Hi all,
Some hint of Ldp to be used in OSM Fixer was trying to find ways that
are in the current form:
osm version=0.5 generator=OpenStreetMap server
way id=4073741 visible=true timestamp=2008-10-30T01:34:45+00:00
user=stev
nd ref=21558066/
nd ref=21558066/
nd ref=21558063/
nd ref=21558064/
The following ways deleted because they are actually one dimensional
ways containing duplicate nodes:
4211255
4269986
4629424
4681902
5638870
5696224
6132110
7459349
8925950
8934066
9234840
9948455
9975882
10046176
10068241
10183859
10474290
11105612
11637539
11708512
12172604
12178293
12184101
Stefan de Konink wrote:
|20869 | 30850313 |
|20869 | 31004817 |
Someone reported that the name of this way changed.
http://api.openstreetmap.org/api/0.5/relation/20869/history
Doesn't work, so I cannot check.
Stefan
___
dev mailing list
We are now also deleting out of sequence deletes and fixing them. The
following ways were never 'propagated' to be deleted in a planet changeset:
10129926
11078370
18991932
32277509
32569822
32709279
32709343
32709354
32709359
32709360
32709397
32720156
32922974
33002141
We recreate the way,
Matt Amos wrote:
On Fri, Apr 10, 2009 at 5:39 PM, Frederik Ramm frede...@remote.org wrote:
Matt Amos wrote:
if the object is new then the version number won't help - a new object
will be created from the placeholder ID in each time.
Right. What Richard meant was probably that the new server
If someone wants to help seeding:
http://thepiratebay.org/torrent/4838395
- Download the torrent into the directory where you downloaded
the planet
- Start rtorrent on the file
- Hash is checked, you start seeding :)
Stefan
___
dev mailing list
+--+--+
| relation | to_way |
+==+==+
|76599 | 25437953 |
| 108088 | 32900081 |
| 108089 | 32900081 |
|19798 | 32534961 |
|20869 | 30850313 |
|20869 | 31004817 |
+--+--+
All fixed again :)
---
For more information:
On Thu, 9 Apr 2009, Tom Hughes wrote:
Frederik Ramm wrote:
Glad they haven't started user diaries yet ;-)
People started doing that a long time ago. I kill them when people
report them.
I have some cloudmade people always spamming me about mapping parties in
the states, now I am not from
On Thu, 9 Apr 2009, Matt Amos wrote:
On Thu, Apr 9, 2009 at 1:05 AM, Tijs Zwinkels
openstreet...@tumblecow.net wrote:
I mean, come on, one of these methods must be able to detect street signs.
:)
i stumbled across a paper about tracking text in videos[1]. maybe some
of the methods
On Thu, 9 Apr 2009, Matt Amos wrote:
On Thu, Apr 9, 2009 at 11:37 AM, Stefan de Konink ste...@konink.de wrote:
On Thu, 9 Apr 2009, Matt Amos wrote:
On Thu, Apr 9, 2009 at 1:05 AM, Tijs Zwinkels
openstreet...@tumblecow.net wrote:
I mean, come on, one of these methods must be able
http://staff.science.uva.nl/~ksande/research/
Review this too.
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev
On Thu, 9 Apr 2009, Tijs Zwinkels wrote:
On Thu, Apr 9, 2009 at 2:29 PM, Stefan de Konink ste...@konink.de wrote:
http://staff.science.uva.nl/~ksande/research/http://staff.science.uva.nl/%7Eksande/research/
Review this too.
Interesting evaluation of color-aware feature extractor. It's
Relations:
67028
108258
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev
I really want to understand this; and to my intel transactions of such
amount of time are just not bound to fail, looking at the timestamp.
http://api.openstreetmap.org/api/0.5/way/28366538/history
http://api.openstreetmap.org/api/0.5/relation/108258/history
Alpinfuchs deleted way 28366538 at
Andy Robinson (blackadder-lists) wrote:
For mappers we need the simplest quickest way of entering the data, and this
works for us. I don’t really care if that makes it more difficult to process
the data in the database, that’s not important while building the database
of data.
