Re: Debian AMD64 - any 32bit compatibility?
On Wed, 2006-08-30 at 09:17 +0100, A J Stiles wrote:
[...]
> The exception to the rule-of-thumb is OpenOffice.org. Although it is Free
> Software, it is riddled with elementary mistakes, starting from the
> assumption that the processor uses a 32-bit word length and a 32-bit address
> space, which have severely impacted upon its portability even to other 32-bit
> architectures.
>
> This fiasco should serve as a strident wake-up call to the Free Software
> Community. Remember, OpenOffice.org began life as the closed-source Star
> Office. The egregious programming errors displayed in its source code were
> tolerated because nobody was checking up on them. How many other
I never looked into OOorg - if only that it is supposedly the hardest
step to actually compile that monster from scratch (which is a necessary
requirement for the first bug fix). And I fear that one cannot enable
most gcc warnings and -Werror - in my experience it is not that hard to
keep the source clean enough (and especially not in pure application
stuff like e.g. a word processor) to enable lots of gcc warnings - and
it actually warned upon bugs on 64bit hardware (e.g. comparisons of
pointer and ints which are usually bugs).
And there is enough other software out there which has enough dirty
programming to be cleaned up (or you leave it in similar to OOorg ....).
> closed-source products contain fundamental design mistakes that nobody knows
> about because the source is kept secret from the users?
"The industry" will see if/when the first Win* version with native[0]
64bit support comes out (and given the history of Win* will never loose
full and transparent 32bit support).
SCNR,
Bernd
[0]: What ever MSFT defines that to mean in their context.
--
Firmix Software GmbH http://www.firmix.at/
mobil: +43 664 4416156 fax: +43 1 7890849-55
Embedded Linux Development and Services
Reply to: