Re: [apps-discuss] draft-farrell-decade-ni-02 - we think this is done...

[Stephen Farrell <[e-mail]>]

On 04/05/2012 10:30 PM, Martin Thomson wrote:
> On 5 April 2012 14:25, Stephen Farrell<[e-mail]>  wrote:
>>> I might want to identify a resource outside of /.well-known/.  I might
>>> also want to identify a resource that is served elsewhere.  I can't do
>>> either with this scheme.
>>
>> Sure. We added that though in section 5. (For URLs anyway.)
>> Be glad to add more text that makes that clear, and happy
>> to say its usable with other schemes if you can offer good
>> words for that.
>
> That doesn't quite do it.  If I understand correctly, you want someone
> to look at a URI and "know" that this particular path component is
> actually a hash.
>
> What I want to do is provide an ni: URI that wraps:
>
> http://ajax.googleapis.com/ajax/libs/jqueryui/1.8.18/jquery-ui.min.js
>
> And have clients (say my browser) load that URI and be able to check
> that the content is the right content.  Without knowledge of URI
> components that should be opaque.  That is, I want a mechanism for
> dealing with mixed content (see plenary) that doesn't require TLS.
>
> This might be done by something like:
>
> ni:sha256:blah:http://ajax.googleapis.com/ajax/libs/jqueryui/1.8.18/jquery-ui.min.jse
>
> Though that completely ruins your nice URI form.

Well, it is less nice but following draft-hallambaker-decade-ni-params
we could add support for:

ni:///sha256;abc...?url=http://ajax...

I've probably got the encoding there wrong but would that meet the
need if properly done?

S.