Would dynamic
On Mon, 6 Apr 2009, Dave Stubbs wrote:
Anyway, Shaun refactored the whole lot for API 0.6. It now atomically
checks relation consistency within the rails models when doing an
update rather than relying on the caller to do it. So this should be
fixed in a couple of weeks a long with everything
Andy Robinson (blackadder-lists) wrote:
Well, maybe if you explain what that means I can respond ;-)
Currently the editor or the server decides what the user sees; I can
imagine that in the 'data is present' way it is far more easy to delete
or update a value that is in unique(k,v) format than
Relations:
108213
Ways: (* = wrong delete sequence, never exported as deleted)
11078370 *
27938082
32277509 *
32709279 *
32709343 *
32709354 *
32709359 *
32709360 *
32709397 *
32720156 *
32813939 *
32828133
32843076
32843303
32846607
32848957
32850436
32852058
32853640
32856292
32860854
32861163
Matt Amos wrote:
i look forward to seeing your tag comparison dictionary, though!
For talk-nl I alread gathered this statistics; since I can relatively
easily provide distributions over that. It is just what you want to view
them for.
Stefan
___
Hi all,
One of the last things I wanted to add was 'native' update support for
database to database transactions. Currently known as OSMChangeset; the
attachment is what I import into the database and the following file
generates it.
Relations
105969
62103
107292
107296
Ways
24175861
27793756
32555395
32575412
32718660
32718666
32719383
32719888
3272
32720010
32725781
32726097
32727711
32729253
32732054
32732080
32732976
32733383
32734873
32734925
32735026
32736694
32749831
32757332
32758397
32760061
32760180
32764004
Matt Amos wrote:
But please keep in mind that SIFT will *not* work for the typical
streetsign because SIFT looks for features, which are typically not
present in the 'easy to understand' signs (because they use colours).
wouldn't SIFT (or variants thereof) work really well against
Martin Koppenhoefer wrote:
Of course this would be a lot of work and discussion in the community,
but I think it's still not too late to do it, and it would help a lot
to bring more structure and logics in the tagging-scheme.
Still will probably still not render the right icons (aligned) at
Ian Dees wrote:
Is there an rsync or similar mirroring system set up for distributing
the planet files? I'd love to host them (and might be able to convince
Amazon or Google to help with the bandwidth and drive space), but
transferring them at 150K/sec is pretty darn slow.
NL has its own
Eddy Petrișor wrote:
Stefan de Konink a scris:
Martin Koppenhoefer wrote:
Of course this would be a lot of work and discussion in the community,
but I think it's still not too late to do it, and it would help a lot
to bring more structure and logics in the tagging-scheme.
Still will probably
Ian Dees wrote:
I think we're talking about different things here, Stefan. I'm looking
at mirroring the planet dump. It turns out Grant already answered my
question in Febuary [1].
I'm talking about that too :) mirror.openstreet.nl (currently syncing
with hypercube).
Stefan
Matt Amos wrote:
maybe what we need is an amenities tag specifically for multiple
co-located amenities?
Again that would be a semantic solution for the 'joint' problem.
Stefan
___
dev mailing list
dev@openstreetmap.org
Matt Amos wrote:
2009/4/4 Martijn van Oosterhout klep...@gmail.com:
Just reading this again made me realise that we could probably make
osm2pgsql faster for the common case just by forking and using
gzip/bzip2 to do the decompression in a separate process (i.e. another
CPU)... Not entirely
Robert Joop wrote:
On 09-04-04 18:57:02 CEST, Ian Dees wrote:
Is there an rsync or similar mirroring system set up for distributing the
planet files? I'd love to host them (and might be able to convince Amazon or
Google to help with the bandwidth and drive space), but transferring them at
Stefan de Konink wrote:
Just starting a tracker I guess?
http://thepiratebay.org/torrent/4824136
IPv6 users; the pirate bay is known not to work on it anymore :(
Stefan
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org
Hi All,
While for now just an experiment, I would like to ask the other
mirroring people to opt in aswel for seeding the last planet mirror. The
task is pretty trivial.
