Re: controversial paper

From: Mok-Kong Shen (mok-kong.shen_at_t-online.de)
Date: 09/30/03 

Date: Tue, 30 Sep 2003 09:36:57 +0200

Mxsmanic wrote:
>
> Mok-Kong Shen writes:
>
> > To be able to process a variable field, you would need
> > indicators and that also consumes space (bits).
>
> Yes, but if there is no constraint on the length of the variable field,
> this is not a problem.
>
> You actually don't even need an addressing scheme that is open-ended at
> both ends. Just allowing for unlimited high-order bits will do. That
> way, if you run out of allocations in 128 bits, you just add more bits
> at the high end. Like patent numbers.

Analogy: In programming in most cases one employs arrays
of fixed dimension (length) rather than data structures
that allows one to arbitrarily extend at run time. Why
is that so?
 
> > If I don't err, in data bases records for efficient processing
> > are in fixed format, eventually with links to variable length
> > informations as needed.
>
> A fixed record that links to a variable-length record is effectively a
> variable-length record.

I was giving (for completeness) the full picture of that
which is done there. In the present case, one doesn't
need variable length records from the very beginning,
as argued.

M. K. Shen

Relevant Pages

  • Re: Is the interleaved marker idea patentable ?
    ... > embedded into the variable field itself. ... from unary encoding to Huffman coding to variable-length strings. ... no clue as to how little you know, to think that this is a novel idea; ... or you are a troll. ...
    (comp.arch)
  • Re: controversial paper
    ... Yes, but if there is no constraint on the length of the variable field, ... > are in fixed format, eventually with links to variable length ... A fixed record that links to a variable-length record is effectively a ...
    (sci.crypt)
Quantcast