create new tag
, view all tags
Another consultant co-worker of mine pointed me to http://www.thebrain.com, which featured a really "neat" applet that essentially gave you some nice eye candy for a navigational aid. It was relatively free-form in the sense that it allowed you to organize links to information (looks like it uses Netscape's Open Directory Project behind the covers).

It raises an interesting discussion point (well, I think so :-). The problem with thebrain.com (and many others) is that it is focused on organizing existing information, but not on collaboration. This seems to be a problem with a lot of tools (duh, that's why we have TWiki :-). My point is that having some way to visually (or spatially) understand navigation is important in TWiki, but we don't really have anything like that. Naturally, because A Twiki web is "flat", the concept of hierarchy doesn't effectively exist, and yet each topic contains links to other topics and references to "parent" (ref-by). We are also finding that it would be enormously useful to have what we call a QuickiTwiki, some type of left-hand "favorites" navigation (hopefullly on a per-web basis).

As we continue to struggle in A-Large-Corporate-Phone-Company that we have brazenly foisted Twiki on, we find ourselves beset by the barbarians at the gate that would rather fight the future than work with it (ah, same old story :-). These people, commonly referred to as Middle Managers and Vice Presidents, seek to preserve The Way Of Things. They demand formal, hierarchical communication and navigation, and so we struggle with keeping TWiki in-place as an informal collaboration tool.

It seems like it would be a pretty easy step to create a navigational tool like that (one that organizes "parents" and "children" of a topic. Is this something that should be considered for TWiki (or more appropriately, as a TWiki.Plugin.WebHome)

-- DavidWeller - 08 Jun 2001

I've actually thought it shouldn't be too hard to create a navigation system similar to The Brain. The problem of course is that it would involve a lot of Ref-By searchs and probably be resource intensive.

-- NicholasLee - 08 Jun 2001

It might not be so hard if we keep a precompiled list of what topics point to any given topic.

On TopicAdd or TopicEdit of a topic A that points to topics B, C, D, spawn a background task to update the precompiled ref-by lists for B, C, and D. I wouldn't bother doing this for IntraWebLinks initially.

-- MartinCleaver - 09 Jun 2001

Oddly, it's intra-web links that interest us the most. Perhaps because we're so new with TWiki that so far only one project (web) is in TWiki.

Also, it's interesting that you mention precompiling. After watching TWiki in action, I'm quite puzzled why it doesn't generate a set of "cache" pages/topics that the web server can use. It's not like links necessarily change on-the-fly. Should this be another topic to discuss? What are the compelling reasons to regenerate a topic each time it's visited?

-- DavidWeller - 09 Jun 2001

You can see a working example of a Wiki navigation map, a bit like thebrain.com - it's discussed at Wiki:VisualTour and implemented using the VisualTour link at the bottom of every page. A bit slow but quite interesting.

(Q for PeterThoeny: Can we enable the InterWikis feature, please, to make such Wiki links easier? Frequently a feature we are discussing has already been implemented by another Wiki, and encouraging people to frequent other Wikis will result in them talking about TWiki there, improving TWiki's visibility.)

Some work that could be built upon to build a navigation map is at TWikiDrawImageMaps - this uses the TWikiDraw applet to generate the image, but there's no reason why this can't be done automatically - this has been discussed before on this web but I can't find the page.

As for the efficiency/caching issue - I think a dynamically built cache would add a lot of complexity, but it might be reasonable to build one overnight or every hour, perhaps.

-- RichardDonkin - 09 Jun 2001

WebMap would be the best to start with. AndreaSterbini uses the same graphing tools as used in Wiki:VisualTour

-- HubertWeikert - 09 Jun 2001

Regarding interwiki plugin: I will install it, stay tuned.

Improving navigation in TWiki is something higher on my priority list. At work we have several webs with 500 to a 1000 topics each. Better navigation within a web is crucial. Navigation has been discussed in NavigationByTopicContext, HierarchicalNavigation and, and as Hubert pointed out, WebMap.

-- PeterThoeny - 09 Jun 2001

When you get to that size it almost requires a human editor to structure the knowledge in the most efficent form. I image this Codev Web could do with better a clearer 'information clustering' interface.

Is anyone aware of a good text to semantic extraction type tool or pointers to discussions on this topic.

-- NicholasLee - 10 Jun 2001

I think a combination of better navigation tools, such as WebMap and HierarchicalNavigation (see ZWiki:FrontPage for one implementation), and more refactoring (discussed recently in TopicsThatDie) is definitely needed. Wikis do seem vulnerable to structure scaling problems - it's very easy to create new pages but quite hard to find them sometimes - I find Wiki:FrontPage quite hard to navigate for this reason, so it would be good to avoid this.

