Schemaless and hierarchical databases - Enterprise Web Developer Community

Message from discussion Schemaless and hierarchical databases
MIME-Version: 1.0
Date: Sat, 16 Aug 2008 03:19:39 -0700 (PDT)
User-Agent: G2/1.0
Message-ID: <d4a4bf92-d8ce-4181-bb6a-4cf983d1a034@y21g2000hsf.googlegroups.com>
Subject: Schemaless and hierarchical databases
From: rtweed <rob.tw...@gmail.com>
To: Enterprise Web Developer Community <enterprise-web-developer-community@googlegroups.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

Isn't it ironic that whilst ISC have spent the last 10 years
desparately trying to hide that "old-fashioned" underlying MUMPS
database and add a whole layer of object orientation and schema to
make it palatable to the marketplace, the rest of the world is
increasingly realising that schema-less and hierarchical databases are
actually what they're looking for in the modern, internet-focused
world.

Try Googling "schemaless database" and you'll very quickly find out
about new databases such as Poseidon (http://www.brainwavelive.com/
developers/applications/poseidon-database.html) and Amazon's SimpleDB
(http://www.amazon.com/gp/browse.html?node=342335011).  Indeed how
even more ironic to see this quote in http://www.satine.org/archives/2007/12/13/amazon-simpledb/
:

"Schemaless - major cool factor for me here; items are little hash
tables containing sets of key, value pairs"

So, far from being old-fashioned and something that turns people away,
the inherent schemaless nature of MUMPS is now something that can be
promoted as a cool feature!  My prediction is that schemaless is soon
to become the new schema!


How about hierarchical databases?  You'll see less about this when you
Google it, apart from sad arcticles explaining the bizarre lengths
folks have to go to in order to store hierarchical data in relational
databases.  However, check out Google BigTable and you'll find a new
implementation of the hierarchical database model, and of course,
because it bears the Google name, is automatically deemed to be cool!

The hierarchical database comes into its own for storage of the
naturally hierarchical structure of XML and, more recently, JSON.
Check out http://www.rpbourret.com/xml/UseCases.htm for  more info on
so-called Native XML Databases.

As both our own eXtc and EWD products have shown, the projection of
the W3C XML DOM is so straightforward and natural that you'd have
thought W3C had designed it to use a MUMPS engine!  As a persistent
storage engine for data exposed and transported as JSON strings, again
the MUMPS model is a perfect fit, and JSON is one of today's super hot
technologies.

So why am I harping on about this?  Because here we all are using a
technology (in both Cache and GT.M) that, far from being something you
should be apologetic about and trying desparately to hide, you should
be shouting about from the hill-tops.  What you've been taking for
granted for years, the rest of the world is now starting to realise is
exactly what they need.

Of course, at M/Gateway we've been saying this for years, eg
http://gradvs1.mgateway.com/download/extreme1.pdf.

So here's the thing: the MUMPS database is a schemaless hierarchical
database.  Both at once! How cool is that?!!  Imagine what you could
do with that?!  Answer: stuff that the rest of the mainstream can only
dream about but drool at the idea of being able to do.  And MUMPS is a
tried and tested technology, not some "new kid on the block" barely
out of a start-up's R&D labs.

So now you know how cool a technology it is, get out there and tell
the world about it....