AW: Re: [Qgis-developer] Announcement: PostGIS Versioning System pgversion and QGIS Plugin

[Düster Horst]

Ivan

As I understand your approach correct your aim is to historize table
records. But that's not the aim of the pgversion approach. pgversion
supports concurrent editing like SVN. To keep all non committed records
it is neccessary to log them in a seperate table via materialized views.


>One more question, how is the view  '<tablename>_version' performing
>in case of huge tables ?

What is a huge table. We use pgversion with tables containing more than
100'000 records and it works fine.

Of course it is possible to combine pgversion with the different
historisation approaches. 

Regards
Horst

------------------------------------------------

Dr. Horst Düster
Stv. Amtschef / kantonaler GIS-Koordinator 

Kanton Solothurn
Bau- und Justizdepartement
Amt für Geoinformation
SO!GIS Koordination
Rötistrasse 4
CH-4501 Solothurn

Telefon ++41(0)32 627 25 32
Mobil   ++41(0)79 511 54 12
Telefax ++41(0)32 627 22 14

mailto:horst.duester at bd.so.ch
http://www.agi.so.ch



-----Ursprüngliche Nachricht-----
Von: Ivan Mincik [mailto:ivan.mincik at gmail.com]
Gesendet am: Sonntag, 7. November 2010 23:28
An: Horst Düster
Cc: qgis-developer at lists.osgeo.org
Betreff: Re: [Qgis-developer] Announcement: PostGIS Versioning System
pgversion and QGIS Plugin

On Sun, Nov 7, 2010 at 3:44 PM, Horst Düster <horst.duester at kappasys.ch>
wrote:
> I just finished the first beta version of my newly released pgvs
support
> system. The idea of this system is to enable the editing of a single
> PostGIS layer concurrently by more than one user at the same time,
> similar to source code versioning systems like CVS or Subversion. You
> will find further informations at:

Congratulations, this is much needed stuff. I wonder why such tools
are not discussed at PostGIS list.

I really admire Your work with one exception, which pushed me to
develop my own tools (based on Your ideas) [1].
What I do not like is, that after enabling version support, all
editing is done against newly created view '<tablename>_version' (and
set of belonging RULES). This works well in all software which doesn't
care if You are loading pure TABLE or VIEW. In my work, we are trying
to support as many GIS client as possible.
For example Manifold GIS is setting read only mode by default, when
loading VIEW (What about ESRI ? or CadCorp SIS ?). In this case, such
tools are unusable. You never knows, when somebody implements similar
behavior in new version of Your software.

One more question, how is the view  '<tablename>_version' performing
in case of huge tables ?


For somebody interested in my work, look here [1] (published source
code is quite outdated)




[1]
http://www.mail-archive.com/postgis-users@postgis.refractions.net/msg118
34.html

Ivan
_______________________________________________
Qgis-developer mailing list
Qgis-developer at lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/qgis-developer

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/qgis-developer/attachments/20101108/87fa9e96/attachment-0001.html