs in half, it would still be hard to justify the cost and danger.
I think Postgres is a lot better, and anybody for whom the equation justifies it should switch. But that isn't going to be everybody every time, and that's fine, because we should make these decisions rationally rather than because it's cool now or it looks like it will be fun.
Most of the people writing these articles are in the best places to switch: noodling on the side, between gigs, or they're students. They're not running an enterprise supporting millions of litigious customers on hundreds of servers running millions of lines of proprietary software that all depend on MySQL. It would be absolute lunacy to switch in that case; even if Postgres were twice as fast, used half the space and cut your costs in half, it would still be hard to justify the cost and danger.
I think Postgres is a lot better, and anybody for whom the equation justifies it should switch. But that isn't going to be everybody every time, and that's fine, because we should make these decisions rationally rather than because it's cool now or it looks like it will be fun.