create new tag
, view all tags

Need ways of Accessing & Ascribing Anchors Better for Fine Grained Addressing

This is a fundamental issue that lies beneath all knowledge tools. I didn't understand the importance of what some people call "fine grained addressing" until studing it and learning from those that have studied it for much longer than I have. I hope I do this incredibly powerful idea justice with my description.

The current TWiki implementation suffers from a similar fundamental problem as do ALL web pages. While new HTML & HTTP mechanisms have come up to address this limitation, none have yet taken hold. The limitation is that the implementations have assumed a "knowledge unit" of a page or topic. This has certainly been one of the best and most clearly understood aspects of web technology. Without choosing exactly this knowledge unit, the web would not be what it is today. I believe this choice has enabled end-user understanding in ways that no other choice could have. The page/topic knowledge unit is clear, it's one big hunk loaded all at once to your computer via the browser. URLs are made up of server names and a hierarichal page name path. TWiki modifies this a bit in adding some "/cgi-bin/view/" stuff (see ShorterURLs) and a "TWiki web" name. The RCS tracks changes based on page/topic. TWiki follows the web convention. TWiki is very successful in important ways.

Anchors have come along to introduce the idea that you can also reference a particular place within an HTML page. These are appended to URLs with a # sign. However these are not assigned by default, they must be manually added. One project that has attempted to address this has been the development of a tool to parse a page and automatically add anchors and give the user little purple numbers so they are aware of these new anchors. I have proposed to integrate this tool with twiki via Plugins.PurplePluginDev.

All of these ideas have in mind the notion that an address is important. There are different ways of giving addresses, even for you own home you could say the third block from a particular corner which is a different address than your street address that may be entirely more useful in some circumstances.

The applicability of this idea came about during a discussion that LynnwoodBrown and I were having on LongTopicNav. My goal in describing this idea is to introduce it and give others an area where they can comment about it's applications to other discussion topics within Codev.

This is a very abstract idea, so I hope that the topic won't drift too far as people find applications of this idea in many different ways. I welcome links to these topics, but discussion of applications should be in other topics.

-- GrantBow - 25 May 2003

Well, if read you correctly, and I'm not sure I do, we are suffering from this because our awareness has already been raised simply by using TWiki - or any Wiki for that matter. We instantly have full text indexing and search capabilities and automatic hyperlinking. WikiPedia is the best example of this I can think of.

Lets look at a counterpoint. A Dictionary or Encyclopedia done 'the hard way'. The classic book 'The Hackers Dictionary' has been on-line for a long time. I have an early edition in hard copy, much thumbed and annotated. I've spent a lot of time following threads and references 'the hard way'. Its an obvious candidate for hyperlinking.

So look at http://www.mcs.kent.edu/docs/general/hackersdict/02Entries and tell me where the hyperlining is? Surely a tool to make the links "automatically" exists?

Things are better at http://www.eps.mcgill.ca/jargon/jargon.html#%3d%20A%20%3d but was this done automatically? Was the text parsed with a specific tool? Probably - or maybe the manual effort of a group of people.

By contrast, a wiki, or perhaps TWiki with some plugin to do a bit of automation and deal with non-WikiWords, would make any entry hyperlinked. Which gets back to Wikipedia.

Now in some ways this begs the question. The question of granularity that is. The hyperlinked hacker's dicionary example had the whole book contents in one page. Not very maintianable for a "live" system! But then its an image of a "dead tree" so it isn't a living, growing document as it Wikipedia or this TWiki. That's why a tool that parses the text and inserts links is sufficient.

So what's the problem?

First, we're spoilt. We're in a priviledged position because we know and can see that things can be better.

Secondly, we already have some compartmentalization with TWiki's "webs". Other Wikis are either compeltely flat or are are wiki farms with harder boundaries.

Thirdly, TWiki has great indexing and search capabilities. These are outstanding.

This gives a distorted view of things.

The underlying non-TWiki problems are somthing else, though. Grant is correct about granularity. However data is not information is not knowledge. Many of the mailing lists I'm on have lots of "me too" comments. Blah!

We're an early generation. We've seen things like Slashdot and still expect to see all the discussion in one place.

Automatic heirarchical Webs might help. Each new thread having its specifics - its "comments" - pushed into a unique sub-web.

So, perhaps more agressive use of the "comments" plug-in would help?

But as Grant says, its navigating long topics that is the present problem. I've looked at Purple. Cute, but I'm not sure it helps the underlying issue. In some ways it adds to it. I don't see the need to address every paragraph!

-- AntonAylward - 26 May 2003

See also Dr.Engelbart's Work for detailed Analysis of use/need

Anton, you are grasping the idea. The concept that I have tried to describe here is more abstract than most examples we might give. Fine Grained Addressing is an issue with any computer mediated communication (CMC) method: web, email, IM, etc. Different forms of CMC have addressed (pun intended) this issue in different ways. There are times such as for document editing that even a single change of a character could need to be "addressed" and discussed, but without a way to point to (address) particular characters in a way that computers understand then we humans continue to do the best we can with the tools we have. For research into this area, see Dr. Douglas Engelbart's research papers at http://www.bootstrap.org/.

-- GrantBow - 27 May 2003

Fine Grained Addressing Could Encourage Discussion

In BulletDiscussion there is an intriguing proposal that effectively puts "add comment" notes after every(?) paragraph. This raises the idea of annotations and feedback. This idea keeps on coming back as a useful necessity. Taking the comment of AreTopicsWebs, and TopicChildren, do you then go the next logical step and say that individual nodes have children and hence can have logical subwebs?

-- TWikiGuest - 21 Jun 2003

Edit | Attach | Watch | Print version | History: r10 < r9 < r8 < r7 < r6 | Backlinks | Raw View | Raw edit | More topic actions
Topic revision: r10 - 2004-05-20 - PeterThoeny
  • Learn about TWiki  
  • Download TWiki
This site is powered by the TWiki collaboration platform Powered by Perl Hosted by OICcam.com Ideas, requests, problems regarding TWiki? Send feedback. Ask community in the support forum.
Copyright © 1999-2018 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.