decoupling from HTML
brad at danga.com
Tue Jun 28 11:35:25 PDT 2005
We'd previously discussed a "magic" like that (cookies.txt, favicon.ico,
etc) but decided against it.
Plus we're not going to let the OpenID core creep into public profile
exchange. Private profile data might be an extension, outside the
core. But public stuff, absolutely not.
On Tue, 28 Jun 2005, M. David Peterson wrote:
> To help maybe answer my own question is there any reason why, if the link
> tag is not found as previously defined to then check the same directory
> specified for an openid.xml file which would be a minimal atom feed such as:
> <?xml version="1.0" encoding="utf-8"?>
> <feed version="0.3" xmlns="http://purl.org/atom/ns#" xmlns:dc="
> <title><![CDATA[<XSLT:Blog />]]></title>
> <link rel="openid.server" href="http://xsltblog.com/openid-server.app"/>
> <tagline>An ongoing weblog of current topics from the XSLT development
> community & other
> XML/XSLT related news items. Hosted, maintained, & edited by M. David
> <id>tag:www.xsltblog.com <http://www.xsltblog.com>,2005://1</id>
> <generator url="http://www.xameleon.org/" version="0.1">Xameleon</generator>
> <copyright>Copyright (c) 2005, m.david</copyright>
> This would allow for the existing spec to work as is with the added
> insurance policy to allow a request for a specific atom xml file to override
> any potential web server issues that serve up other page formats based on
> the client.
> This would also add bits of other information pertaining who the user is,
> etc... possibly for a richer user experience but thats obviously beyond the
> scope of OpenID but within the scope of standards compliance which will
> allow for such information to be more easily consumed at the visiting site.
> On 6/28/05, M. David Peterson <xmlhacker at gmail.com> wrote:
> > Oooh... that is definitely something that should be reopened and evaluated
> > to determine a solution that covers not only HTML/XHTML but, as mentioned,
> > common data feed formats + SVG, XUL, XAML, and/or any other XML format which
> > in quite a short space of time will be seen more and more as the default
> > markup (and as such, namespace) for a variety of "pages" served up by
> > default when a particular framework (e.g. XAML) is known to be available
> > on the client. In fact when you download and install the latest
> > Indigo/Avalon WinFX beta release you are given the option to set IIS 6.0to render the XAML version of a particular page when the client is capable
> > and the server contains the proper xaml page in the directory requested by
> > this client.
> > I realize that the page that is contained at the specified location does
> > little more than act as a way to locate the proper validation information
> > and as such can be HTML/XHTML without effecting the rest of the applications
> > contained on that particular server. But there are enough justifiable
> > reasons (technical and marketing) for a web server to be allowed to only
> > serve, for example, RSS and Atom data feeds, that I believe this will become
> > fairly common place in a fairly short period of time. In fact services like
> > FeedBurner already allow the ability to host and serve up your data feeds
> > and as such the requirement to actually have a public site doesnt exist,
> > instead using a simple tool to create and post your entries directly to
> > FeedBurner for publication.
> > What then?
> > On 6/28/05, Brad Fitzpatrick <brad at danga.com> wrote:
> > >
> > >
> > > On Tue, 28 Jun 2005, Mario Salzer wrote:
> > >
> > > > If now HTML/XML parsers check for the presence of <html>,</head>
> > > > or <body> tags before actually reading out the two <link> tags,
> > > > they deprive themselves from supporting any other XML formats
> > > > which were especially designed with 'html compatibility' in mind.
> > >
> > > The reason the head checks were done in Net::OpenID::Consumer was to
> > > prevent people from hi-jacking other people's webpages by leaving
> > > comments/posts (which their software didn't strip) containing link tags.
> > >
> > > But even a regular expression to search for any link tag after we
> > > ascertain that the document isn't HTML (say, no <html> or <body> or
> > > <head>), that's still kinda lame. I suppose workable, though, if the
> > > regexp allows a namespace... but I'd in that case prefer a full-on XML
> > > parser so we can match on the /correct/ namespace.
> > >
> > > In practice, though, I imagine OpenID will be tied to HTML/XHTML, and I
> > > think that'll be fine.
> > >
> > > If you have a proposal though that doesn't add tons of complexity for a
> > > couple geeks doing XML+XSLT to make their homepage, I'm all ears.
> > >
> > > - Brad
> > >
> > --
> > <M:D/>
> > M. David Peterson
> > [ http://www.xsltblog.com/ ][ http://www.xmlblogs.net ]
> M. David Peterson
> [ http://www.xsltblog.com/ ][ http://www.xmlblogs.net ]
More information about the yadis