Wrong trade-off. The time is DEVELOPMENT time, not PERFORMANCE

Wrong trade-off. The time is DEVELOPMENT time, not PERFORMANCE

Yes, there's a time vs. memory trade-off. But you're measuring the wrong time. The time that is getting traded off is development time.

On the desktop, GNU/Linux systems are basically in "catch-up" mode, trying to implement a complete GUI desktop with highly honed, mature, and featureful applications faster than the current dominant vendor. Clearly, OSS/FS permits a radically different process that seems to enable faster development, but people still want things faster still. So how do you do it? Other trades are possible; you could sacrifice quality (but no one wants that) or reduce features (many users don't accept that). The obvious thing to do is to use approaches that save development time (use higher-level languages, pre-integrated massive libraries, etc.) at the expense of memory use.

That doesn't mean you should ignore memory use. Indeed, any general solutions that reduce memory use significantly, without vastly slowing down all development, should definitely be employed. If people will track down the primary "memory hogs", and fix them, the world would be a better place. But let's be clear about the reasons for the memory use. It's not that people are stupid; smart people are making conscious decisions that, at least at the time, other issues were more important.