29 |
document contents are named by cryptographical hashes [ref GUID paper]. |
document contents are named by cryptographical hashes [ref GUID paper]. |
30 |
Assuming an index of all documents and links |
Assuming an index of all documents and links |
31 |
on the local computer and all systems that can be reached through |
on the local computer and all systems that can be reached through |
32 |
the currently available network connections. Keeping such an |
the currently available network connections. Since no centralized |
33 |
index for the public Internet requires a decentralized, |
scheme can scale to such an index of "everything," keeping one |
34 |
|
for the public Internet requires a decentralized, |
35 |
self-organizing distributed system (the IPTPS'03 definition |
self-organizing distributed system (the IPTPS'03 definition |
36 |
of "peer-to-peer" [ref]). |
of "peer-to-peer" [ref]). |
37 |
|
|
38 |
|
We propose to use this same system for publication on the 'net |
39 |
|
and storage on a local computer. This means that |
40 |
|
when a local document is published or a public document |
41 |
|
is saved locally, it retains its identifier; as long as a |
42 |
|
document is also accessible through the network, keeping |
43 |
|
a local copy is merely an efficiency measure (keeping |
44 |
|
a permanent cache). Obviously, in such scheme, as long as link |
45 |
|
remains accessible, it never breaks just because the documents |
46 |
|
it refers to are moved to a different location. |
47 |
|
|
48 |
|
The Squirrel system [ref] is a peer-to-peer network of shared |
49 |
|
browser caches, where a web page can be retrieved from |
50 |
|
any computer in a local network that has a copy in its cache. |
51 |
|
Our distributed index for document and link retrieval would |
52 |
|
have this as a natural side-effect, except |
53 |
|
that it would also extend any data that has been permanently |
54 |
|
downloaded on one of the network's computers. |
55 |
|
|
56 |
|
A system as we propose would necessiate deep changes in applications. |
57 |
|
As Vitali [ref Versioning hypermedia] notes, any implementation |
58 |
|
of Xanalogical storage necessiates this, as "[n]o approximate, |
59 |
|
good-enough solutions" for the management of global identifiers |
60 |
|
"can conceivably be considered acceptable in this case." [#]_ |
61 |
|
As such, we are cannot meet the minimum definition of an |
62 |
|
open hypermedia system, as given by Davis et al [ref]: |
63 |
|
We do impose storage of markup on applications and we cannot |
64 |
|
generally use data created by tools that are not aware of our system. |
65 |
|
|
66 |
|
... |
67 |
|
|
68 |
|
|
69 |
|
|
70 |
|
.. [#] We have built a system for editing Xanalogical text |
71 |
|
in a non-aware editor such as ``emacs``, attempting to |
72 |
|
determine the user's changes through structure matching [ref]. |
73 |
|
This seems to work reasonably well for insertions, |
74 |
|
removals and rearrangements, but is hopeless if the user |
75 |
|
uses copy&paste between arbitrary documents. |
76 |
|
|
77 |
|
|
78 |
|
|
79 |
|
|
80 |
|
|
81 |
|
|