dream hardware

castironpi at gmail.com castironpi at gmail.com
Wed Feb 13 02:11:59 CET 2008

On Feb 12, 4:51 pm, "Martin P. Hellwig" <x... at xs4all.nl> wrote:
> Bjoern Schliessmann wrote:
> > Jeff Schwab wrote:
> >> The only "dream hardware" I know of is the human brain.
> > Nah. Too few storage capacity, and too slow and error-prone at
> > simple calculations. The few special but very advanced features are
> > all hard-wired to custom hardware, it's a real nightmare
> > interfacing with it.
> > Regards,
> > Björn
> Yes and don't try to do a direct interface if you are not absolutely,
> certainly, 100% sure, that the understanding of the API is mutual.
> If not, be prepare to handle exceptions. Some you can ignore in a
> try/except clause like SyntaxError, MemoryError, RuntimeError and
> ReferenceError.
> Others should be more interpreted like a progress indication, for
> example: ArithmeticError which may be raised like a plain StandardError.
> But if you are lucky and get things more or less running watch out for
> LookupError, BaseException, EnvironmentError and NameError, does can
> ruin your day and may even be fatal.
> Absolutely fatal are: ValueError and TypeError, nothing really you can
> do against, even if you can catch these errors the program will usually
> still halt on a SystemExit.
> And if you think that is the worst can happen, he!
> There is still: AttributeError, ImportError, IOError and OverflowError.
> And it the end if all that didn't happen there is quite a chance either
> you or the other go for a NotImplementedError.
> If that didn't happen you always get after random() time an EOFError,
> however the good thing is that the process continues due to os.fork()
> (not available on all platforms).
> Don't count on return(0).
> --
> mph

Fortunately, import curses comes linked in the assembly.

More information about the Python-list mailing list