[whatwg] Idea: Links w/o end anchors (is possible)
ppj at concept67.net
Wed Jul 22 03:38:36 PDT 2009
I've just joined the list, hopefully only to make a single suggestion. I
couldn't find this idea in the archives, so apologies if I'm way behind
where everyone else is going.
Brief background history:
A couple of years ago in 2006-7, a group I was with were interested in
linking without anchors at the end of traversal (in basic HTML). We
wanted to be able to link to any text in within a document. A friend,
Matt Schneider, creator of PurpleSlurple, had/passed on the idea that
only a short string of ~100 chars was required to uniquely identify any
document. Playing with it, I discovered that even shorter strings could
be used to uniquely identify places within a web page.
I created an overweight prototype proxy server, capable of approximate
match measurement, transclusion chunk separation... many features, zero
adoption rate, it seemed.
Unofficial spec and code (in Ruby) can be found on the internet archive
Unknown to me at the time, in India, someone I only know as
Nataranjan created LiveURLs (based on some work in Ahoy) and introduced
them in a Firefox plugin called WebMarker.
LiveURLs differed in that they use a checksum to encode the search text,
so it isn't open to direct user scrutiny but is more compact. Unlike my
XPunt prototype WebMarker was more lightweight and didn't support
approximate match indication etc.
Both these implemetations suffered from the issue that two users had to
have the relevant piece of software installed to share the links.
However, together those prototypes prove that the technology to do
linking without anchors exists, and is relatively straightforward.
Back to the present. The way around this shared software issue is simply
to have browsers include something like the following as basic
link traversal behaviour.
The Goal: Links w/o anchors.
The Strategy: Two stage process.
1) get an extra 'search' attribute on to the <a> tag in HTML so that we
e.g. <a href='...' search='...'>link text</a>
2) If there's take-up, then later on push for adding a date-time of creation attribute to <a>. This will add link history to the internet.
The way (1) works is someone sticks the basic href to a page in the href
attribute, and then sticks the text they want to link to in the search
attr. The browser fetches the page, and as a secondary action (at user
option) searches for the
text. The simple option is that it just searches for the plain string,
maybe later it can do all the fancy approximate match stuff that I put
in the XPunt prototype in '06/07.
Since we know those search strings don't have to be very long to find
the unique location, it shouldn't burden the document text very much.
Just thought it would be good to put these ideas to the list.
I'm hoping everyone just gets how much this could advance the world.
Peter P. Jones
More information about the whatwg