[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]

Re: Request for Some Enlightenment re Ted's Position on Files/Directories



On Tue, Jan 14, 2003 at 12:33:30AM -0600, Jeff Rush wrote:
> I realize that Ted is quite busy, and that he gets LOTS of silly 
> questions all the time, but could he spare a moment to set me straight 
> on one Xanadu issue that I've pondered over the past few years until I 
> get headaches! ;-)  I just can't see how to operate w/o files/directories.

Since Ted is traveling at the moment, I'll take a stab at answering.

> > Hierarchies and files must not be part of the mental structure
> > of documents.
> 
> I understand why they're BAD, but I don't understand how to do without 
> them.  I've watched many non-technical people struggle with filenames, 
> paths and drives.  But hierarchy is SUCH a pervasive concept within 
> information management circles; rows, shelves, volumes, et. al.
> 
> [1] if documents don't have (file)names, how do I return to them later? 
> How do I tell friends to go read them?  How do I put references to 
> them in magazines and books?  How do I advertise them on TV and billboards?

Ted never said that documents shouldn't have names.  On the contrary, the
very definition of a "document" in Xanadu is any collection of information
to which you have attached a name.  At any time, you can take any of
the information in the entire Xanadu docuverse, collect and arrange
it in any way, and by giving it a title declare it as a new document.
Titles need only be unique to you.  To tell someone else how to retrieve
your document (once you have published it by notifying the Xanadu system
that you wish it to be publically available) you only need to give them
the publisher ID and title.

> [2] what replaces the web's URL space? With the (admittedly 
> hierarchical) URL, I can print out the reference; I can read it to a 
> friend over the phone; I can manually modify it and navigate around the 
> server's structure in order to find other things/repair lost things.

URNs.  Universal Resource Names, which TBL has sheepishly admitted he
very unfortunately didn't start with in the first place.  You can only
navigate around URLs when you have outside knowledge of the filesystem
structure on the server - for example, Unix servers are case-significant
while Windows ones are not.  Database-driven servers may use arbitrary
object IDs.  While you can sometimes do this with some degree of success
today, you can't rely on it and it is a hack to get around a lack of
navigational facilities in the existing web.

> *Within* a (Xanadu) browser, you could eliminate/hide URLish refs, but 
> how does Xanadu deal with external, non-browser handoffs of document 
> references, without using SOME form of hierarchy?  Certainly the URL 
> hierarchy doesn't have to correspond in any way to the storage 
> mechanism, but still it seems a useful concept, but then I'm a techie 
> born and bred in file hierarchies.

I'm not sure I follow the question.

> In your _Literary Machines_ book, tumblers are used, but those seem to 
> be a form of filename albeit a string of digits.  They are like ISBN 
> numbers for a publisher, and hard to read over the phone or scribble 
> (reliably) on paper.  No one is ever going to put them in TV ads.

Tumbler addresses are to Xanadu document names as IP addresses are to
domain names.  They are only intended for internal use, but of course
the more technically inclined can use them directly if necessary.

> Also the original vision for tumblers of server.owner.document.position 
> would seem to be a hierarchy itself; one which, as a developer on the 
> Udanax-Green code, I've struggled to deal with and thought that it 
> imposed an unnecessary rigidity on (meta-)document organization.

It is difficult to see how else to guarantee global uniqueness and
preserve the transfinite arithmetic properties of tumblers.  Furthermore,
there is a logical containment relationship being expressed in this
design.  The Xanadu definition of a document is far more flexible than
most; can you describe the boundaries of the organisational rigidity
you mention?  What restrictions do you see?

Hope that helps,
		Andrew
-- 
mailto:xanni@xxxxxxxxxx                         Andrew Pam
http://www.xanadu.com.au/                       Chief Scientist, Xanadu
http://www.glasswings.com.au/                   Technology Manager, Glass Wings
http://www.sericyb.com.au/                      Manager, Serious Cybernetics
http://two-cents-worth.com/?105347&EG		Donate two cents to our work!
P.O. Box 477, Blackburn VIC 3130 Australia	Phone +61 421 775 358