#MAPBENDER IRC Log - 2009-06-08

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
07:02:41 seven_: I know that ur, astroidex and Testbaudson will not be here today.
07:04:06 mandsch: hi seven
07:04:35 mandsch: ok, holidays or bolsena already ...
07:06:04 nazgul: hi...
07:06:15 nazgul: quite a number on holiday : )
07:06:28 mandsch: yeah, looks like ...
07:10:16 seven: I got active and did some marketing.
07:10:22 seven: at least at last...
07:10:34 seven: submitted two talks for FOSS4G
07:10:46 seven: will give one at GSDI 11 in Rotterdam
07:10:59 seven: Workshop for the Greek OSGeo chapter in two weeks
07:11:13 seven: and marketing material ( case study ) on prtr.
07:11:46 seven: PRTR is not exactly a Mapbender highlight but the whole project is cool.
07:12:00 mandsch: what is prtr?
07:12:09 seven: http://www.prtr.bund.de
07:12:14 sigq: Title: PRTR - Pollutant Release and Transfer Register ( at www.prtr.bund.de )
07:12:36 seven: The first German Federal Administration that uses OSM data
07:12:57 seven: the stack is core: PG/PostGIS, MapServer, Mapbender.
07:13:00 mandsch: hmm, nice.
07:13:15 seven: 36 servers in an ESX cluster serve the OSM WMS.
07:13:19 seven: Cool hardware...
07:13:37 mandsch: wow!
07:13:38 seven: largest install that I have been working on in the past years.
07:13:57 mandsch: are you administering the wms server?
07:14:03 seven: Yes.
07:14:11 mandsch: cool...
07:14:38 seven: They need much better cartography, but the raw stuff already works fine.
07:14:52 mandsch: just wanted to ask ...
07:15:01 seven: The interesting bit is that the OSM data comes right out of the databas.
07:15:05 mandsch: have you tested with agg rendering?
07:15:10 seven: We had to do a lot of generalization.
07:15:21 mandsch: ah, yes yes, it's awful slow ...
07:15:27 mandsch: but with 36 servers ...
07:15:34 seven: AGG rendering start below 1:40.000 only because images are size *10
07:15:38 seven: slow?
07:15:49 mandsch: not this service
07:15:51 seven: Ah, the WhereGroup OSM WMS is unsuable...
07:16:00 mandsch: but when using the osm data raw from the db ...
07:16:14 seven: maps are 36k with GIF / no AGG
07:16:20 seven: and up to 500k with PNG and AGG
07:16:41 seven: for people with ISDN access that is painful.
07:16:47 mandsch: ;- )
07:16:54 seven: We should detect speed and offer two versions.
07:16:56 mandsch: we need tiles ...
07:17:11 seven: First we need slippy maps.
07:17:23 mandsch: ok, that's a nice start ...
07:17:25 seven: So that maps stay until the new ones are loaded.
07:17:53 : * seven does not like tiling, no reprojection, fixed scale levels. Ugh.
07:18:21 : * mandsch neither, but he don't thinks that osm data is made to be served on-the-fly ...
07:18:27 seven: Hm.
07:18:36 seven: Probably not.
07:18:54 mandsch: but i easily convince of the opossite .
07:19:21 mandsch: is image size the only thing or also preocessing time with agg?
07:19:25 seven: It would be cool to be up to date on a daily basis.
07:19:44 seven: interestingly preprocessing with AGG does not make a difference.
07:20:08 mandsch: have you tried quantization?
07:20:13 seven: We actually just switch formats from GIF to PG. Rest does MapServer.
07:20:29 seven: mandsch: What is quantization?
07:20:41 mandsch: hmm, not sure if i spelled it correctly ...
07:20:48 seven: -> GIF to PNG
07:20:49 mandsch: it reduces the colors
07:20:52 seven: yes.
07:21:05 seven: but it seems that MapServer detects first,
07:21:06 mandsch: bit you still have the agg look
07:21:11 seven: which format.
07:21:14 seven: and then does not use AGG
07:21:26 mandsch: not, it won't use agg with gif
07:21:34 seven: Exactly. MapServer does not use AGG
07:21:48 seven: its not a Bug but it is a Feature. Maybe.
07:22:08 mandsch: we'll but it's cool.
07:22:25 seven: If we manage to make a daily diff with updated data
07:22:42 seven: then it would make a lot of sense that we use WMS and not canned tiles.
07:22:54 seven: Always the newest data...
07:23:01 mandsch: true ...
07:23:07 seven: We want to add a button "Edit now..."
07:23:14 seven: that links to OSM / Potlatch
07:23:30 seven: so that people can directly hack the map - and see it next day.
07:23:55 : * seven dreams on: Add a OSM editor to Mapbender...
07:24:10 seven: Who would found such a hack?
07:24:11 mandsch: har har ..
07:24:38 mandsch: you'd just need to make a osm-wfs ...
07:24:48 seven: Probably yes.
07:25:16 mandsch: then have 136 servers with geoserver ... ;- )
07:25:30 seven: Ouch.
07:26:12 mandsch: ok, christoph and Astrid are onholidays ...?
07:26:14 seven: I don't know. We would need a full OSM dump in one PG instance. No generalized data.
07:26:29 seven: Christoph is on holidays this week.
07:26:39 seven: Astrid is busy and travles tomorrow.
07:26:52 seven: Astrid is very busy currenlty.
07:27:01 mandsch: good, isn't it?
07:27:04 seven: One can tell when she does not show up here.
07:27:19 seven: We had the FOSSGIS e.V. meeting this weekend in LinuxHotel in Essen.
07:27:23 seven: She was there too.
07:27:31 mandsch: I talked to Marco already a bit...
07:27:34 seven: Busybusybusy.
07:27:36 seven: OK.
07:27:59 seven: We shoudl blog about this or write a report to fossgis-talk but who will do it... ?
07:28:20 mandsch: about the fossgis weekend?
07:28:25 seven: Yes.
07:28:48 mandsch: There were a lot of people i guess ...
07:28:49 seven: Can you ping Marco Lechner to add a title to #fossgis IRC chat?
07:28:52 seven: he is admin
07:29:04 mandsch: yep i can do ...
07:29:17 seven: We were 13 in total.
07:29:27 seven: Good progress, was a productive meeting.
07:30:08 mandsch: I#m still curious what will happen with intergeo ...
07:31:01 mandsch: I talked with Olaf about it, the other day ...
07:34:42 mandsch: seven: I have to go. Are you around this week?
07:36:29 mandsch: ok, have a nice evening. CU.