3 ms·
The thing is, DRY has some downsides. If you've pretty much nailed down the functionality of the old thing, especially if it involves some kind of manual testin
by oblio 6y ago
The thing is, DRY has some downsides. If you've pretty much nailed down the functionality of the old thing, especially if it involves some kind of manual testing/open beta test with users, the last thing you want to do is touch the old code.
I'm not defending them, but sometimes something is good enough and you have to move on.
- reader_mode 6y agoThis doesn't sound like an issue with DRY as much as an issue of monolithic design. If driver version X worked but X+1 doesn't on your driver - don't update the driver. But when the driver is in the kernel trunk then that option isn't that simple.
- atq2119 6y agoIn the context of the Linux kernel, you've never truly nailed down the functionality of the old thing. Interfaces external to the driver evolve over time (which bleeds into changes internal to the driver), requirements change, people want to add new features that old hardware ought to support -- all this adds up and puts evolutionary pressure on the drivers. Code duplication really hurts here. It's possible that the AMD developers are from a Windows culture where they don't have to worry about this because Microsoft gives them a stable interface. Though even there, I find it questionable whether it's a win. Surely the occasional bug will surface that affects drivers for older hardware as well?