Re: Figuring out easier readonly interfaces

On Mon, Oct 14, 2013 at 9:05 AM, Anne van Kesteren <annevk@annevk.nl> wrote:

> On Mon, Oct 14, 2013 at 4:59 PM, Mark S. Miller <erights@google.com>
> wrote:
> > On Oct 14, 2013 6:58 AM, "Anne van Kesteren" <annevk@annevk.nl> wrote:
> >> We have the same notion elsewhere though. E.g. a <input readonly>
> >> cannot be modified by the user, but can be modified through script.
> >
> > In this case, the "only" still describes a meaningful restriction, where
> the
> > restricted client is the user.
>
> That is the same notion Robbert is using. The user cannot modify the
> object, but the user agent can.
>

The user agent is the browser, which can clearly modify any object it
implements. Are we talking about user vs script, as in your <input
readonly> example? Can the user modify a DOMRect object? If yes, and a user
cannot modify a DOMRectReadOnly object, how can a DOMRect be a kind of
DOMRectReadOnly?

I thought the question was: What do we name the supertype that a read-write
Foo, a read-only Foo, and an immutable Foo would have in common? Clearly,
this is one of the questions in this thread, as it also relates to future
extensions to Maps, Sets, WeakMaps, etc... whose clients are scripts but
not users. Other possible namings:

BaseFoo, AbstractFoo, CommonFoo,
FooBase, FooAbstract, FooCommon

but I like the idea of some variation on *read* or *query* to explain what
these have in common.


>
>
> >> Attributes annotated in IDL with "readonly" have the same behavior.
> >
> > Could you provide a link to the relevant part of the webidl spec, and to
> an
> > example of a webidl that makes use of this ambiguity? Thanks.
>
> http://dev.w3.org/2006/webapi/WebIDL/#dfn-read-only


The text there is self contradictory. First, it says
"An object that implements the interface on which a read only attribute is
defined will not allow assignment to that attribute." which supports
only-means-only.

But then it says: "This[inherit] can be used to make a read only attribute
in an ancestor interface be writable on a derived interface."

They can't both be true.



>
> http://xhr.spec.whatwg.org/#xmlhttprequest (see any readonly attribute
> defined here)
>

Looked. Where are these overridden to make them read-write?



>
>
> --
> http://annevankesteren.nl/
>



-- 
    Cheers,
    --MarkM