Why doesn't search spin off TWO threads, one to search the local computer, and the other to search online, if the user has that feature enabled?
This way both threads can work and local results can be returned while the other thread tries to search online. This has two WONDERFUL benefits:
- You will get local results faster, regardless of the state of online search, - Local search will work period, even if you're not online.
This is... common sense shit? Computers can multitask; there's no reason both searches can't happen simultaneously. Sometimes I wonder what they're smoking at Microsoft; I've been wondering this on and off for 25+ years now.
Fail gracefully is something should always be part of the design of any web based service.
Who said that first? Was it the Microsoft team that designed this [wikipedia.org] or the Apple team that designed this [wikipedia.org] or this [wikipedia.org]?:-)
In a semi-related true story, I once got a fatal error message from Tcsh that said, "Assertion Botch. This can't happen!" as the shell failed. Man, talk about a paradox.
"the design of any web based service" is compared to a non recoverable kernel panic? I really don't understand. If web search is enabled, it returned no results, not a kernel panic. At least it should have been in the summary if true.
I'll go. The failure mode for a certificate was considered permanent failure and never tested. It was never considered by anyone part of this team. Not part of the design, not part of the architecture.
Microsoft consistently fails at code reuse, bubbling up one error as something else and likely inadvertently hiding the source of problems. If you find the right debugging tool, you can usually see the problem.
And ultimately, it's the choice to hide potentially complicated messages to prevent confusion. A sear
Federal grants are offered for... research into the recreation
potential of interplanetary space travel for the culturally disadvantaged.
Can someone explain... (Score:3)
Why doesn't search spin off TWO threads, one to search the local computer, and the other to search online, if the user has that feature enabled?
This way both threads can work and local results can be returned while the other thread tries to search online. This has two WONDERFUL benefits:
- You will get local results faster, regardless of the state of online search,
- Local search will work period, even if you're not online.
This is... common sense shit? Computers can multitask; there's no reason both searches can't happen simultaneously. Sometimes I wonder what they're smoking at Microsoft; I've been wondering this on and off for 25+ years now.
Re: (Score:2)
Fail gracefully is something should always be part of the design of any web based service.
Re:Can someone explain... (Score:4, Insightful)
Windows being graceful only amounts to Ballmer wearing a tutu to work again.
Re: (Score:2)
Fail gracefully is something should always be part of the design of any web based service.
Who said that first? Was it the Microsoft team that designed this [wikipedia.org] or the Apple team that designed this [wikipedia.org] or this [wikipedia.org]? :-)
In a semi-related true story, I once got a fatal error message from Tcsh that said, "Assertion Botch. This can't happen!" as the shell failed. Man, talk about a paradox.
Re: (Score:2)
"the design of any web based service" is compared to a non recoverable kernel panic? I really don't understand. If web search is enabled, it returned no results, not a kernel panic. At least it should have been in the summary if true.
Re: (Score:2)
If web search is enabled, it returned no results, not a kernel panic.
It still should have returned local results.
Re: (Score:3)
You could have left it at:
Fail gracefully is something should always be part of the design.
Re: (Score:3)
Why doesn't search spin off TWO threads, one to search the local computer, and the other to search online, if the user has that feature enabled?
Because without service outages, unpaid interns would never get any breaks.
Re: (Score:2)
Do you really expect them to be able to write something multi-threaded properly?
Re: (Score:2)
I'll go. The failure mode for a certificate was considered permanent failure and never tested. It was never considered by anyone part of this team. Not part of the design, not part of the architecture.
Microsoft consistently fails at code reuse, bubbling up one error as something else and likely inadvertently hiding the source of problems. If you find the right debugging tool, you can usually see the problem.
And ultimately, it's the choice to hide potentially complicated messages to prevent confusion. A sear