[sclug] Re: sclug digest, Vol 1 #148 - 6 msgs

[Neil Haughton]

>
>
>Patents are "open" in the sense of being available for public
>scrutiny and peer-review.  That's not the reason why they're "open" - it's
>more to do with placing limits on the time that an inventor can exploit
>her or his invention, and ensuring that the inventors secrets don't die
>with them, but it seems to me that in theory at least, that particular
>purpose and public interest should be well served by the system.
>(though in practice, I expect the wording of patents tends to be designed to
>give as little away as possible and generally make life difficult for anyone
>trying to copy the invention, and consequently difficult to review).
>
>  
>

I'm not a patent expert, but I have some experiece of patents in 
traditional engineering design. Patents don't dislose the exact details 
of how something is made, just the principles that the inventor wants 
protected. In fact they tell you nothing about how well something is 
being designed or made, and this will surely apply to software patents 
too; the patent itself need open nothing of the implementation to public 
scrutiny.  That a particular code technique or invention is patented 
will have no effect on the safety or otherwise of the implementation.

To take a practical and topical example (I'm using Mozilla Mail here), 
the 'inventor' who thought of applying Baynes' Theorem to the filtering 
of junk mail might patent the idea (it probably is patentable in the 
US), but that would not prevent him (or invited colleagues) from writing 
code that was full of loopholes, or applying the theorem poorly or 
inefficiently. Similarly the application of some mathematical idea to 
the control of an engine speed might be patented, but that would not 
prevent it being coded in a buggy way which resulted in the occasional 
'inexplicable' crash - it would merely prevent another organisation from 
using the same mathematics for the same purpose, possibly more safely.

Regards,

Neil Haughton