Re: Neo4j and Topic Maps

Quoting Patrick Durusau <patrick-Q/[email protected]>:
> Murray,
>
> I haven't looked at Graham and Kal's BrightstarDB, mostly because I  
> would have to have a MS Server setup (as I understand it) and that I  
> don't have. And probably not going to invest in the  
> hardware/software to create such a platform.
>
> If you have or can obtain access to such a platform, I would love to  
> hear about BrightstarDB.

Hi Patrick,

Yes, I appreciate that Graham and others are doing that work but that's
the same problem I'd have, as I'm on Linux, using Java.

> On Neo4j, I am curious what you think is required to write a "..new  
> engine on top of Neo4j directly"?
>
> Being mindful that I don't think of "merging" as re-writing anything  
> but presenting results to a user as though rewriting had occurred.

I've been looking at a couple of different approaches.

One is using a separate document store (perhaps off the shelf) and
mapping the documents and links using Neo4j as an explicit "Dynamic
Link Manager" or DLM (ala Arbortext); another is building the document
store into Neo4J and from that having an implicit DLM. The idea of a
"new" Topic Map engine would be to build it implicitly into Neo4j,
using the graph structure literally as the Topics and Associations
of a Topic Map. The Topic Map engine would *be* the DLM, and vice
versa. If there's any innovation in my idea it's having the document
store itself be an expression of a Topic Map. The one question that
arises is how to keep the classification system/ontology separate
from the documents, though in the TM model they'd just be Occurrences.

I wish I had 10 million dollars and a year -- I'd build the thing. But
I know a lot of people on this list who are in a similar situation...
My last project took me about eight years and I finally ran out of
steam. Not sure I'd want to repeat that. This time I'd want a team.

> That is one of the weaknesses of my diagram of a topic map tool  
> chain. I under specified what would/could be done by a topic map  
> engine.

That doesn't sound like you.

> BTW, using Neoj would not be implementing the TMRM. That could be,  
> not necessarily would be, implementing a different data model with  
> different rules, from the TMDM. Hopefully disclosed so that the  
> results could be merged with the results of using other legends.

Yes, I realise I misspoke. What I meant was the idea of that implicit
graph structure within the Neo4j graph database exposing an effective
Topic Map, and whatever changes might be necessary to be able to
consider it an expression of the TMRM, without consideration of an
explicit data model. If a data model were in order the TMDM would of
course make the most sense.

That is, unless I stick with TM4j...

[shrieking in background heard]

Murray

...........................................................................
Murray Altheim <murray10 at altheim dot com>                       = =  ===
http://www.altheim.com/murray/                                     ===  ===
SGML Grease Monkey, Banjo Player, Wantanabe Zen Monk               = =  ===

       Boundless wind and moon - the eye within eyes,
       Inexhaustible heaven and earth - the light beyond light,
       The willow dark, the flower bright - ten thousand houses,
       Knock at any door - there's one who will respond.
                                       -- The Blue Cliff Record