W3C home > Mailing lists > Public > www-html@w3.org > September 2002

Re: XFrames - Resurrected Zombies? (was: Comments on HLink)

From: Ian Hickson <ian@hixie.ch>
Date: Sat, 28 Sep 2002 19:14:45 +0000 (GMT)
To: (wrong string) äper <christoph@paeper.de>
Cc: Steven Pemberton <Steven.Pemberton@cwi.nl>, "www-html@w3.org" <www-html@w3.org>, "www-html-editor@w3.org" <www-html-editor@w3.org>
Message-ID: <Pine.LNX.4.21.0209281837210.15733-100000@dhalsim.dreamhost.com>

On Fri, 27 Sep 2002, Christoph [x-unknown] Päper wrote:
> 
> If you've got a quite simple frameset, navigation (fixed included via source
> attribute), subnavigation and content, it'll look like this:
> 
> (1) http://example.org/#frames(subnav=foo-2,content=2/bar-4)
> 
> The "next topic" link in 2/bar-4 would look like this:
> 
> (2) <a href="bar-5">next topic</a> or
> (3) <a href="bar-5" target="content">next topic</a> or
> (4) <a href="/#frames(content=2/bar-5)">next topic</a>

Not (4), that would end up nesting a new frameset inside the parent one
(unless I've missed something). To do 4 you'd need to target="" the top
level (using target="_top" or whatever the modern equivalent is).


> What'll my address bar look like if I cicked (4)?
> 
> (5) http://example.org/#frames(content=2/bar-5) or
> (6) http://example.org/#frames(subnav=foo-2,content=2/bar-5) or
> (7) http://example.org/#frames(nav=nav,subnav=foo-2,content=2/bar-5)

Assuming I am right, it would look like

   http://example.org/#frames(subnav=foo-2,content=#frames%28content=2/bar-5%29)

If your assumption (that it replaced the entire frameset) was correct, it
would look like (5):

   http://example.org/#frames(content=2/bar-5)

...as far as I can tell, and subnav will be blank. (The nav frame will
still be populated from the source attribute.)


> Isn't the use of the sharp sign, traditionally used to precede a
> fragment identifier (IDs), a misuse?

No, the string following the # sign is media-type specific according to
RFC2396. It is in fact a very correct use. Of course I would say that,
since I was one of the people who suggested its use back in February...


> When do which characters (comma!) in uri* have to be encoded?

Any commas, parentheses, and percentage signs in the URIs have to be
escaped, otherwise they could be misinterpreted as parts of the parent
XFrames URI. Note that you can't rely on nested brackets being
balanced, because the following are possible URIs:

   http://www.example.org/foo#bar,nav=baz
   http://www.example.org/foo#frames(a=
   http://www.example.org/foo#frames(a=%29b=)

...and you wouldn't want to end up with the xframes processor getting its
levels confused. So whenever inserting a URI into an XFrames URI, you MUST
escape commas, brackets, and percentage signs (specifically, percentage
signs that are escaping brackets, commas, or other percentage signs
escaping same).

The spec doesn't mention any of this, it probably should.


> http://www.example.org/~user/#frames(navigation=en/nav.xhtml,
> head=images/logo_big.png,
> banner=http://ads.singleclick.net/ads/so,me,weir,d123.str/ing4567?as,
> content=/en/category/sub/topic42.xfm#frames(footnotes=topic42-fn.xhtml,
> text=topic42.php#heading17?highlight=keyword))

That would have to be written as:

   http://www.example.org/~user/#frames(navigation=en/nav.xhtml,head=im
   ages/logo_big.png,banner=http://ads.singleclick.net/ads/so%2Cme%2Cwe
   ir%2Cd123.str/ing4567?as,content=/en/category/sub/topic42.xfm#frames
   %28footnotes=topic42-fn.xhtml%2Ctext=topic42.php#heading17?highlight
   =keyword%29)

That is, the nested commas and brackets are escaped. (There are no
doubly nested characters here so no %25s in this example.)


> I won't write something like that (at least navigation (if standard
> language), head and maybe banner would only be in source
> attributes), but there are people who want to and will
> (automatically generated of course). I really don't want to see that
> in my address bar, my bookmarks or my search engine results.

You see that now anyway on any site with session IDs etc. I don't
think it is that big a deal.


> I don't see the point where this whole thing is better for search
> engines. If they wanted, they could have followed the values of the
> src attributes of the frame elements already as if they were hrefs
> of as. Some may do so, dunno.

I agree with this... I don't see how XFrames particularly help with
search engines really. It does make it easier for them to return
results to framesets, but that seems to be all.


> Btw.: What's the supposed file extension for framesets xml, xfm or
> frm? All of these are used in the draft. And what'll be the MIME
> type?

There is no such thing as a file extension on the Web, only MIME
types. Hopefully the MIME type for XForms will be text/xml or
application/xml.

-- 
Ian Hickson                                      )\._.,--....,'``.    fL
"meow"                                          /,   _.. \   _\  ;`._ ,.
http://index.hixie.ch/                         `._.-(,_..'--(,_..'`-.;.'
Received on Saturday, 28 September 2002 15:14:54 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:15:52 GMT