Re: Re-connecting

On Wed, 2005-02-23 at 06:02, Terry Loveall wrote:
>
> On Tue, 22 Feb 2005 22:12:28 -0500 (EST)
> Mark Slicker <[email protected]> wrote:
> > (snip)
> > My critique of hierarchical file systems is that they put too much burden 
> > on the user, to give each file a name and place within the hierarchy. Also 
> > they are not very general or flexible. I think the system could be much 
> > more active in recording information about data objects, and the system 
> > could make this information available in a general way to applications.
> 
> > To which Terry (I presume it was Terry at this point) replied:
> Forth, with a nested vocabulary structure, can provide the basis for
> 'attribute typed' data storage. The basic managed 'physical' tree structure is
> provided by the vocabulary nesting.
> (snip)
> ... for communication oriented apps, e.g. web browser or email client, it
> would probably be worth the effort to provide a rigorously defined
> 'meta-object' format that essentially allows simple replication of creation
> time/date, modified time/date, size determination, owner ...
>
> But all of the above could be applied to conventional OS design. See some of
> the recent write-ups on beOS. With colorForth, the opportunity is at hand to
> try it again.
> Regards,
> Terry Loveall
>
> Nick here:
> How about start simple with the CF list, which is not large but 
> already gives me - and Google! - trouble, trying to find who said 
> what when & where?  I think Forth could easily do:
> 
> 1. A "passive" program that reads the complete list (plus incoming 
>    messages) but only to store the lot as plain text documents in a 
> single folder, filed by date in an "archival stratum" - to be searched
> but never to have its temporal order disturbed. 
> 
> 2. An "active" program that reads 1. and "digests" it, by compiling
>    words whose initial definitions would be date of first appearance :
>  : word#n   date sender thread keywords ;  
> Subsequent appearances would make use of Forth's ability to redefine 
> a word by a link to the previous definition of that same word :
>  : word#n   newdate sender thread keywords word#n ;
> Names of senders would of course also be words - important keywords.
> Keywords could be selected either by native human intelligence or
> by a numerical filter e.g., from among all the words in the message
> to select any that are new or uncommon, plus a few that are already 
> in the dictionary.  The new words of course would have to be added 
> as soon as they were read - otherwise could not be used as keywords
> for the message in which they first appeared:
>  : newword  date sender thread ; 
> This digest of words with dates would be used to help search 1.
>   
> Most long words can be handled by most Forths, with their width of 31 
> characters, but CF users might need to watch out; I seem to remember 
> reporting (when?) a confusion by truncation of the end character from
> two words which I had carelessly called Longword#1 and Sameword#2.  
>
> Regards
> Nick