[OSGeo-UK] Managing OSMM data updates with free software

[Walsh, Joanna]

dear all,

I promised to move a Twitter conversation about change-only updates to MasterMap data here to the list. Half-term is over and i am back at the desk. Here goes.

The story so far. At Registers of Scotland we are setting up a sort of in-house, custom OS On Demand-type service, there is a business requirement to make maps with no internet access if needed. 
There is a license for a proprietary software package to handle imports and updates for OS MasterMap. Our architect urged us to use this package if possible, specifically because it claims to automatically handle "change-only updates" to MasterMap. The goals are as much automation as possible and as little technical debt as possible. 

We tried it out and i really don't like the look of what it does with the database. It looks like an Oracle Spatial database which someone has minimally ported to Postgres but not really tested for retrieval. The tables are created with *no integer primary keys*, and forcing case-sensitivity with double quotes in table and column names has upset TileMill quite badly :/ It looks like there's going to be lot of analysing and indexing work needed to be done to make that data queryable at decent speed. This is not the behaviour i expect of a package whose marketing pitch is that it takes the pain away - if we frot the data, we lose the seamlessness of the update support anyway.

I downloaded the Astun Loader from github. It Just Worked TM, and because it hands over the PostGIS conversation to OGR, what comes out appears right and doesn't need optimisation.
https://github.com/AstunTechnology/Loader

So what am i asking the list for now?

- A way to reassure our architect that the open solution WILL lead to much less technical debt in the medium term 
- A straw poll of others in the OSGeo UK community. Some of you will already have in-house processes in place for dealing with change-only updates. If anyone's built that on the Astun Loader, are there any strong reasons that couldn't be contributed to github? (Embedded configuration / no benefit of releasing?) 

Politically; i think we could contribute to a project that had started, but i don't think we'd be able to start one. Whoever did would get the benefit of lower maintenance burden and be able to bake in their requirements. 

There's a parallel thread here concerning MapBox/TileMill/CartoCSS. I'm sketching out some CartoCSS stylesheets for MasterMap. I am not a cartographer or a designer, just doing the logic.
Anyway, the stylesheet is in part a model of what comes out of the Astun loader, closely tied to the PostGIS schema, so these are two sides of the same coin, right? 
It'd be interesting to see a few common public styles. We may need a minimal Land-Line style and it would be hilarious to see an OSM style. 

Thoughts apprec.,


Jo 
 
P Please consider the environment - do you really need to print this email?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/uk/attachments/20131022/a67fb9d8/attachment.html>