[whatwg] data: scheme in form submission

[Olav Junker Kjær]

Ian Hickson wrote:
> Well, the definition in the spec does what any submission system needs to 
> do -- it gives the data to the target. 

Yes, that was the part I didn't understand :-)
The spec just describes how escape and embed the dataset in the URI, and 
then it says: "Then, access the resulting URI". But what does it mean to 
access a data URI?

> I'd rather not leave things undefined, because at the end of the day 
> people will just come and ask me what they should implement anyway, and 
> then I'll have to come up with something which will end up implemented but 
> not specified anywhere.

You are right, undefined is a bad idea.
However, by describing specific rules for how to handle data-URI, the 
spec gives the impression that this scheme is somehow a useful feature. 
If I understand you correctly, you are saying that the data-scheme is 
not really useful as an action target, and that its only described in 
the spec for completeness? In that case, there are many other schemes 
<http://www.iana.org/assignments/uri-schemes> that authors could 
theoretically use in an action-attribute.

Maybe the spec could just say that the form throws a 
URI_SCHEME_NOT_SUPPORTED exception when an unsupported scheme is used.

Olav Junker Kjær