I give Turpin the $M roadmap, was re: ***F*** removable media.

jbone at place.org jbone at place.org
Sun Oct 26 22:29:46 PST 2003

For Turpin:

To your first point:  FreeNet.  Now, granted, you might not regard 
FreeNet as a total solution due to lack of market share, so to speak.  
But regardless --- the tech is there, waiting for users.  To the second 
point:  FreeNet.  Cf. above.  To the third point (a) you've got it, 
80/20 and (b) the long-term solution requires Deepfile-2.  (That's 
Deepfile minus 2, the things we argued and designed in the conf. room 
over the Baskin Robbins over three years ago --- problem being a 
go-to-market problem / chicken-or-egg problem.) Cf. OceanStore, etc.  
And numerous commercial plays, mostly science projects that despite 
significant VC funding have blown up and will continue to blow up.  
Rich unstructured-content metadata is slowly worming its way into the 
enterprise.  Unf., ultimately only Microsoft can deliver it as a 
platform.  (No more Netscape debacles in the filesystem arena, there 
are some savvy M$ researchers --- starving for smart product managers.)

Bottom line is this:  the forces are in play.  It's (i.e., the ultimate 
metadata / KM / platform capability) an inevitable consequence of the 
technological forces that have come / interacted before.  It is --- 
like any good idea (e.g. Internet scale notification, presence, instant 
messaging, etc.) just a matter of time.  The key is whether the people 
that understand and can solve this problem can be --- and continue to 
be --- financially incented to do so while the market catches up.  
Classic innovators dilemma.

If I were you, Turpin, I'd spend every waking hour working on "Peabody" 
(tm --- Turpin) as an end-user tool --- in Python, leveraging RDF, with 
emphasis on releasing to Windows, OS X, and Gnome in that order with 
close attention to integrating under the hood w/ Friedman's Dashboard / 
cluestreams.  Not as an endpoint, but as a functionality-multiplier, 
regardless of platform.  You can't make $B on it, but you could easily 
retire through licensing the efficient algorithmic engine --- or if 
more eleemosynary, become the Linus of KM.  If you wind the narrow way 
through representation, retrieval, composition, and --- most important 
--- coordination properly, it's within reach.  Has been for several 

That's likelier than anything to get us further down the path w/o the 
hobbles of commercial / investment / profit motives.  And a good 
showing there early could make serious jack --- or at least move the 
bar way, way up.  How many hours a day are you spending hacking 
Peabody(tm)? ;-)

Don't worry so much about the storage metaphor.  Abstraction is your 
friend --- find the sweet spot, hint it's namespaces, leverage REST and 
RDF.  Yeah, today is *FUNDAMENTALLY* broken.  Work around it.  
Solutions will come --- but not until there's a profit incentive or 
somebody like you w/ the wherewithal to avoid the online committee 
debacle and just sit down and *DO* an end-user implementation.


BTW --- RT --- you looking at the Documents / Tables / etc. unification 
stuff coming out of Meijer?  Smart kid.  Read his stuff.

More information about the FoRK mailing list