Thomas Hruska wrote:
> Ranju V wrote:
>> Yep.  Something is definitely off with Google today.  Even my harmless
>> little site is being flagged as being possibly harmful!  Grrr!
> 
> Some Google software engineer is likely injecting coffee and adrenaline 
> into his/her veins as I write this ;)
> 
> "This site _may_ be harmful" is, however, a grammatically-accurate 
> statement.  Maybe a computer program took it too literally and just 
> applied it to the Internet as a whole.  Humans and computers just 
> weren't meant to be together.
> 
> And now it looks like the problem is resolved.  The engineer can now go 
> back to sleep.

http://googleblog.blogspot.com/2009/01/this-site-may-harm-your-computer-on.html

In case anyone was interested.


A possible solution to this problem is a fairly simple design change:

Have a set of second-stage test servers to which updates get deployed 
_first_.  _Require_ all employees (and offer the option to external 
"beta testers") to use the second-stage test servers.  Automatically 
roll (on a rolling 6-hour cycle) an update out to the live site if it is 
okay on second-stage. The StopBadware.org updates would get pushed out 
to second-stage and only a few people know about it instead of the 
entire world, but they still know about it within minutes :P

In case anyone here will eventually be running a VERY large company any 
time in the future.


BTW, stuff like this happens ALL the time.  As a C/C++ developer, you 
WILL get yelled at when "it" breaks for not having thought of that "one 
something-or-other issue".  It is important to quickly fix the problem 
(without panicking and introducing new problems in the process), take 
the yelling in stride, and add the issue to your test suite/alter how 
you do things so it doesn't resurface in the future.

You will NEVER catch all issues because you see several trees instead of 
seeing the forest OR, if you see the forest, you will NEVER see each 
tree.  Catch-22.

In other words, this will happen again to Google as it will happen to 
each of us (not the same problem, but you know what I mean).  The best 
we can do is mitigate the problems by using such things as test suites, 
doing our best, getting 8 hours of sleep (less sleep = more mistakes!), 
peer reviews, and using tools that look for specific generic problems 
(e.g. out-of-bounds array accesses).

-- 
Thomas Hruska
CubicleSoft President
Ph: 517-803-4197

*NEW* MyTaskFocus 1.1
Get on task.  Stay on task.

http://www.CubicleSoft.com/MyTaskFocus/

Reply via email to