cocoon-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <>
Subject Re: [offtopic] Mad Ideas Re: Direction of links
Date Sat, 13 May 2000 10:57:13 GMT
Robin Green wrote:
> Jeremy Quinn <> wrote:
> >When you put <link href="blah.xml">Blah</link> in your document, you
> >something you can easily break, it is even worse if your link has to
> >contain meta data about the destination.
> >
> >What happens when you want to find out "what links to me?" if you have one
> >way links scattered throughout your documents.
> >
> >OK, I am rambling .....
> No, I think you're onto something.
> Take a deep breath, this is totally insane -
> How about a totally object-oriented quote "filesystem" unquote, down to the
> character level with every required level of semantic structure inbetween
> (XML, files, collections of files, you name it), with dynamic optimisation
> (a la Hotspot for data as well as code) for most often used queries etc.?
> That way everything on your system could be viewed as being part of a
> systemwide "object database" and addressed as such (it'd have to be one heck
> of a powerful object database, but still).
> Base it on the principle that actual rate of stored "information" increase
> (in the Shannon sense - useful data not redundancy) coming from the real
> world is often quite small in many applications (it's limited by the speed
> of people typing at keyboards, or scanners, barcode readers etc.) and you
> can index data as you input it (to an extent - obviously there are large
> combinations of ways in which you might want to index data).
> If you set it up right (or if the heuristics worked right you wouldn't even
> have to configure it) it would give you "instant" SQL, XQL etc. query power
> over your XML "files" - and _anything_ on your system - even in-memory Java
> objects, given sufficient security permissions - not just your conventional
> databases.
> Obviously that's some way off (!), but it's something I'd like to look into
> in my PhD (if I make the grade). That and/or Intentional Programming, which
> kind of fits in to object databases because programs in IP are represented
> as trees of objects IIRC.
> Mix in molecular nanocomputing ( ), and this might be
> _almost_ feasible! ;-)

You are talking about an XML-oriented file system with XPath and
XInclude capabilities.

Add revisioning and ACID properties and voilĂ : the perfect substrate for

Guess what? I was having the exact same thoughts about the need of
something like this.... but it's way beyond my actual scope.

Anyway, if you want to write it and start a project around here, I'm
_sure_ others will be enthusiasts :)

Stefano Mazzocchi      One must still have chaos in oneself to be
                          able to give birth to a dancing star.
<>                             Friedrich Nietzsche
 Missed us in Orlando? Make it up with ApacheCON Europe in London!
------------------------- http://ApacheCon.Com ---------------------

View raw message