Download into your mirror directory;
http://torrents.thepiratebay.org/4824136/planet-090401.osm.bz2.4824136.TPB.torrent
Grant Slater wrote:
Stefan de Konink wrote:
http://thepiratebay.org/torrent/4824136
Dude, that is funny.
When the next release coming out?? ;-)
should have put as description:
--== 1337 Mega Zero-Day Warez! ==--
Leaked direct from the Netherlands
OSM Libré Mega GIS Road
D Tucny wrote:
Then shouldn't that be two nodes? with the atm one having a layer=1 tag?
Similar to having a bar above a restaurant above a bank at the same
position but on different floors...
There are better examples probably; what should a supermarket that has a
post agency be talked
Tijs Zwinkels wrote:
I just posted my proposal to the wiki:
http://wiki.openstreetmap.org/wiki/GSoC_Student_Applications_2009#Draft_application:_Automatic_Street-Sign_Detection_and_Reading
.
It might still have a few rough edges. Feedback is welcome!
A few specific questions:
- Do you
Tijs Zwinkels wrote:
On Fri, Apr 3, 2009 at 8:51 PM, Stefan de Konink ste...@konink.de
mailto:ste...@konink.de wrote:
I would go outside JOSM, and make this a webapp that can be included
as a layer in any editor. WMS like.
I'm not sure whether this is very suitable
Matt Amos wrote:
2009/4/2 Stefan de Konink ste...@konink.de:
Iván Sánchez Ortega wrote:
El Jueves, 2 de Abril de 2009, Will Nordin escribió:
I am trying to convert the planet-090319.osm (approx 107GB) to gml. I've
tried the Java and Python solutions, but both end up thorwing OutOfMemory
Dave Stubbs wrote:
And while currently possible in the database, it doesn't work in the
current API, so you can't actually enter them :-)
Just download an object, duplicate the tag and reupload. Sounds pretty
possible to me :)
In a few weeks time it just becomes official, and any existing
Matt Amos wrote:
On Thu, Apr 2, 2009 at 10:33 PM, Stefan de Konink ste...@konink.de wrote:
Dave Stubbs wrote:
And while currently possible in the database, it doesn't work in the
current API, so you can't actually enter them :-)
Just download an object, duplicate the tag and reupload. Sounds
On Wed, 1 Apr 2009, Tijs Zwinkels wrote:
- Any opinions about / ideas for my proposal?
- Anybody who would be interested in mentoring this project? Pretty please?
- Let me know, and I'll arrange things with the right people.
Could you tell us what the difference is with probably any
On Wed, 1 Apr 2009, Tijs Zwinkels wrote:
On Wed, Apr 1, 2009 at 4:26 PM, Stefan de Konink ste...@konink.de wrote:
I know this is most definitely not the first time that geotagging photo's
with gps data has been attempted.
You are right... ;)
http://code.google.com/p/gpicsync/wiki
The following ways have been altered:
5969406
10507178
32349559
32504849
32657837
32657847
32657856
32657863
32657865
32657868
32658963
32663070
32664970
32667930
32669918
32671444
32673646
32674030
32674192
32674668
32677368
32679329
32686827
32686829
32686834
32690114
32691563
32691615
32691682
Marc Schütz wrote:
Use as delimiter, (since it can never
be used in key or value tag)
I'm pretty sure it can; AFAIK there are no limitations on what may occur in
keys and values (besides NUL).
We are running an XML system, that prohibits the entry of lessthan and
greaterthan in plain
Matt Amos wrote:
On Tue, Mar 31, 2009 at 2:33 PM, Stefan de Konink ste...@konink.de wrote:
Marc Schütz wrote:
Use as delimiter, (since it can never
be used in key or value tag)
I'm pretty sure it can; AFAIK there are no limitations on what may occur in
keys and values (besides NUL).
