[Date Prev] [Date Next] [Thread Prev] [Thread Next] Indexes: Main | Date | Thread | Author

RE: [ba-ohs-talk] :xu: Serious deep addressing (was Re: ![Gzz] Re: [ba-unrev-talk] Starting Point for Collab Tool



> Not just every paragraph, but every character, must be tracked.
>   When you're almost there, why stop?
>
> I would recommend another look at the tumbler mathematics by
>   Roger Gregory and Mark Miller, described in *Literary Machines*
>   (you can find an illegal copy on the Web someplace, scanned as
>   muiltipage fax=TIF=jfax).    (01)

I've not yet located the scan, but there is a page about this here :
http://udanax.com/green/febe/tumblers.html    (02)

The W3C specs have gone down a similar path with XPath and XPointer as ways
of addressing parts (down to individual characters) of an XML doc e.g.
#xpointer(id('boy-blue')/horn[1])element(boy-blue/3)    (03)

The biggest difference to my eyes would be that the W3C system is 'lumpy',
particularly as the documents will be addressed through reference to a
server - I don't think it is possible to do a 'smooth' relative reference
from one arbitrary point in a doc on one server to another point on another.
(I'm not entirely sure how you'd sequence the servers either - aardvark.au
first?)    (04)

I'm curious, has any work on this kind of system been done in the
compression/crypto worlds? I'm sure this is ground well trodden, but if you
had a doc starting with "abcdef..." then you can make words up as (start
letter, number of letters) pairs.    (05)

Cheers,
Danny.    (06)