www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - [OT] The One Hundred Year Data Model

reply Justin Johansson <no spam.com> writes:
Perhaps off topic for this NG, though certainly a good topic for LtU, 
but nevertheless D people might have some interesting insight on the 
topic of data models (for programming languages).

So I'll begin with saying, "Forget the Hundred-Year Language"
c.f. http://www.paulgraham.com/hundred.html
and 
http://tapestryjava.blogspot.com/2008/12/clojure-hundred-year-language.html

and drop the notion of the "Next Big Language" per se.

Let's take a step back and instead ask what might be the Next Big Data 
Model or the Hundred-Year Data Model in the same vein as Paul Graham 
contemplates (link above) a Hundred-Year Programming Language.

While discussions about programming languages, syntax, static vs dynamic 
typing, etc are about ubiquitous and can be as emotional as political 
and religious ideological discussions, it seems (to me at least) that 
in-depth discussions about data models are few and far between.

Apart from the Third Manifesto (the relational database data model) made 
famous in decades past
http://www.thethirdmanifesto.com/
there have been few advancements in abstract data models since then.

While there may be others, the only significant new data model in the 
last decade that I know of is the XQuery 1.0 and XPath 2.0 Data Model (XDM)
http://www.w3.org/TR/xpath-datamodel/

With the above preamble, I would like to ask members of the D community 
  to contemplate about what the ubiquitous data model of the future, 
perhaps the Hundred-Year Data Model, might be in shape or form, taking a 
programming language agnostic position.

Cheers,

Justin Johansson
May 15 2010
parent reply Spacen Jasset <spacenjasset yahoo.co.uk> writes:
Justin Johansson wrote:
 Perhaps off topic for this NG, though certainly a good topic for LtU, 
 but nevertheless D people might have some interesting insight on the 
 topic of data models (for programming languages).
 
 So I'll begin with saying, "Forget the Hundred-Year Language"
 c.f. http://www.paulgraham.com/hundred.html
 and 
 http://tapestryjava.blogspot.com/2008/12/clojure-hundred-year-language.html
 
 and drop the notion of the "Next Big Language" per se.
 
 Let's take a step back and instead ask what might be the Next Big Data 
 Model or the Hundred-Year Data Model in the same vein as Paul Graham 
 contemplates (link above) a Hundred-Year Programming Language.
 
 While discussions about programming languages, syntax, static vs dynamic 
 typing, etc are about ubiquitous and can be as emotional as political 
 and religious ideological discussions, it seems (to me at least) that 
 in-depth discussions about data models are few and far between.
 
 Apart from the Third Manifesto (the relational database data model) made 
 famous in decades past
 http://www.thethirdmanifesto.com/
 there have been few advancements in abstract data models since then.
 
 While there may be others, the only significant new data model in the 
 last decade that I know of is the XQuery 1.0 and XPath 2.0 Data Model (XDM)
 http://www.w3.org/TR/xpath-datamodel/
 
 With the above preamble, I would like to ask members of the D community 
  to contemplate about what the ubiquitous data model of the future, 
 perhaps the Hundred-Year Data Model, might be in shape or form, taking a 
 programming language agnostic position.
 
 Cheers,
 
 Justin Johansson
 

effectively a dimension. This has been around for a while and it popular in data warehousing applications. Data is king these days and extracting information from it easily is what everyone wants to do that has a lot of data. I am surprised to find that there don't really appear to be many open source libraries to support this sort of thing. Like for example sqllite in the RDBMS field. There is jbase with it's multi valued fields and so on, but as far as I can tell it is built on top of a relational model which is more OLAP. Whereas I think I am thinking more along the lines of MOLAP.
May 15 2010
next sibling parent =?ISO-8859-1?Q?=22J=E9r=F4me_M=2E_Berger=22?= <jeberger free.fr> writes:
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Spacen Jasset wrote:
 What about the dimensional database model. Where each attribute is
 effectively a dimension. This has been around for a while and it popula=

 in data warehousing applications. Data is king these days and extractin=

 information from it easily is what everyone wants to do that has a lot
 of data.
=20
 I am surprised to find that there don't really appear to be many open
 source libraries to support this sort of thing. Like for example sqllit=

 in the RDBMS field. There is jbase with it's multi valued fields and so=

 on, but as far as I can tell it is built on top of a relational model
 which is more OLAP. Whereas I think I am thinking more along the lines
 of MOLAP.

Funny, just after reading your post I discovered mdds which looks like what you're describing: http://code.google.com/p/multidimalgorithm/ Jerome --=20 mailto:jeberger free.fr http://jeberger.free.fr Jabber: jeberger jabber.fr
May 15 2010
prev sibling parent sybrandy <sybrandy gmail.com> writes:
 I am surprised to find that there don't really appear to be many open
 source libraries to support this sort of thing. Like for example sqllite
 in the RDBMS field. There is jbase with it's multi valued fields and so
 on, but as far as I can tell it is built on top of a relational model
 which is more OLAP. Whereas I think I am thinking more along the lines
 of MOLAP.

First, OLAP is actually data-model independent. You're thinking ROLAP (Relational-OLAP. E.g. Star schemas in a relational database). Second, I'm not sure that a cube (E.g. the structure behind MOLAP) is the wave of the future as there are column-oriented data stores, such as Infobright, that support ROLAP with great performance and compression while still supporting the relational model. In fact, Infobright uses MySQL, so you can even use the same MySQL drivers to access the database. Casey
May 15 2010