[Oberon] Re: pe.URI.Mod & heuristic translators & Dbase.

[easlab at absamail.co.za]

> cg> Watson.ShowDef URI.Mod == nill

Peter wrote:
> Open pe.URI.Mod and * mark it.  
> Open the Watson.Panel.  Click the Show Def * 
> button.  The definitions will be dsplayed.
> 
I assumed that URI was an existing module, and
just wanted to see 'what it does', which is usually
quickly done by using Watson to see what IDs 
have been chosen for the procedures, and perhaps 
some comments. This without having to mount my
partition which contains the sources, which are 
in any case not yet all de-compressed.

> cg> System.Directory UR*Obj == nill
> 
> System.Directory *UR* ~
> 
> cg> URI seems to suggest use of DNS ?
> 
> It is just a reimplementation of ESC
> and UnESC.  No direct involvement of 
> DNS.
> 
OK, thanks, I'll test it later.

=========
IMO one reason why the common dumbed-down
literal/verbatim [how to do] instruction method eg.
   cd /dogDirectory
   rm *Big *Small
is inferior to the 'what to do' method eg.
    "rename all *Big files to *Small files, in the 
    dogDirectory", 
is that it is less robust.

'What to do' allows potential mistakes to be 
avoided, by using the intelligence of the reader.

I fetched your web-page which is fine with N-O.
And Illustration my point above, your reference to
 <PeterE'sPartiton>URI.Mod is a typical mistake
which is best avoided by not trying to 'program
the user/reader'.

As a youngster I found that it is near impossible to 
build working circuits from detailec instructions, 
without knowing 'what you are doing', in order to 
be able to test and correct the instruction mistakes
and inevitable ambiguities.

Extending this principle: heuristic translators
are potentially valuable [cost-effective] 
applications.  Even if a C to Pascal/Oberon
translator is not possible without some human
intervention, a Bablefish-like product is valuable.
And cost-effective if it's data-driven, like the 
editor vim, which can be programed for any
laguage syntax to eg. colour rerserved words;
or the <old 4 lettered> assembler which could
have various instruction-sets added as they 
became available.

Right now I'd like a translator for:
   p(\mathbf x |\mathbf y, \theta) = \frac{p(\mathbf y,
   \mathbf x | \theta)}{p(\mathbf y | \theta)} =
  \frac{p(\mathbf y|\mathbf x, \theta) p(\mathbf x
   |\theta) }{\int p(\mathbf y|\mathbf x, \theta)
   p(\mathbf x |\theta) d\mathbf x}
which I'm guessing is latex ?
And with N-O's FontEditor useful latex and other
translators seem possible.

Of course Linux ETH-Oberon solves these problems.

The explosive success of wikidepia demonstrates
how not only 'exact answer applications' have value.
The value added by heuristic [fuzzy] human input is
great.  Today's global value: knowledge/competence
cannot be exactly quantified - like the square root of 9.

Re. cost effective modularity: can Dan Parnete's work
on data bases be used together with the existing
'send and fetch NewsGroup headerLists & articles',
to eg. 
* save retrieved headers in threads; keyed on subject
   sub-keyed on <some other criteria>
* order them by date...etc.
* classify them by Groups ...etc ?


== Chris Glur.