Re: Mumps and post-relational

From: Dawn M. Wolthuis <dwolt_at_tincat-group.com>
Date: Tue, 22 Jun 2004 08:17:06 -0500
Message-ID: <cb9bgq$isf$1_at_news.netins.net>


"Laconic2" <laconic2_at_comcast.net> wrote in message news:Bo2dnYZBb9bujkXdRVn-gw_at_comcast.com...
> Over in news:comp.databases there is a thread where what's being
discussed
> is a product called "Cache". Among the things it's being called is
> "post-relational". Somewhere in the discussion, it says that Cache has
> historical roots in MUMPS, and that MUMPS is multidiminsional.
>
> Is this true?

Cache' is a MUMPS implementation. From a timing perspective, it is a peer to PICK, although there does not appear to be any history of common specifications. It has historically been a layer within applications in the health care vertical market. I believe it is multidimensional in a similar way to PICK being multi-dimensional. By that I mean that it is NOT a MOLAP tool,but permits non-1NF structures (as it should;-)

But it is NOT, I repeat NOT, POST-relational from a timeline perspective. Both PICK and MUMPS implementations that have been done after relational modeling arose have called themselves post-relational. While the implementation itself comes after relational models, the data model for Cache' is from MUMPS and that pre-dates at least the popularization of relational modeling.

I recall readin a site that someone pointed me to once that talked about PICK and MUMPS and I don't see that offhand, but could look for it again if important to you. Cheers! --dawn Received on Tue Jun 22 2004 - 15:17:06 CEST

Original text of this message