So what should one do with an assertion failure? You’ve detected a data structure is inconsistent. What do you do? Just carry on and hope nothing else important was corrupted?
Panic and write a crash dump so that people can notice and fix the problem. It’s difficult to imagine any alternative when the program is no longer in an anticipated state.
If you look at the quantity of code that goes in to the Linux kernel (or really the rate at which it goes in), it’s quite clear that no single individual could possibly review it all. The fact that he delegates review work to trusted subsystem maintainers is simply a matter of practical necessity.
This is less so admitting anything and more publicly blaming the other guy for his bad patch.
which to me read as criticism of Torvalds for failing to take personal responsibility for the bug. Do you expect him to actually do that for every flaw in a 20+MLOC kernel of which he’s written (or, as was my point, personally reviewed) a tiny fraction? And as for the “publicly” part – of course it’s public, it’s all public. If anything, it would strike me as kind of weird to suddenly take a piece of relevant discussion off-list (i.e. private).
I think the point is, regardless of who’s responsible for the bug, there are better ways to acknowledge a bug slipped out than publicly shaming the person who wrote it.
The success of Linux shows that’s wrong.
Maybe the issue is that people (you?) really need to stop forcing their own cultural expectations on persons from other cultures?
Linus’s handling of the mailing lists is deplorable. This isn’t a question of “cultural expectations"—he’s just an asshole. In my opinion, the extent of his assholishness is overstated, but the effect unfortunately probably is not.
Linux was successful because it was in the right place at the right time. BSD was plagued by legal issues, GNU was GNU, commercial unices only ran on staggeringly expensive hardware, DOS and (to an extent) Mac OS were very primitive, and most other competitors were already on the way out.
The success of Linux and Linus' personality are orthogonal. There’s no rule saying a person with bad people skills can’t run a successful project, and I wasn’t making that claim.
Also, I’m not sure culture has much to do with it. I don’t think all Finns are abrasive and abusive.
So what should one do with an assertion failure? You’ve detected a data structure is inconsistent. What do you do? Just carry on and hope nothing else important was corrupted?
Panic and write a crash dump so that people can notice and fix the problem. It’s difficult to imagine any alternative when the program is no longer in an anticipated state.
This is less so admitting anything and more publicly blaming the other guy for his bad patch. It’s a shame people still give Torvalds attention.
If you look at the quantity of code that goes in to the Linux kernel (or really the rate at which it goes in), it’s quite clear that no single individual could possibly review it all. The fact that he delegates review work to trusted subsystem maintainers is simply a matter of practical necessity.
I don’t really know what you’re responding to, but I agree with you nonetheless.
I was responding to
which to me read as criticism of Torvalds for failing to take personal responsibility for the bug. Do you expect him to actually do that for every flaw in a 20+MLOC kernel of which he’s written (or, as was my point, personally reviewed) a tiny fraction? And as for the “publicly” part – of course it’s public, it’s all public. If anything, it would strike me as kind of weird to suddenly take a piece of relevant discussion off-list (i.e. private).
I think the point is, regardless of who’s responsible for the bug, there are better ways to acknowledge a bug slipped out than publicly shaming the person who wrote it.
But Linus having poor people skills is old news.
The success of Linux shows that’s wrong. Maybe the issue is that people (you?) really need to stop forcing their own cultural expectations on persons from other cultures?
The success of Linux is much more about economics than technical merit.
Linus’s handling of the mailing lists is deplorable. This isn’t a question of “cultural expectations"—he’s just an asshole. In my opinion, the extent of his assholishness is overstated, but the effect unfortunately probably is not.
Linux was successful because it was in the right place at the right time. BSD was plagued by legal issues, GNU was GNU, commercial unices only ran on staggeringly expensive hardware, DOS and (to an extent) Mac OS were very primitive, and most other competitors were already on the way out.
The success of Linux and Linus' personality are orthogonal. There’s no rule saying a person with bad people skills can’t run a successful project, and I wasn’t making that claim.
Also, I’m not sure culture has much to do with it. I don’t think all Finns are abrasive and abusive.