Include fragments [was RE: [uf-discuss] a.include mimetype]

[Chris Messina]

Actually, there's a distinction to clarify, which Joe brought up:
cross-domain transclusion (which Tantek says isn't what we're talking
about) and local transclusion -- i.e. from the same referencing page.

But I actually think that the remote case is more interesting from a
standpoint of inheritance and authority. Another pattern worth looking
at (and mimicking) is cited blockquotes. Since I can quote a remote
webpage and then show the source in the semantics of the blockquote
tag, you can actually go and see whether what I've quoted is a)
accurate given the original context and b) whether the source changed
(thus reducing the authority of one side over the other).

In this conversation, it's useful to create a local instantiation of
the data (eg "Microsoft's offices") linked to the equivalent expanded
version (100 Market St, san francisco, ca). Therefore parsers can use
the local data (and, humans can read this "default" view) and in the
case of uncertainty, refer to the remote source.

Does this view make sense?

Chris

On 7/10/06, Joe Andrieu <joe at andrieu.net> wrote:
> On 7/10/06, Tantek Çelik <tantek at cs.stanford.edu> wrote:
>
> > <span class="location"><a href="#scandinaviah" class="include"
> > type="text/html-frag">Scandinavia House</a>
> >
> > I think this may be another legitimate use of "includes" - to replace
> > abbreviated local content with a more extensive chunk from another
> > location.
>
> This usage is more of an "include" compared to the current usage,
> which,arguably, is more like a "repeat" pattern, i.e., this information
> already exists elsewhere on this page and I don't want to repeat it, so go
> find it with this reference.
>
> "Including" references to fragments on other pages creates the problems
> Chris brought up: that the fragment, which may be from some other author and
> may be arbitrarily formed or invalid or impermanent.
>
> I have been dealing with this in some small way with the hCitation
> microformat, which does not include a mechanism for actually referencing a
> specific fragment within an online source. My current approach is to manage
> the excerpt myself, but use the hCitation for presenting the source of the
> excerpt. In this type of fragment reference, it is not possible to gaurantee
> that the fragment is well-formed or even wrapped in a single referenceable
> entity, e.g., within an identified DIV element.
>
> Have there been any other discussion or conclusions about how to reference
> arbitrary HTML fragments in any microformats? For example, how to delimited
> a section of content within a URI?  hCitation is what I have been thinking
> about, but it seems to potentially make sense for the include pattern as
> well, if we embrace the external reference usage.
>
> Both Xlink[1] and MS' clipboard format[2] resolve this conceptual problem,
> sorta, but neither help with the arbitrary external HTML reference.
>
> However, there are a different bunch of inherent problems for parsers and
> for impermanent references.  It might be cleaner to require that "include"
> in the current usage be sourced on the same page, and create another
> reference pattern, "external" (or something) that can incorporate approaches
> for the trust/validity/permanance issues.
>
> -j
>
> [1] http://www.w3.org/TR/xlink/
> [2]
> http://msdn.microsoft.com/workshop/networking/clipboard/htmlclipboard.asp
>
> --
> Joe Andrieu
> joe at andrieu.net
> +1 (805) 705-8651
>
>
> _______________________________________________
> microformats-discuss mailing list
> microformats-discuss at microformats.org
> http://microformats.org/mailman/listinfo/microformats-discuss
>


-- 
Chris Messina
Agent Provocateur, Citizen Agency &
  Open Source Ambassador-at-Large
Work  / citizenagency.com
Blog  / factoryjoe.com/blog
Cell  /  412 225-1051
Skype / factoryjoe
This email is:   [ ] bloggable    [X] ask first   [ ] private