Re: BIDI?

On Wed, Apr 27, 2011 at 1:38 AM, "Martin J. Dürst"
<duerst@it.aoyama.ac.jp>wrote:

> Hello Slim, others,
>
>
> On 2011/04/27 14:59, Slim Amamou wrote:
>
>> hello Mark,
>> - why not enforce the use of the scheme for uri detection? (in contexts
>> where the scheme is not obvious, like for example the URL bar of the
>> browser)
>>
>
> The scheme has very similar problems to the top-level domain: The list of
> schemes grows the same way the list of top-level domains grows. My guess is
> that the list is a bit shorter, and grows a bit less quickly, but it's not a
> big difference, at least for the moment.
>
>
>
There is a critical difference, though:  the bar to minting a new URI scheme
is very low, especially if you do not register the scheme (many do not), but
the bar to having a new TLD recognized is much higher (even if it is
dropping).  Despite that, I prefer to have the scheme, if present, be
fundamental to the algorithm--as there are some well-known schemes with
limitations to the permitted code-points beyond those acceptable in URIs
(e.g. URNs, which do not permit some characters other wise permitted).  I
suspect we may also see IRI schemes with similar limitations in the future.
 An internationalized version of URNs, for example, might still restrict the
set of delimiters even as it opened up new code points for Name-specific
strings.

regards,

Ted Hardie