[Geoserver-devel] Upcomming Geoserver Release

Chris,

Yes, the two big things I'm planning on for the next release are:

1) SRS support (for the WMS)
      I dont expect this will be too difficult, but there's two issues:
      a) Geotools CRS support; I think this has improved very quickly,
         but Martin has said that we're losing a lot of information
         by using the .properties file instead of the actual epsg
         database. The database isnt integrated into geotools yet
(there
         is a JIRA for it, though). The implication is that our SRS
         xforms could be "poor" because we dont have access to all the
info.

      b) It also means that *every* layer will have to have a correct
         SRS associated with it. I was thinking about having the
         FeatureType config in Geoserver throw an error/report if the
         SRS wasnt valid (or looked suspicious), but that might be a bit
         too aggressive. I'm just trying to get users fix it before
they
         ask "why isnt my WMS working?".

2) GEOSERVER_HOME

     You've seen a few messages about this recently. I'm going to do a
bit more research and then do something simple - probably something in
the web.xml and the "-D..." option.

You also mentioned about documentation being spread between the "html"
documenation and the Wiki documentation. I've been sticking most of
the new documentation in the Wiki because it much easier to add,
integrate, and update. Plus, users get to make changes to it.

I'm sure you're one of the few people who use's your WEB-based server's
when you're not on the web. I can included a site-snaged version of
the wiki with the distribution, but it kinda defeats the purpose of the
wiki.

In particular, the SRS help page, i believe, will be heavily frequented
by users since it is a bit confusing. I was originally going to put a
help page and another page with all the actual SRSs in it. I thought
having two links might be confusing, but its probably a good idea.
I'll make the list of actual SRSs inside geoserver so you dont need the
web to access it.

dave

----------------------------------------------------------
This mail sent through IMP: https://webmail.limegroup.com/

dblasby@anonymised.com a écrit :

a) Geotools CRS support; I think this has improved very quickly,
   but Martin has said that we're losing a lot of information
   by using the .properties file instead of the actual epsg
   database. The database isnt integrated into geotools yet (there
   is a JIRA for it, though). The implication is that our SRS
   xforms could be "poor" because we dont have access to all the info.

Geotools is already capable to connects to an EPSG database, but the user has to download and install the database himself on his machine. The Access database is the easiest (it is just a matter of setting up an ODBC datasource in the Windows Control panel). PostgreSQL work too, but need a little bit more steps to install.

Note that it is possible to avoid the need to install an EPSG database on the user side if connection to a remote server is acceptable. User could connect to a distant PostgreSQL database.

GEOT-46 in JIRA is just about embedding the EPSG database in Geotools as an HSQL database. It should not be very long (providing we have a volunter) since it is not about adding the capability to read an EPSG database; Geotools already has this capability.

  Regards,

    Martin.

Le Vendredi 8 Avril 2005 01:18, Martin Desruisseaux a écrit :

dblasby@anonymised.com a écrit :
> a) Geotools CRS support; I think this has improved very quickly,
> but Martin has said that we're losing a lot of information
> by using the .properties file instead of the actual epsg
> database. The database isnt integrated into geotools yet (there
> is a JIRA for it, though). The implication is that our SRS
> xforms could be "poor" because we dont have access to all the info.

Geotools is already capable to connects to an EPSG database, but the
user has to download and install the database himself on his machine.
The Access database is the easiest (it is just a matter of setting up an
ODBC datasource in the Windows Control panel). PostgreSQL work too, but
need a little bit more steps to install.

Note that it is possible to avoid the need to install an EPSG database
on the user side if connection to a remote server is acceptable. User
could connect to a distant PostgreSQL database.

GEOT-46 in JIRA is just about embedding the EPSG database in Geotools as
an HSQL database. It should not be very long (providing we have a
volunter) since it is not about adding the capability to read an EPSG
database; Geotools already has this capability.

We already talk of that on the begining of june 2004 : I provided scripts for
getting EPSG database under HSQLDB (EPSG_v65_4_HSQL.tgz was attached to the
mail - subject: HSQL). At that time, the development of GT2 using GeoAPI
prevented to use this. The HSQL database weights 4M.

I also used the EPSG database under postgreSQL with the following
/etc/.java/.systemPrefs/org/geotools/cs/prefs.xml :
-----
<?xml version="1.0" encoding="UTF-8"?>

<!DOCTYPE map SYSTEM 'http://java.sun.com/dtd/preferences.dtd’>

<map MAP_XML_VERSION="1.0">
  <entry key="EPSG Factory"
value="org.geotools.cs.CoordinateSystemModifiedEPSGFactory" />
  <entry key="EPSG connection"
value="jdbc:postgresql://localhost:5432/epsg?user=epsg_reader&amp;password=*****"
/>
  <entry key="JDBC driver" value="org.postgresql.Driver" />
</map>
-----
I don't know if it is working again with the new GT2 ?

  Regards,

    Martin.

