In-Reply-To: <436200FB.4090400@ccgis.de>
References: <7def66b8e277ec3dcfb0e90f001194fa@zeesource.net>
	<20051027094158.W82309@storm-user.niwa.co.nz>
	<bf8760a42ccac8bb461af93a0f2cc589@zeesource.net>
	<200510262102.27259.xzilla@users.sourceforge.net>
	<e480a4bdc7f680be0521f818a404defa@zeesource.net>
	<436200FB.4090400@ccgis.de>
Mime-Version: 1.0 (Apple Message framework v623)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <abe9efb7c3915c6783c3d22434aa5804@zeesource.net>
Content-Transfer-Encoding: 7bit
Cc: Robert Treat <xzilla@users.sourceforge.net>,
	pgsql-www@postgresql.org, Brent Wood <b.wood@niwa.co.nz>,
	postgis-users@postgis.refractions.net, MAPSERVER-USERS@LISTS.UMN.EDU,
	freegis-list@intevation.de, cavallini@faunalia.it,
	mapbender-users@lists.sourceforge.net
From: Claire McLister <mclister@zeesource.net>
Subject: Re: [Freegis-list] Re: [GENERAL] Map of Postgresql Users (OT)
Date: Fri, 28 Oct 2005 10:45:48 -0700
To: Arnulf Christl <arnulf.christl@ccgis.de>

Arnulf,

  Thanks for your message. I looked at the mapbender site, and it looks 
like it has the basic functions except the password protection, which 
you will probably need at some point:-) The performance from here 
(California) was slow, and the map image did not look as good as 
Google's. I suppose we have better map images available?

  I could not find Cunninghams talk on P2P style wiki's, but seems to me 
that it would be too much work for a simple web page. Ideally, we'd 
just want to have a static map with possibly Javascript popups. That 
way, the performance can be good and the web site doesn't become that 
onerous that the pg www group kicks us out.

  So, a question to you and Palo: Can you generate such static map 
(along with Javascript code) from Mapserver?

Claire

On Oct 28, 2005, at 3:44 AM, Arnulf Christl wrote:

> Claire McLister wrote:
>> Good point. This might actually be a problem. Google Maps API 
>> requires each server that is serving the map to be registered with 
>> Google, and send the corresponding key when making the Javascript 
>> request.
>> Unless the mirrors can each send their own keys, this will not work.
>> Claire
>
> Hi, more cross posting...
>
> Sounds like this task should be solved using WMS and WFS services. 
> Google is cool but in order to foster both Open Source and standards 
> (OGC) we would suggest to use e.g. MapServer as WMS to produce the 
> maps and GeoServer as WFS to manage geomtries. Those OWS services 
> could then be included in any OWS client, web interface, etc. 
> Everything is there and ready to go, the effort to get it to run 
> should be minimal.
>
> Our clients usually operate PostgreSQL databases with several million 
> geometries - those 500 to 600 markers could be hosted on any 
> antediluvian box and still be fast. You will know what to about 
> mirroring, we would probably suggest to implement P2P - read Ward 
> Cunninghams ideas  int he keynote at the Wikimania conference this 
> summer...
>
> We have done the user mmapping Mapbender users with tooltips, direct 
> link to the website, etc. You can have a look at it on the project 
> homepage http://www.mapbender.org
> Its done Wiki-style so that users can enter their own position (use 
> the blue flag) or remove or edit any position (use the i-button). No 
> user accounts yet but security and authentication is in place and have 
> just not yet been necessary (still waiting for spatial spam...)
>
> We will spread word about spatial data management using 
> PostgreSQL/PostGIS with MapServer, GeoServer etc. at the 
> http://www.opendbcon.net (database fundamentals) on Nov. 8. and 9. in 
> Frankfurt/Main, Germany.
>
> This idea is on short notice but i would be really very excited if i 
> could show the map there! We suggest the following alterantive 
> actions:
>
> - It should be easy to add a geometry_column to the existing 
> PostgreSQL database which alredy has the positions. Add a GeoServer 
> WFS to access the geomtry and a MapServer for WMS display.
> - Alternatively send Paolo or us or both the geo-positions and we add 
> them to our existing OWS infrastructure (that will takt a few hours of 
> work only).
>
> Beause there is not so much time until the conference this would just 
> be be a prototype which should in the long run move to the PostgreSQL 
> or PostGIS homepage or Wiki or both. Yet another idea:  Mediawiki (the 
> Wikipedia software) is also right now introducing geometries to the 
> Wiki database - obviously also using PostgreSQL/PostGIS. This would be 
> another cool multiplier.. and there we also meet with Google again.
>
> Best, Arnulf.
>
>
>> On Oct 26, 2005, at 6:02 PM, Robert Treat wrote:
>>> On Wednesday 26 October 2005 18:24, Claire McLister wrote:
>>>
>>>> On Oct 26, 2005, at 2:08 PM, Brent Wood wrote:
>>>>
>>>>> There has recently been a thread on the Postgres user list about a 
>>>>> web
>>>>> based postgres user/developer map. Claire has built a Google map 
>>>>> based
>>>>> system, getting locations from IP addresses.
>>>>
>>>>
>>>> Just so people know, this map can be found at:
>>>>
>>>>   http://www.zeesource.net/maps/map.do?group=456
>>>>
>>>
>>> I'm going to loop in a few more people here :-), namely the pg web 
>>> team who
>>> will probably be able to give pointers on site integration. One 
>>> thing to keep
>>> in mind is that the postgresql website is statically mirrored onto a 
>>> number
>>> of different servers, so any solution we come up with will ideally 
>>> allow
>>> itself to that. I believe a google maps system can do this (we can 
>>> mirror the
>>> javascript code and the location/data file on all mirrors and the 
>>> google map
>>> will work with it as is), which is why I originally went with that 
>>> type of
>>> solution.
>>>
>>> -- 
>>> Robert Treat
>>> Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL
>> _______________________________________________
>> Freegis-list mailing list
>> Freegis-list@intevation.de
>> https://intevation.de/mailman/listinfo/freegis-list
>
>
> ---------------------------(end of 
> broadcast)---------------------------
> TIP 9: In versions below 8.0, the planner will ignore your desire to
>       choose an index scan if your joining column's datatypes do not
>       match