h a l f b a k e r yTastes richer, less filling.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
Tufte meets Gibson meets Xerox. Not "Send me email when this page changes", more "greet me in the morning with a rendered
virtual environment (casually representing page interconnections, ownership, size, traffic) in which things that changed recently are
rendered in slightly more garish colors."
Netomat
http://www.netomat.net/
As cited by [efader]. [egnor, Jun 08 1999]
bubbles
http://www.pobox.com/~jutta
see annotation. [jutta, Jun 08 1999]
An Atlas Of Cyberspaces
http://www.geog.ucl.../casa/martin/atlas/
Martin Dodge, Cyber-Geography Research (!), Center for Advanced Spatial Analysis, University College, London. Plenty of eye and brain candy. [jutta, Jun 08 1999]
Interpolation and Gamma Correction
http://www.fh-furtw...ch/gamma/gamma.html
Why hue blending usually works so badly, even in tools (like Photoshop) that are otherwise quite good... [egnor, Jun 08 1999]
Mirror Worlds
http://www.oup-usa....sbn/019507906X.html
As cited by bristolz, who ought to learn to use the "link" command. [egnor, Jun 08 1999]
WebWader
http://jeanpaul.lef...er/screenshots.html
Doesn't use whizzy graph layout, but does attempt to offer a "site overview", designed mostly for Webmasters. [egnor, Jun 08 1999]
map.net
http://maps.map.net/index.html
The latest attempt. More money than usual, but the visualization is disappointing. [jutta, Jun 08 1999]
A weird example.
http://desktopcnc.com/
I ran into this page a while ago (you can guess what I was looking for) and the navigation system in use reminded me of this idea. It seems terrible at first glance, but after a while it seems merely weird. Status bar (ab)use is key to the experience. [egnor, Jun 08 1999]
Color vision
http://www.handprin.../HP/WCL/color1.html
Do you think you understand how color works? You don't. [egnor, Jun 08 1999]
[link]
|
| |
there is a somewhat similar application called netomat. do a netsearch for it and you should find it..
it is built in java and very abstract.. enjpy |
|
| |
This idea is a little vague. There has been research on alternative ways to visualize the Web, but I'm not sure quite what you're after... |
|
| |
(Thanks for cleaning the spam.) |
|
| |
This is not a question of research - I don't really want anything technically new or hard. It's more a practical matter of finding out which combinations of existing techniques "work" as a user interface. |
|
| |
The weird red, blue, and yellow dots on my homepage are visualization renderings of the hypertext structure that the link points to. (Or pointed to about four years ago.) They show only data on that site. Darkness is age; radius is size; primary color is blue for an index, red for a list of outside links; yellow for a document that's mostly itself. Once you "understand" that language, you can tell a lot of things about the tree below with very little space. |
|
| |
I would like that degree of abstractness, combined with very small renderings of the pages (so I recognize them if I've been there before), combined with strong visual clues about things that changed, and all of it pre-loaded so I can browse familiar sites quickly without having to wait. |
|
| |
Oh, I always wondered what those were. |
|
| |
"Finding out which combinations of existing techniques 'work' as a user interface" still sounds like "research" to me, but that's just a word. |
|
| |
Is the red/yellow/blue color is assigned by threshold (ratio of interior links to exterior links to text content), or by fiat, or what? It seems like it would be interesting to vary the hue continuously as per the ratio, but perhaps that looks icky. |
|
| |
Would you imagine something like a hyperbolic tree layout? It's hard to imagine fitting even a small site into a single static screen legibly. |
|
| |
Colors: By threshold. Of course the thresholds are assigned by fiat... Yes, hue could be made gradual, but I didn't think it was worth it. Nothing I've ever done with anything near gradual hues ever worked (and I keep trying, to wit, the early halfbakery categories). |
|
| |
Hyperbolic trees: something like that, but it would of course have to be generalized towards non-tree structures. To get a feel for possible approaches,
spend a few hours in the linked-to Cyberspace Atlas. |
|
| |
Legible: As I lose legibility on one scale, I'll want to replace the lost information with something emergent. If one can't see what's on the page, one should still be able tell how big / recent / textual it is. If one can't see the individual pages anymore, one might still be able to tell whether it's a big flat site with lots of little database entries, or a relatively well balanced tree, or a long chain of predictable responses. And so on, all the way up. |
|
| |
If this is done, is fast, and still allows for gradual zoom (just like different features of the earth become apparent as one zooms out in "powers of 10"), one might be on to something. |
|
| |
Colors: Have you tried using gamma-corrected hue blending? You might find that the result is more attractive and comprehensible. (Or maybe you've already been there and done that.) See link. |
|
| |
Have you seen WebTV's "history" mechanism? It uses a screen full of thumbnails. Pretty cool, really, and (especially for the device) better than the list of page titles one usually sees. |
|
| |
Reminds me (a lot) of David Gelertner's book, Mirror Worlds. http://www.oup-usa.org/isbn/019507906X.html |
|
| |
Gelertner: I hate reading technological prophecies and doubt I'll be able to stomach this one, but I'll make an attempt, so thanks for the reference. |
|
| |
Colors: what I meant by it "not working" is not any technical rendering problem, but simply that when it comes to expressing a continuum, almost anything - light, saturation, structure, position - seem to be easier to process by a human viewer than changes in hue. |
|
| |
I didn't know webtv kept thumbnails; interesting approach. |
|
| |
One thing that is clear from looking at the different information landscapes is that some sort of auxiliary service would be needed that provides a meta-browser quickly with contextual information, without forcing the pages (or even their headers) themselves to be downloaded. |
|
| |
We were almost there before the Web took off (remember "archie" queries?); I wonder whether a site overview / index format will re-evolve (as an XML schema, no doubt), or is already being pushed somewhere; and I wonder what would be needed to actually propel that service into widespread distributed deployment. |
|
| |
Site overview: DAV (which is touted as "next-generation" HTTP by some) lets you query and retrieve "properties" for documents. They define some properties (basic concepts like "author", "last modified date", etc) and leave the door open for others (properties are stored and returned as XML). |
|
| |
So, that's a likely place for your "site overview" to start, though the currently defined properties don't include some of the things you'd probably like to see (link topology, text content, etc). Unfortunately, in the extreme, you'll probably want "properties" (such as the density of particular search terms of interest to you) that can only be generated by examining the original content...| — | egnor,
Oct 23 2000, last modified Nov 30 2000 |
|
|
| |
Fortunately, I never understood the whole "agent" and "search term" monitoring bit anyway and don't think it has much of a place in visualization; I'm quite happy to use more discrete tools for that. |
|
| |
(Alternatively, search engines could provide a mark-up of the visualized space as they do now, and could be integrated with the meta-browser. But that's really not something I can picture as more helpful than a simple set of weighted query results.) |
|
| |
The second one is of particular note. I think you might want to revisit the notion of hue-coding at some point, using the information in this page. (On the other hand, you'd have to be a serious color freak, like me, to make it through everything they have to say.) |
|
| |
Specifically, I'd like to see examples of your failed hue-coding experiments. I think I vaguely remember such a thing in the early Halfbakery, and I suspect that gamma correction may have been biting you (I remember that the intermediate colors seemed awfully "muddy"). |
|
| |
Anyway, the point is that human color vision is extremely quirky; if you can take advantage of the quirks, you might be able to devise a hue-coding scheme that people do in fact immediately "grok". |
|
| |