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_ide95&alloc_id396&op,ick
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel

regards,

didier

Quoting dblasby@anonymised.com:

Chris,

Yes, the two big things I'm planning on for the next release are:

1) SRS support (for the WMS)
      I dont expect this will be too difficult, but there's two
issues:
      a) Geotools CRS support; I think this has improved very
quickly,
         but Martin has said that we're losing a lot of information
         by using the .properties file instead of the actual epsg
         database. The database isnt integrated into geotools yet
(there
         is a JIRA for it, though). The implication is that our SRS
         xforms could be "poor" because we dont have access to all
the
info.

At this point you know this stuff better than I, but my original thought
was to try to give users options. Properties file for defaults, but
then allow them to connect to a database if it's available. Or to
download an embedded database jar. Could we maybe do that with
hibernate? An all in one jdbc connection and epsg database? Put it as
an extra, not in the main distribution. But it would also be nice to
use say the PostGIS srs table if there is a postgis datastore defined.

      b) It also means that *every* layer will have to have a correct
         SRS associated with it. I was thinking about having the
         FeatureType config in Geoserver throw an error/report if the
         SRS wasnt valid (or looked suspicious), but that might be a
bit
         too aggressive. I'm just trying to get users fix it before
they
         ask "why isnt my WMS working?".

Well, I'm definitely of two minds of this. It kind hits right between
two of our stated goals: 1) support and encourage standards, and 2) be
as nice to users as possible. If we go for the former, then we don't
let users submit bad data, invalid srses and all. We don't want users
sharing bad data on the internet - I feel like this is one of Paul's
pet peeves about ims servers, they have bad coordinate system stuff,
because they make it so users don't have to configure. Ok, I think I
talked myself into going with 1). We just need to make it easy, have a
nice migration strategy, explain everything upfront, apologize, and put
that faq question first (well, second after INSTALL JAI!).

2) GEOSERVER_HOME

     You've seen a few messages about this recently. I'm going to do
a
bit more research and then do something simple - probably something
in
the web.xml and the "-D..." option.

Cool. I was thinking about this too, it's not worth spending a whole
ton of time on, getting it exactly right. Get it working reasonably
well for this version, and we can learn from it and apply in geoserver
2.0, when we get to redo as much config as we want. We're still bound
at this point by being backwards compatible. So yeah, just a minimum
to make it possible, we don't need to migrate users en masse,
perfectly, at this point.

You also mentioned about documentation being spread between the
"html"
documenation and the Wiki documentation. I've been sticking most of
the new documentation in the Wiki because it much easier to add,
integrate, and update. Plus, users get to make changes to it.

I'm sure you're one of the few people who use's your WEB-based
server's
when you're not on the web. I can included a site-snaged version of
the wiki with the distribution, but it kinda defeats the purpose of
the
wiki.

I'd disagree with that statement. This of wiki as the source code, and
the site snagged version as a release. Just like for code releases we
should clean it up before releasing. Online users can get at the
latest version, others can get the released version. Just be sure that
the site snagged version looks good.

In particular, the SRS help page, i believe, will be heavily
frequented
by users since it is a bit confusing. I was originally going to put
a
help page and another page with all the actual SRSs in it. I thought
having two links might be confusing, but its probably a good idea.
I'll make the list of actual SRSs inside geoserver so you dont need
the
web to access it.

You also could consider putting it on the sourceforge site, so you
wouldn't have the confluence problems.

Chris

dave

----------------------------------------------------------
This mail sent through IMP: https://webmail.limegroup.com/

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real
users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel

----------------------------------------------------------
This mail sent through IMP: https://webmail.limegroup.com/

Richard didier a écrit :

We already talk of that on the begining of june 2004 : I provided scripts for getting EPSG database under HSQLDB (EPSG_v65_4_HSQL.tgz was attached to the mail - subject: HSQL). At that time, the development of GT2 using GeoAPI prevented to use this. The HSQL database weights 4M.

I remember that very well; I still have the mail in front of my eyes every days (it still in my incoming box). My problem is that I was hopping a volunter for setting up an HSQL database, test it, apply the needed modification if the SQL syntax is a little bit different (which is often the case with different SQL database), bundle the database in a Geotools module, take the opportunity to recreate the HSQL database with the latest EPSG version (which is now 6.6). I can help with some aspect of that, but was hopping a volunter with commit right as EPSG database module maintainer.

I also used the EPSG database under postgreSQL with the following
/etc/.java/.systemPrefs/org/geotools/cs/prefs.xml :
-----
  <entry key="EPSG Factory" value="org.geotools.cs.CoordinateSystemModifiedEPSGFactory" />
  <entry key="EPSG connection" value="jdbc:postgresql://localhost:5432/epsg?user=epsg_reader&amp;password=*****"/>
  <entry key="JDBC driver" value="org.postgresql.Driver" />
-----
I don't know if it is working again with the new GT2 ?

We are going toward a more javax.sql.DataSource oriented solution, which is the Sun's recommanded way. Thinking again about it, it is possible to avoid completly preference settings (preferences has been reported to be a cause of trouble on Unix/Linux boxes).

  Martin.