[whatwg] "Content-Disposition" property for <a> tags

Eduard Pascual herenvardo at gmail.com
Mon Jun 6 10:27:34 PDT 2011


On Mon, Jun 6, 2011 at 6:59 PM, Dennis Joachimsthaler <dennis at efjot.de> wrote:
> Yes, I was trying to refer to the verbosity. There's no html attributes
> with dashes in them as far as I know, except for "data-", which are user-
> defined. This would kind of break the convention a little. I could think
> about having "contentdispo" or some shortname like this, it would fit
> "better" to what we currently have in html.

Maybe "disposition" could work? For the HTTP header, the "content"
part indeed refers to the content of the response; but on the case of
a link, the attribute would be referring to the linked resource,
rather than the actual content of the element. So it's more accurate,
we reduce verbosity, and we get rid of the dash, all of this without
having to make the name less explicit nor relying on an arbitrary
abbreviation (ie: why "dispo" and not "disp" or "dispos"? Since there
isn't a clear boundary, it could be harder to remember; but dropping
the "content-" part seems more straight-forward).

> Again, html convention: Currently html only has one "statement" in every
> attribute, except for things like events (which is javascript) and style
> (which is also ANOTHER language: css).

Well, meta elements with a http-equiv attribute normally have a full
HTTP header (including parameters if needed) in their content
attribute, so I see no issue in taking a similar approach. After all,
HTTP _is_ another language (or protocol, to be more precise, but
protocols are still a kind of languages).

> Seems cleaner to me if we stay to the "standard" and not change the syntax
> "rules".
HTTP is also a standard. So we could stick to it. It all boils on a
choice of which standard we honor above the other. Seeing that HTTP is
an actual standard, rather than a mere convention, and we are actually
"borrowing" a feature from it, it looks like the "winner" to me.

> Please tell me if I missed anything here!
>From the top of my head, @class is defined to be a space-separated
list of class names. Sure, it is a simpler syntax, but it's still a
multiple content attribute. I think there are some more cases, but I
can't recall any right now.

Regards,
Eduard Pascual



More information about the whatwg mailing list