We
Matt Amos wrote:
if this is really annoying you that much, please take a holiday for 3
weeks and come back. the referential integrity, tagging and other
problems you're having to deal with will be gone* and the world will
be a happier place!
* hopefully. i guarantee nothing ;-)
As written
Matt Amos wrote:
On Tue, Mar 31, 2009 at 4:11 PM, 80n 80n...@gmail.com wrote:
Let me see if I understand correctly what you are saying. You think that
the currently proposed migration from the old server and old schema to the
new server with a new schema that includes some referential
Stefan de Konink wrote:
The following ways (79) are broken:
8171646
The above way seems to work now again.
way id=8171646 visible=false timestamp=2009-03-31T15:36:33+01:00
user=TomH/
Thanks for that; could you see what was wrong with the server?
Stefan
Richard Fairhurst wrote:
Stefan de Konink wrote:
Could anyone with knowledge in ruby *pretty please* take a look
at the source code and invent some wheel that Potlatch will never
'reupdate' duplicate k/v-pairs by itself.
Could you _please_ tell me, given that it has been explained to you
Stefan de Konink wrote:
Could anyone with knowledge in ruby *pretty please* take a look at the
source code and invent some wheel that Potlatch will never 'reupdate'
duplicate k/v-pairs by itself.
Today RichardF informed me that 'this' feature exists for Potlatch and
therefore the problems
On Mon, 30 Mar 2009, Martijn van Oosterhout wrote:
On Sat, Mar 28, 2009 at 4:29 PM, Stefan de Konink ste...@konink.de wrote:
But...it appears that there are cases where a node has been deleted from
the database while being referenced by a way.
Welcome in the fantastic world of OSM :D I
Thomas Wood wrote:
I believe the recommended method to get a fully consistant dump of the
data was to download the dump, then apply the next day's changefile.
Or at a constraint to the dump select file that it will not fetch from
the current table, but rather from the last history entry (that
Could anyone with knowledge in ruby *pretty please* take a look at the
source code and invent some wheel that Potlatch will never 'reupdate'
duplicate k/v-pairs by itself.
So to make everything clear:
1) I don't care that the wrapper layer screws up [so be it]
2) But if the object is touched
Hi Ben,
Ben Supnik wrote:
I wrote (yet another) small C program to split the planet XML into
tiles. The only thing that I think is interesting about my tiler (vs.
the exisging ones) is that it can run the whole planet on a 32-bit machine.*
How do you handle the LargeFile allocation?
D Tucny wrote:
Again, this isn't so simple... doing this would remove some errors
(deleted objects in a relation) and introduce new ones (relations that
are missing critical components or otherwise don't make sense)...
Nothing is simple; but even in the 0.6 API there will no keyboardl/chair
Shaun McDonald wrote:
On 27 Mar 2009, at 12:14, Rolf Bode-Meyer wrote:
2009/3/26 Matt Amos zerebub...@gmail.com:
if we were to convert all the tables to innodb we could add
transactions, but this would mean significant API downtime. happy for
us, then, that we're rolling out 0.6 Real Soon
Matt Amos wrote:
This is for the main DB server. Part of the 0.6 transition will
include moving from MySQL to Postgres for the sanity of the database
admins, and improved data integrity.
...and still 0 on a real benchmark.
if you'd like to help out it would be great if you could do some
Dave Stubbs wrote:
Some of them are due to different bugs, but in general yes. The
relations that reference node 0, 1 and other numbers below about 20
are a complete mistake (bug fixed ages and ages ago now) -- those
nodes should be removed from the relations.
It seems that downloading a
Stefan de Konink wrote:
Dave Stubbs wrote:
Some of them are due to different bugs, but in general yes. The
relations that reference node 0, 1 and other numbers below about 20
are a complete mistake (bug fixed ages and ages ago now) -- those
nodes should be removed from the relations
The amount of 'potlatch' in the attached document is in my opinion again
a bit too high. This ran on the planet export of yesterday. So I'll
check it in today.
