|
Posted by Paul Bramscher on 11/07/68 11:53
I built a system a couple years ago and encountered these issues as
well. I began to see links in a web site as a directed graph
(http://en.wikipedia.org/wiki/Graph_theory). Breadcrumbs then become
semantically meaningful subgraphs. The meaning is the god's eye view
(webmaster's) and provided as a navigational aid to users. It's not at
all the same as what the user actually did to get from A to B (which may
have been a terribly convoluted route). It's also not meant to be a
comprehensive fully-connected graph (there may be many ways to reach
certain pages, but you select only the more important ones as breadcrumbs).
So "breadcrumb" is really a poor name for this thing. A cairn-marked
trail is probably more accurate.
I settled on a recursive method which allows the webmaster to fully
control the tree (move things around, add nodes, delete subtrees, etc.)
Because of this potential for movement, I decided also -- deliberately
-- not to encode the page's hierarchy anywhere in its URL. Best to make
the URL blind to the breadcrumb navigation in my experience. Give every
page a unique numerical identifier, I think, but make it flat.
JDS wrote:
> On Fri, 21 Jul 2006 09:57:33 -0700, Colin wrote:
>
>> What you are talking about I would call a directory or page heirarchy.
>
> See, now, I've always interpreted breadcrumbs to be a fixed directory
> hiearchy, and have nothing to do with the user's current browsing session.
>
> Home > Products > Gizmos > Rotating Sprockets
>
> Makes sense to me that that thing there is a representation of the
> website's hiearchy.
>
>
> Although I can understand the "user's current browsing session" type of
> breadcrumb implementation.
>
> Of course, the two styles of breadcrumb require quite different
> implementations.
>
> Also, one of the things I don't like about the "current session" style of
> breadcrumb is the logic problems: what happens when I browse from "Home"
> to "Gizmos" to "Right Handed Doohickeys" and then back to "Gizmos"? But
> just clicking on links, not using the breadcrumb itself. Does the
> breadcrumb look like this?:
>
> Home > Gizmos > Right Handed Doohickeys > Gizmos
>
> Huh? What then? Also, what if the person comes in directly to the Right
> Handed Doohickeys page from an outside page, say, a search? Does "RHD's"
> get the top billing in the breadcrumb?
>
> (Breadcrumb):
>
> Right Handed Doohickeys
>
> For me, those sorts of issues make it most sensible to stick with the
> "breadcrumb as website hiearchy" model.
>
> Also, on that note, the "breadcrumb" need not necessarily have anything to
> do with the "URL". When I have implemented breadcrumbs, it has always
> been with the use of a data flag -- in a database or in the page code --
> that indicates, "This page's parent page is XXXX". In the case of
> "Rotating Sprockets" or "Right Handed Doohickeys", the flag would be equal
> to "Gizmos" (or, rather, the unique identifier for the page "Gizmos").
>
> Using "parent page" identifiers, one can build the breadcrumb recursively
> all the way back up to "Home".
>
> later...
>
[Back to original message]
|