Other navigational ideas are:

  • Make it easy to categorise topics and search by categories (which TWiki already supports of course),
  • Encourage use of the Ref-By link, or a 'Related Pages' section that is used on every page (good when people use search engines). It's quite hard to get used to navigating with Ref-By, since most websites don't have this feature, only Wikis.
  • Easy finding of topic names when you know XYZ has been discussed can be fairly hard - a popup tool that helps you (a) search quickly for the topic name and (b) list all topic names, as per PopupPageIndexForEditing and IndexWhenEditing would be good. I think a tool that does both would be very useful, as long as it does not require JavaScript to work.
  • Encouraging people to search for an old discussion topic before creating a new one would be good, too - perhaps a bit of text to this effect in WebHome, under both features and bugs, with a search box right there, would help?

For more ideas on generic usability, please see UsabilityIdeas.

-- RichardDonkin - 10 Jun 2001

Before anyone gets too thrilled by The Brain's eyecandy, they thought so highly of it, they went and got themselves a patent (US# 6,031,537) for it in both the US and the UK. Frankly, I seem to recall educational films back in the 1960's when I was in high school that used that exact same metaphor, but, as is typical, the Patent Office has the memory of a titmouse. There are also several other things that predates the The Brain, among them a thing called "Mind Mapping" from (maybe as far back as) the 50's and some software products too. Too bad, but if anyone knows of any films or other media that clearly uses a very similar metaphor (the selected "thought" pivoting to become the central "thought" (The Brain patent terminology), please let me know and we can ammend The Brain's patent. AFAIK, that can be done without cost and lays a trap should the company ever try to enforce the patent. It used to be vanishingly rare that patents got tossed out, now it's an all too common occurance due to the abysmal research (or lack thereof) of the Patent Office.

On the general topic of information organization, collaboration and weblogs, I encourage any who have not visited to check out Memes.Net. Don't go planning a short visit wink One thing in particular worth noting (and stealing!) is the little combobox/button at the bottom which allows a user to select the page style they prefer (I think a cookie with that info also gets set).

-- DavidLeBlanc - 10 Jun 2001

Well, I believe conceptmapping predates mindmapping by quite some time, but uses a similar but less constraining model. It was invented in 1960's by Novak and Gowin. http://conceptmapping.net/ (powered by TWiki) is a site that I set up for people to talk about it. It has some links to every piece of conceptmapping and mindmapping software I know of, it hasn't really taken off in terms of contributions from people other than me.

-- MartinCleaver - 11 Jun 2001

One way that TheBrain manages the complexity is by making only links one or two levels away large. As the distance increases, the labels get smaller. Try limiting the graph to only one or two and see how that works.

-- JeffSparkes - 12 Jun 2001

That is one thing that actually really bugged me. As far as i've been able to figure out, The Brain will only show 2 levels out from the center at any given time. While it's good to limit focus at times, there are other times when the whole gestalt is wanted.

-- DavidLeBlanc - 13 Jun 2001

{meta} changed TopicClassification from FeatureBrainstorming to TWikiVsOtherProducts. What is the proper protocol for this kind of change? Eg. Should I ask somewhere else first? {/meta} It's all there to be used! - MM

-- MattWilkie - 04 Oct 2001

I've been doing a little digging here. The Brain comes in Personal and Enterprise (EKP) flavours, it's the EKP version I'm the most interested in. The website is a very slick and there is a very smooth bandwidth hogging flash movie available for the EKP version. Noticably absent from the website and PDF brochures are technical and implementation details; What you see is what you get: purchase information, testimonials, marketing copy, and a smidgen of tutorials and FAQs for the personal version.

I emailed them with a list of questions about costs, server and client requirements, file formats and data sources supported out of the box, methods for integrating pre-existing non-web data (eg. ms-word docs) and managing the relationships for same, and finally skillsets needed by the system administrators and what their typical tasks entail.

A representative replied with more marketing: Because of its scalability and that each client's information infrastructure is different, we like to engage clients into a very thorough sales process so that we can truly understand their requirements and criteria for deployment ... Until I get a better sense of your requirements and current information infrastructure, I won't be able to provide you a solid proposal. Please note that we have a very thorough sales process that provides you with ROI analysis, custom demos, and more, all with limited investment required

Which basically says to me that they dont want to commit to anything until they know what you want to hear and how much you will pay. Further dialog with them seems to supports this rather cynical assessment, but hey, maybe that's just me.