I'll also run the 'hey, go away, empty way'-script but those are still
to be detected.
Stefan
Richard Fairhurst wrote:
Stefan de Konink wrote:
The amount of 'potlatch' in the attached document
As has been explained to you, you mean the amount of API ;)
Yeah yeah... I knew this would trigger you to a response; My next mail
includes the b0rk3d ways and relations. I can already tell
Stefan de Konink wrote:
I'll also run the 'hey, go away, empty way'-script but those are still
to be detected.
http://kinkrsoftware.nl/contrib/osm/brokenways.txt.gz
The 'hey borked relation'-script output:
http://kinkrsoftware.nl/contrib/osm/brokenrelations.txt.gz
I'll download the ways
Dave Stubbs wrote:
2009/3/25 Stefan de Konink ste...@konink.de:
Richard Fairhurst wrote:
Stefan de Konink wrote:
The amount of 'potlatch' in the attached document
As has been explained to you, you mean the amount of API ;)
Yeah yeah... I knew this would trigger you to a response; My next
Andy Allan wrote:
On Wed, Mar 25, 2009 at 3:07 PM, Stefan de Konink ste...@konink.de wrote:
It's a known problem with a fix on the way with 0.6.
Is this already checked that it will be fixed using 0.6? Since currently
the API doesn't seem to be the problem, the lacking referential
Matt Amos wrote:
yeah, the wiki page could probably do with some updating. but you
remember this thread last month, right?
http://lists.openstreetmap.org/pipermail/dev/2009-February/014024.html
the issue with things referencing deleted items should go away because
the transactions wrap the
Stefan de Konink wrote:
So I collected all the 'problematic' things. I think by what you
mention, it can be trivially fixed just by fetching all nodes and
reinserting them again?
s/nodes/ways
___
dev mailing list
dev@openstreetmap.org
http
Scott Shawcroft wrote:
Please let us know what you think. We firmly believe that distributing
the data over a number of computers is a far better solution than one
single supercomputer.
This conclusion (divide and conquer) is right for fetch. What was your
update performance?
Did you
Scott Shawcroft wrote:
Our update performance shouldn't be too different. We simply send the
update request to all the node machines.
And your node machines do not cache their partition results? (Thus is a
scan always required?)
By within do you mean a bounding box query? Could you be
ls -1 | while read i; do
grep -v tag $i | grep -v /way | grep -v /osm /tmp/toupload/$i
sort -u $i | grep tag | grep -v 'k=created_by' /tmp/toupload/$i
echo 'tag k=created_by v=OSM Fixer/' /tmp/toupload/$i
echo ' /way' /tmp/toupload/$i
echo '/osm' /tmp/toupload/$i
done
The following
Ceriel Jacobs wrote:
On Tue, 17 Mar 2009 18:35:18 +0100 Stefan de Konink wrote:
I have set this up completely inside a webserver. Cherokee, no php, just
the python script that renders the thing on demand. Runs in production
for months now.
snip
vserver!10!domain!1 = tile.openstreetmap.nl
On Thu, 19 Mar 2009, Grant Slater wrote:
Patched planet files will become available shortly, but they too take
time to generate.
Are you saying you are going to regenerate all previous planet files?
Isn't that a bit waste of resources?
Stefan
___
On Thu, 19 Mar 2009, Shaun McDonald wrote:
Isn't that a bit waste of resources?
From a legal perspective no.
