[Info-vax] HP wins Oracle Itanium case

[Paul Sture]

On Tue, 21 Aug 2012 18:19:47 +0000, ChrisQ wrote:

> On 08/21/12 17:40, David Froble wrote:
> 
>> I've found debuggers to be quite useful. Many times they have pointed
>> out something that was right in front of my nose, and I didn't see it.
>> Great harm to the ego.
>>
>> But yeah, you really need to understand the code.
> 
> Most manufacturers embedded tool chains come with an ide and debugger
> these days and they are sometimes quite usefull for finding a subtle
> problems. However, there is a tendency in some quarters for the debugger
> to be used first to find a problem, rather than analysing where the
> problem is from program behaviour and reference to the source. Over
> reliance just takes away the instinctive ability, if that doesn't sound
> too pretentious. Keep the mind sharp, problem solving is what it's for
> and all that :-).

The one thing I miss with the demise of line printer fan fold paper is 
the ability to print out source code or compilation listings for 
debugging purposes.  With source code typically less than 80 characters, 
the white space at the RHS of the 132 character printout gave you plenty 
of room to pencil in notes and new bits of code.

One place I worked had a quiet room full of desks where you could retreat 
from the hubbub of chatter and phones, taking your listings with you.  I 
found this a really good way not just to concentrate on the problem in 
hand, but to avoid the impulse to make a minor change, recompile, make 
another minor change, recompile and so on.

> I start out with the maxim that code should should work right first
> time, but that requires up front effort. Obviously it doesn't always,
> but if the effort is put in first, then any bugs should be easier to
> find, with or without a debugger...

I still find printing source code a useful way to review code.

-- 
Paul Sture