no trees (even more on what the heck is that thingamy?)

Left, right, right, left, right... right? Nope, I'm lost... lemme see... left, right left, left...

Ever "lost" a file only a week after having created it? It's in Documents > Clients > UK > nah, not there, click, click... in UK > Clients > Correspondence... not... let's try Spotlight... hmm, what was that file name again?

Pretty useless those tree structures. That's why the thingamy does away with them altogether.

And the alternative? Multiple tags filtering.

Tags everywhere representing anything adding up to a real knowledge-bearing-name-space.

Choose "UK GAAP, P&L, Expenses, Production cost" and see only the accounts for that. De-highlight UK and highlight US and see the corresponding selection now for US GAAP.

Invoke a report and choose "Contacts, US, Suppliers" and see Mike, Bob and Ted, de-highlight US and highlight France and immediately get Pierre, François and Michel on your screen.

In a task to choose co-worker set task to give operator / assignee all employees - select the tags "C++, German speaking, Beer drinker" and see the list shrink.

No hierarchy, tree structures, no folders. Efficient browsing of all data when navigating, reporting, working, selecting...

Why not go all the way and do an operating system with no file structure and no applications?

Next step, a bit later, just have to refine this stuff a bit first... :)

Comments

[NikoN]

I know this is not the point of your post, but you can actually "tag" files for Spotlight. Just open Get Info and write tags in the Spotlight comments field. Works for me... (Although even more than Spotlight I use the shareware app LaunchBar. With that I can go for days without ever having to browse for a file.)

[sig]

Yep, the world is starting to drift that way - and I'm glad, helps me ;)

But as you know, a loong way to go - like hierarchies, taxonomy...

And perhaps quite important: With tree structures we map from representor to the species, it be file, plant, web page, person... look at the "path" in terminal, in plant names, url and employee titles. Move the file, index.html or person or "reclassify" the plant and all links and businesscards are broken!

Map it the other way around I say (a bit like the accounts in thingamy as you've seen) ;)

[David]

Is this, then, like the difference between how, say, iTunes and Windows Media?

iTunes lets you select genre/artist/album and your library will kind of filter down based on your selections. WM forces you to go through the tree/search paradigms.

[sig]

Have to admit I do not know WM at all - long time since I dabbled in the Windows world :)

And not quite the way iTunes does it either - the closest example I can get to is our own early test here http://thingamy.com/thingwork/ which shows nothing when you "get there" - then try to highlight a tag, then another one, de-highlight and play around and you'll get the idea (Note: Too cool for MS IE though, try Firefox ;)

[David]

Yes... it's exactly like that difference. Of course it's completely extensible. e.g. not limited to one type of object (a tune) and three dimensions of tagging (genre, artist, album).

It would seem to me the real challenge with using this for enterprise data would be automatic tagging... e.g. if you have 2,000,000 invoice records and want to tag them with 5-10 different tags... that would have to be automatic. Is that where the SI comes in?

Anyway... cool stuff!

[sig]

You have a point there when volume gets big :)

Now, you do not always have to tag - there are other "place holders" as well including all the properties and metadata (who, when, where, costs, amounts etc) in a class (class being the cookie cutter to make objects - like invoice objects if they need to be objects!).

In addition we have a nifty behind-the-scenes "instruction" that can be used in a flow to manipulate an object, including adding or removing tags.

And of course - most important - we really have no clue yet what and where the practical obstacles starts to be too high - but we're ready for that too! When we hit such we'll tweak a bit, add a useful intsruction or other feature - then wait for the next obstacle.

We may try to think over all eventualities - but thinking we'd ever be able to see all, nah, better to be prepared to sort that out when we know - and spend time now to enable us to react fast and effciently when we "know" ;)

[flx]

I follow this site for a while now and while many of the ideas (like tags, or contexts - choose the name) have been around for a while, nobody in my knowledge has put them into one working system. I hope to see the results once in a while. And I will keep on reading here.

[Hamish]

Personally I use Copernic Desktop, and it "just works..." Concept is the same, tags generated by indexation.