Please; why don't we just ignore all previous planets and create a
new planet and start to produce daily (full) history? Where we dump our
old history excluding Teleatlas data
On Thu, 19 Mar 2009, Brett Henderson wrote:
Osmosis will provide the capability to do this from 0.6 onwards. If
disk space isn't an issue on the planet server I'll set it up as soon as
I get time after 0.6 goes live. It has the ability to go back in time
and produce full history diffs for
http://openstreetmap.org/api/0.5/node/344793814
As you can see here andnav seems to allow 'null' tags. I'll add this to
the OSM fixer scripts so it will get removed. Likewise for stuff like
this:
http://openstreetmap.org/api/0.5/node/344020916
Most likely I'll run an OSM Fixer session today to
Shaun McDonald wrote:
Sorry if that wasn't clear. I meant when the API 0.6 goes live and the
database is upgraded to the 0.6 schema then the Osmosis tasks supporting
0.6 data will support creating full history change files. The Osmosis
version is currently 0.30.2, I'll release a 0.31
On Thu, 19 Mar 2009, Ceriel Jacobs wrote:
I suggested to Grant to actually do a benchmark instead of just
picking
the next product from the stack.
When benckmarking different database systems, don't forget to include
Virtuoso.
Here a TPC-D benchmark comparison (lower is better):
1.
Udo Giacomozzi wrote:
Instead I'm thinking about writing some multithreaded FreePascal or
C++ FastCGI server that does it much more efficiently (by using
libmapnick internally) but I still need to have a look at mod_tile as
it might already provide a highly performant solution.
Investigate
Udo Giacomozzi wrote:
Hello Stefan,
Friday, March 20, 2009, 12:55:27 AM, you wrote:
Instead I'm thinking about writing some multithreaded FreePascal or
C++ FastCGI server that does it much more efficiently (by using
libmapnick internally) but I still need to have a look at mod_tile as
it
Grant Slater wrote:
Recent planet files available from: http://planet.openstreetmap.org/
will be disrupted for the next few days. Apologies for the disruption.
Hopefully be fixed within the next 48 hours.
Reasons? Look over here:
Udo Giacomozzi wrote:
I just set up my own Mapnik server that renders map tiles on demand
and saves them on disk so that they are efficiently cached. Currently
this is done using a PHP script (FastCGI handler with 5 instances)
that handles 404 errors and invokes a python script to render the
Jon Burgess wrote:
On Wed, 2009-03-18 at 00:48 +0100, Stefan de Konink wrote:
Dane Springmeyer wrote:
I should stress though that Mod_tile is the optimal solution for tile
rendering for most OSM related purposes.
I wonder if an Apache module is able to provide caching. I doubt it can
Hi all,
On many requests I have created an osmparser.c that instead on real
files exclusively operates on stdin. Why would this benefit you? Most
multiprocessor systems do nothing with their second processor, so in
this case one processor can do bzip2 -d -c theplanet.osm.bz and the
other
Robert (Jamie) Munro wrote:
Yes. Patents are the right to exclude others from doing something that
was your idea.
Thankfully anything related to mathematics cannot be patented in Europe
:) So @#$* the USA :)
Stefan
___
dev mailing list
Robert Scott wrote:
On Sunday 15 March 2009, Stefan de Konink wrote:
You can extend it to NURBS without breaking tools ;) Think about the
control point approximation ;) So every embedded device that isn't fast
or doesn't have a fpu will use the ugly linear version,
And more people
Udo Giacomozzi wrote:
Is there a good reason why all map databases seem to be fixed on
straight lines and do not support some kind of bezier curves? After
all, streets most of the time have curves and only at junctions or in
other rare situations have corners.
Using bezier curves would at
Udo Giacomozzi wrote:
Saturday, March 14, 2009, 4:44:10 PM, you wrote:
SdK Basically my storage model thinks everything in simple linear lines.
SdK Adding weight will give you NURBS ;) Infinite weight gives you an
SdK absolute point. I guess you get the idea now.
Not sure I really
Udo Giacomozzi wrote:
Of course I realize that now the OSM data format can't be extended to
curves without breaking existing renderes and tools...
