Re: CODASYL-like databases

From: paul c <toledobysea_at_ac.ooyah>
Date: Thu, 03 Apr 2008 13:58:48 GMT
Message-ID: <so5Jj.22857$rd2.9034_at_pd7urf3no>

> The last standard for the network model was NDL; I still have a copy
> of the document. The real problems were (1) They could not be
> optimized since they had fixed access paths to the data (2) They were
> procedural and not declarative, so the programmer did all the work.
> (3) Everyone had a different language instead of a Standard like SQL
> to use (NDL came too late).
> ...

They have an unnecessarily complicated data interface, are basically impossible for an end-user to handle without a lot of expert help and programmers new to an app have a steep learning curve. Everybody has to know what kinds of 'records' have the data they want and then decide which of many operators and options are needed for access, then organize some other code and place the operators very carefully in that, then likely have to cope with numerous arcane environmental exceptions/return codes and then construct the kind of 'record' they wanted in the first place. The same data in two different databases might be kept in different record types and need completely different operators.

How could a 'standard' possibly improve that situation? Received on Thu Apr 03 2008 - 15:58:48 CEST

Original text of this message