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

Re: Some Questions to Xanadu

Hash: SHA1

On Sun, 31 May 1998,  Soeren Grenzdoerffer 
<leonardo@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:

> Andrew Pam wrote:
> > 
> > On Tue, May 26, 1998 at 03:46:35PM +0200, Soeren Grenzdoerffer wrote:
> > > a)
> > > Ted Nelson wrote in several publications that there is no censorship in
> > > Xanadu.
> > 
> > That is certainly one of the very strong goals.
> > 
> > > How will you handle forbi[dd]en topics like bomb-building or
> > > child-pornograp[h]y?
> > 
> > The important question is, who says they are forbidden?  In each country
> > there are different laws and customs and different things are forbidden.
> > 
> > > In Germany they trying to disconnect server in the WWW with such topics from
> > > the net. I can't imagine that the government would allow Xanadu publishing
> > > such topics.
> > 
> > We hope that they will not be able to prevent it, since Xanadu is intended
> > to be a global system.  Of course servers in a particular country need not
> > carry material that would be illegal in that country, but that does not
> > prevent other servers in other countries from carrying the information.
> > This is of course exactly the existing situation with the Web.
> But is there not the danger that different gouverments ban Xanadu
> because such topics (and I don't think about small countries but
> countries like USA or the coutries of Europa. (OK. I know only the
> discussion about that in germany but I can imagine that the
> discussion is in other countries too.))
> > 

This is a serious point, and its a good thing its been brought up. 
The WWW has caused plenty of such issues, and they become more 
serious still in some ways with Xanadu. OTOH, as I said before, since 
'policy' on such issues is set by the client software, what may 
happen instead is that laws requiring certain filters get passed (not 
at all a desirable thing, but better than an outright ban). 

> [snip b), c), d), e)]
> > > f)
> > > Without removing any document out of Xanadu how do you want to handle
> > > 'junk-information' ?
> > 
> > What is junk information?  This is largely subjective.
> > Also, we plan to have Hierarchichal Storage Management, so that
> > infrequently accessed information will migrate to off-line media which
> > will not be loaded on-line again until requested by someone.
> IMO is 'junk-information' small pages to a topic which are linked to
> a big document, but has only a small phrase, a Link and a small
> picture on it (e.g.'I like Ted Nelson and his Xanadu-System', a Link
> to xanadu.net and a small picture of Ted Nelson). No big
> information, but if you're searching something about the topic you
> find that link you can follow.

Well...the short, and somewhat dishonest, answer is to point out 
since Xanadu doesn't work like HTML does - it is 'chunk' oriented 
rather than page oriented (a link points to a specific datum, not to 
the layout as a whole) and the links are exterior to the data (and 
can point to any portion of the data specifically needed) - it should 
be easier to search through and filter the data for the kind of 
information desired. This ignores the means by which this filtering 
would have to be done, and it presents what (to me) the most serious 
question about Xanadu - how  one can find specific information on a 
broad topic. 

For example, a search on the word 'Napoleon' may bring up links to 
biographies of Napoleon Bonaparte, or ones of Napoleon III, or the 
French Revolution, or the Napoleonic Wars, or the Franco-Prussian 
War, or Corsica, or military theory, or the Code Napoleon, or 
napoleon pastries...or any of the hundreds of people, places, and 
things named 'napoleon'. Now, the way Xanadu indexes documents should 
help eliminate false leads (at least, as I understand it), and the 
client software will usually be intelligent enough to at least filter 
and sort the information, so that it present it by 'topic' or 
relationship or whatever. However, 'topics' are arbitrary, and 
either the client software would have a built in 'topic index' 
(giving it a bias of the kind Xanadu is supposed to help eliminate) 
or the user would have to give (extensive) sorting criteria (which 
could be defaulted, but how?).  Even assuming that presentation 
problems can be solved, the experience of the Web is that some 
searches will bring up hundreds, even thousands of links - and Xanadu 
is supposed to be much, much larger. What solutions have been 
proposed to handle information overload like this?

That having been said, I should point out that Xanadu is intended to 
do much more than the Web does;  the goal is a general system for 
accessing and manipulating information of all kinds. It can be used 
for HTML pages...or databases, or spreadsheets, or  'word 
processing', or *any*thing involving 'persistent data'. Its not meant 
to replace the Web; its meant to replace file systems. Web-type 
'pages' and searches over them would be a special case use, but it 
would be used for - well, everything. I personally think that it 
could be used for other 'system services' as well...but the user 
should need to know about those kinds of implementation details, and 
it really isn't an intended use of Xanadu (the only reason I think of 
it like that is because I'm a bit too much of a technoid).

Version: PGP 5.5.x - QDPGP 1.71
Comment: http://community.wow.net/grt/qdpgp.html

"I personally am far too abundantly endowed with the three cardinal
virtues of programming, to wit, laziness, impatience, and hubris, 
to undertake any such project." - Gregg Reynolds, on comp.lang.sather
Schol-R-LEA;2 *** Now 'scholr' at 'slip.net'