You can extend it to NURBS without breaking tools ;) Think about the
control point approximation ;) So every embedded device that isn't fast
or
rajan vaish wrote:
I am Rajan,a CS major from India.I interned online for One Laptop per
Child creating a Geography teaching tool using GeoRSS,MapServer and
OpenLayers in summer'08. I am really looking forward to participate and
work for OSM this year and there are few things of my interest
rajan vaish wrote:
Hi Stefan,
I created a *web* based tool actually,the desktop version was for
Windows (since there was a news of Microsoft showing interest in OLPC
with Windows ) ,this is the tool,I produced
http://vaish.rajan.googlepages.com/Atlas.exe ,make sure that you have
rajan vaish wrote:
Though,I really don't understand why will there be a problem when using
a Sugar version or for that sake any platform,since OSM is something web
based too and is basically browser dependent (Firefox in this case) ,or
may be I am not getting your question properly,I own an
Grant Slater wrote:
The API downtime scheduled for the 0.6 API transition has been postponed
due to delays acquiring the new database server.
So it is impossible to buy a machine for 15k? Only one response: wow!
Stefan
___
dev mailing list
Grant Slater wrote:
Large imports in the pipeline.
Partitioning is a scalable solution to that, not buying new hardware.
Now it is nice you put 32GB (extra expensive) memory in there, but
most likely your hot performance would be far better with more (cheap)
memory than more disks. At the
Matt Amos wrote:
At the time I wrote my paper on OSM Dec2008, there was
about 72GB of CSV data. Thus with lets say 128GB you will have your
entire database *IN MEMORY* no fast disks required.
in 8Gb kits? that would be *extra* expensive (about £8,680 according
to froogle).
Some people are
Stefan de Konink wrote:
Wow... (serious wow) I have never seen the database THAT expanded unless
I was using an XML database.
And now I think of it; that is probably because *I* wasn't able to
download the history tables. That makes sense; but does it make sense to
have the history tables
Grant Slater wrote:
But as detailed below by Stefan, the internal block fragmentation is a
serious issue, which needs to be fixed first.
I am also still very sceptical about SSD MTBF on DB server load levels.
Write 1 bit = Full SSD block write.
Big community site in NL reported less than a
On Mon, 9 Mar 2009, sly (sylvain letuffe) wrote:
None does download correctly.
Most likely the 500 is there because the Rails script doesn't reply a
valid content-type. Thus the error is not really passed through.
Stefan
___
dev mailing list
I haved heard some French people were able to use their governments data
from a WMS where a WFS was not available. In The Netherlands we have a
situation like it; now since I hate tracing since kindergarden I
proposed several solutions for the GPX/NMEA problems many years ago in
the form of
Dick Marinus wrote:
Well, I don't have any questions but I'm open to suggestions :-)
Please let me know If you like the idea.
Hey :D
Great to see you here :) Yes I love the idea especially as a read only
variant. You might have figured out what server we use for mirroring,
and maybe just
Frederik Ramm wrote:
Dick Marinus wrote:
I'm currently investigating if I could create an OSM server by only
using HTTP.
The OSM server *is* using only HTTP.
No he means *ONLY* http, so without databases or altered files.
Stefan
___
dev mailing
Richard Fairhurst wrote:
In brief, one of the performance problems we have is that when you ask
for all the data within an bbox, Rails retrieves all the nodes from
the db in one hit, creating an object for every single node. If you
have 5000 nodes in one bbox, that's quite a memory
Frederik Ramm wrote:
Iván Sánchez Ortega wrote:
all this should be mostly handled by the db server with appropriate
foreign key constraints, not by the application code
+1
Rails guys: we're using InnoDB tables for a reason, you know.
I'm not a foreign keys guru myself but I think it may
Iván Sánchez Ortega wrote:
El Miércoles, 11 de Febrero de 2009, Stefan de Konink escribió:
I'm not a foreign keys guru myself but I think it may be difficult to
have one with the relation_members as designed currently.
I have solved that by splitting the members in 3 distinct tables
201 - 300 of 564 matches
Mail list logo
