> On 18 jan 2015, at 18:00, Alan DeKok <[email protected]> wrote:
> 
> https://blogs.oracle.com/nico/entry/normalization_insensitivity_should_be_the
> 
>  They recommend storing the data / filename entered by the client as-is.  
> Then for comparisons, each client reads the filename, does normalization, and 
> compares the normalized strings.

Very english/us centric text when conclusions are drawn on how often issues 
might arrive.

That said, the thinking is not bad, but requires quite some thinking on various 
scenarios.

>  This is the same conclusion we got to in the RADEXT NAI document:
> 
> https://tools.ietf.org/html/draft-ietf-radext-nai-15
> 
>  Is there a reference which makes this recommendation clear?  It would be 
> good to help people create good solutions.
> 
>  If there isn’t a reference, should there be one?

There must be much more work behind something like this, like looking at the 
various normalization forms, when matching actually happens (risk for creation 
of two different files, when later only one of them can be accessed -- and it 
might even be a random one of them) etc.

   Patrik

Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail

_______________________________________________
precis mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/precis

Reply via email to