I did get one concrete answer:

   Basic QA approved setup is the following, but we can operate
   on other platforms and systems

      Client: Windows 98/NT/2000
           Internet Explorer 5.0 and above
           Netscape 4.7 and above

      Server: Windows 2000 Server
           IIS 5.0
           MS SQL Server 2000
           Allaire JRun 3.0

In a follow up 20 minute phone conversation with the sales representative quoted above I was able to get a little new information: some Brain EKP installations can be "in the six figure range" for cost and that licensing is determined on a combination of number of clients and "connectors". A connector is a conduit between The Brain EKP server and the data source. For example you can connect to a filesystem, a database (most flavours), a webserver, etc. It is designed for broad band but dialup access over 56k will be okay but 28.8k is iffy.

In a further hour long conversation with a different representative I was finally able to glean some hard{er} information:

  • a minimalist installation with one connector (to NT server file system) and full text search and indexing will cost $100k-$200k

  • a read-only client is $120, a r/w client is $250 (read/write refers to the ability to create/maintain links, relationships and notes about the documents. Modifying the documents themselves has nothing to do with The Brain)

  • there is no ready cut demo, they are all custom built per client

  • Relationships for non-web documents are manually built and maintained (eg. this is not a turn key solution for the hordes of documents stashed away over the last 10 years on your file server) though you could use the SDK and RDBMS backend to build your own autolinking/relating methods.

  • The required skill set for the administrators will lie somewhere in between those of an NT server admin and a full fledged Database Administrator.

  • the default and recommended RDBMS is SQL Server but any SQL compliant database should work.

So in the final analysis, or at least much of one as can be done with vague generalities and without actually seeing the product, what The Brain EKP has that TWiki doesn't is: feel free to flesh out the list

  • relational database backend (dunno enough about them to comment)
  • full text searching of all linked documents, nearly regardless of file format (TWiki with Namazu would go a long way to address this feature; SearchEngineVsGrepSearch , SearchAttachments )
  • well developed and refined graphical viewer of the relationships (the spinny words and rubber band links thing, which might be addressed with WebMap )
  • a well packaged interface (addressed with skins, or at least the ability is addressed. TigerSkin is cool but it has a few rough edges still)

When it comes right down to it, the only thing about The Brain which I'm really greedy for is the relationship viewer.

And finally there is something which is absolutely essential to any application which attempts to address this problem space and that doesn't come with either TWiki or The Brain: a MikeMannix who is willing and able to champion MasterRefactor -ing, and people willing to help.

-- MattWilkie - 07 Dec 2001

Well, Matt, thanks for all of your...unequivocal support. I hope my rant-to-refactor ratio doesn't get out of control and cause you to wonder... MasterRefactor went from a spontaneous (rather ambitious) idea, to a clear idea at this point. But, besides having limited time recently, I've been kind of stuck on the approach, not to entering thoughts, but to following through by tackling refactoring to support them.

An interesting find just now added a missing - well, obvious, but ignored - perspective for me on MasterRefactor. You may've read it already. It's a link from RichardDonkin in HowToGetInternalBuyInForTWiki, a very brief interview with Ward Cunningham called Tips for Site Operators - Launch Tips for the Wiki Apprentice.

Without getting into a long comment, beyond the obvious, it made me conscious of a basic skill and required task to really get a Wiki rolling on a full-on collaborative mode. Wiki is at least as much about Media and publishing, self-image and peer review, as it is a "business tool". c2 didn't get that way just by chance (check, say, slashdot, for comparison).

I will start to "refactor" topics this week, for the first time.

-- MikeMannix - 27 Dec 2001

This is a 6 years old discussion topic and I know that we now have the TouchGraphPlugin but I guess that this OSS found on sourceforge is worth the look hypergraph. It's a The Brain like navigator. From the official website:
_HyperGraph is an open source project which provides java code to work with hyperbolic geometry and especially with hyperbolic trees. It provides a very extensible api to visualize hyperbolic geometry, to handle graphs and to layout hyperbolic trees.

As soon as you want to look at large data volume that has a hierarchical structure, you will find hyperbolic trees very useful - they show more data than standard tree representations like your favorite explorer, and they have a great look and feel._

-- EricCharikane - 21 Sep 2007

Edit | Attach | Watch | Print version | History: r23 < r22 < r21 < r20 < r19 | Backlinks | Raw View | Raw edit | More topic actions
Topic revision: r23 - 2007-09-21 - EricCharikane
  • 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.