Tangled in the ThreadsJon Udell, May 10, 2000
Visualizing data, and scratching the digital pix itch
This week I visited a new site, WebBrain.com, which uses an information viewer called TheBrain to visualize the Open Directory category tree.
The same technology for visualizing associations is in a downloadable version of TheBrain, which you can use to create and publish concept maps. An effective use of this can be seen in Carl Malamud's Mappa Mundi. See for example http://mappa.mundi.net/Internet-Drafts/, and follow the SiteBrain link.
This is very nifty stuff. Call me a stick-in-the-mud, though, but I can't help thinking that the real challenge is data entry and categorization. Give me a dataset that somebody has neatly tagged by topic, using a nicely-controlled topic vocabulary, and I'll spin out lots of nice ways to visualize that info. But who's going to create that dataset, and how?
I haven't yet used TheBrain's concept-map-authoring capabilities enough to judge whether it's really more powerful than what can be achieved with disciplined use of bookmarks, or any other simple way of categorizing contents. To me, the real effort is just in doing the categorizing, which is intellectual work that is simply not automatable, though we'd all like to pretend that it is. Tools can't do that work. But they can help, so I put the question to the newsgroup: "Is TheBrain a useful knowledge categorizer?"
I've been using TheBrain for a year or so. It's been real assist to me as an organizer for links, files, ideas, and snippets. Where it beats a simple bookmarking system is the multiplicity of parents and siblings you can attach to a node.
It's the finest implementation of a mind mapping tool that I've run across, although there are some open source efforts underway that I hope will catch up.
I agree with you Jon, that categorization is the major hurdle for most users of such organizers. I think a multi-linking tool like TheBrain helps. You're not always fighting the one-parent requirement of a typical bookmarking system.
If you can home in on a node from more than one direction, you don't struggle so much with decisions like "argh, should this go under 'Jon Udell' or 'zope-xml'?" And it's easier to find things later.
Michael, and several other correspondents, noted that TheBrain has been awarded a broad patent for its dynamic GUI. Perhaps Apple should have gotten there first?
Jeffrey P Shell:
Is this different from Apple's HotSauce MCF [Meta Content Framework] of a few years ago? MCF, I believe, went on to become RDF. HotSauce gave you a 3-D fly-through of graphs of data. It was pretty neat. Back in the days of Apple Research Labs.
Indeed, HotSauce did seem resemble TheBrain, though its current home doesn't show evidence of much activity.
If you're interested in this sort of thing, you'll want to check out Carl Malamud's Mappa Mundi, an e-zine that's devoted to visualization of the Internet. Here's the index to the map-of-the-month list.
And if you haven't seen the Peacock Maps visualization of the whole Internet, you'll want to have a look. It's stunning.
image from http://www.peacockmaps.com
Finally, Dave Caplinger pointed us to the Atlas of Cyberspaces, yet another fascinating cyber-exploratorium.
Scoping out digital cameras
Seeing all these pictures reminded me that I've been meaning to hop on the digital camera bandwagon, so this week I've been doing some research. As always, the newsgroup was an invaluable source of information and feedback.
As I looked around, it seemed apparent that serial-port picture transfer was going to be a non-starter. Life's too short to wait for those pictures to transfer, right? Yet some really nice-looking cameras, such as the Nikon Coolpix 950, don't do USB.
One interesting solution I came up with is printer-based. The Kodak PhotoSmart P1100 supports both CompactFlash and SmartMedia cards. You can print directly from the cards, or you can use the printer as a USB cardreader for image transfer to your computer.
Jonathan Brickman:I did the serial-port thing for a client once. Worked very well. Astonishingly reliable and fast, given the two-conductor minimalist RS-232.
Most people using digicams heavily get CF readers for their machines (which may use USB). This means you don't have to drain camera batteries to do download, the CFs look just like disks, etc. You would especially want a CF reader if you had more than one CF card... it'd be a shame to take lots of pics on lots of CF cards, and be stuck swapping them all back into the camera to download them using some stupid interface.
I am just getting into digital photography and use a CF<->PCMCIA adapter with my laptop to read the cards. It is fast and easy. I can then fix up the images and print them when I want to.
Of course, you need to have a laptop computer to do this.
Not necessarily. I've heard of a few people who've gotten the PC Card slots for their desktops and done it that way. It makes sense, in a way. PC Cards are easy to install and hotswappable. I love the fact that my SCSI card is a PC Card, because I can take my laptop over to the scanner and plug it in without shutting down. And I just upgraded my HD with an Apricorn EZ-Gig, which makes your old drive into a PC Card accessible drive. Plugged it into my wife's NT machine and it just worked. PC Card is cool.
Robert White:I've got an Olympus with a 4 meg card and serial. Takes a couple of minutes to transfer. Then it takes me hours, or even a couple of days, clipping and sharpening and tossing the 40 or so originals. So it really depends on how you're using it. Most of mine end up as low-res HTML, so I'm CPU bound rather than IO bound.
Marcus Rimmer:I've just got a Sony DSC-F505 and while the camera is great the USB support only works on Win 98.
I am using an Olympus D-450 Zoom and am very pleased with the high res photo quality of 1280 x 960.
Olympus uses the thin SmartMedia memory cards. The camera comes with a handy floppy disk adapter. Transferring images is as simple as inserting the SmartMedia card into the "floppy disk" and insert that one into the floppy drive.
Lots of options! In the end, I went with the Kodak P1100; it'll be here shortly and we'll see how that goes.
Althought I've got lots of ideas for how I'll want to use a digital camera, Randy Switt suggested something really intriguing:
The Kodak DCS series is very flexible, and it even supports a programming language so you can write scripts for it. You can do things like turn it into a motion detector/burglar alarm (take a picture every x seconds, compare with last one, discard last one if no change above a certain amount).
We do some tracer studies in 2-D effect boxes, and using the scripting I can set it up to take photos in any series fashion I need. Typically the best setup for me is a logarithmic scale, with pictures every few seconds at first, gradually lengthening to every few hours.
And here I'd been thinking the camera would get me off the computer and out into the real world of trees and ponds and bugs -- the real kind, not the software kind. Hmm. Maybe, on second thought, a scriptable camera isn't such a great idea after all...
This work is licensed under a Creative Commons License.