To this day, at least in the geospatial environment, every professional with neutral thinking recognizes that free software is as mature as commercial software, and in some respects superior.
The standards strategy worked very well. Although its balance of updating in the face of the energy required by technological evolution is questionable, perhaps it was the basis for ensuring success in other efforts such as community, philosophical approach, economics and other ideas that were used to justify the model, Which are also necessary.
However, selling Open Source solutions is not easy in business or government environments, for many reasons that partly originate in competition but also as an inevitable result of the weaknesses of the model, which must evolve and coexist with proprietary software. Decision makers are asked questions like:
If one morning we see a problem resulting from updates from other platforms, in aspects such as security Who responds to the moment we need support, and at what price to leave it budgeted?
Given the range of alternatives in language, libraries, client solutions, web solutions, what combination should we choose to ensure compatibility quasi total?
OpenGeo Suite is a solution that not only takes advantage of all the maturity of the available tools, but also aims to respond to these weaknesses of the model. In addition to giving the community a solution with which they can boost their development initiatives, it creates a common thread for the components involved to guide their evolution and, for companies OpenGeo Suite provides the seriousness that requires deciding for open source. While there are other companies, after a while of trying this alternative I have no other than to recognize the high capacity and creativity of the thinking behind Boundless, the company that created this solution.
Let's take a look at the OpenGeo Suite approach:
What tools does OpenGeo Suite include?
Having so many options for solution is not bad, it is normal, although it complicates somewhat how to ensure the selection of tools in integral production processes. An erroneous selection can be costly if we realize when we have already invested in research, development, training and above all unrecoverable time.
For example, only in language development we have a puzzle resulting from the needs of the community, many of them doing exactly the same, others emulating in another flavor, some with unique characteristics in simple routines that we would like to have them all. Let's look at this separation by functionalities and languages; Although I must be honest, categorization is not exclusive and in some cases it is difficult to distinguish the border:
- At the client level, the most popular context is: QGis, Grass, ILWIS, SAGA, Kapaware, based on C ++. gvSIG, Jump, uDIG, Kosmo, LocalGIS, GeoPista, SEXTANTE, based on Java. MapWindow by its side on ActiveX based on .NET.
- In libraries we have: GDAL, OGR, PROJ4, FDO, GEOS on C ++. GeoTools, WKB4J, JTS, Baltic based on Java. NTS, GeoTools.NET, SharpMap on .NET.
- As for data bases, Postgres is the dominant indisputable, although there are other solutions.
This shows us that it is possible to mount a system almost in any environment. Moreover, many of them, although they were born in a language, now support others. Also many of them were born as a client but are able to manage web data and in cases like Open Layers it is possible to develop in a web environment almost everything that is done in a client tool.
What combination of free software use?
OpenGeo suite was decided by Qis As a desktop client, who at this point already deserves a category of articles in Geofumadas. For web chose GeoServer as data server that operates on Tomcat, Jetty as a Java runtime environment, GeoWebCache for the tessellation and OpenLayers as a library although the latter option does not have a license plate considering solutions such as Leaflet that is growing with great success especially by its model Based on Plugins and its potential with mobile applications. You can see that they could go by a single line of language but I would like to see the matrix of analysis that has led to this definition.
Let's be clear, anyone can implement these solutions per individual. What OpenGeo contains is an installer with versions of these components with improvements for efficient tedious routines; for example:
- The installer makes the assembly neatly. Being able to choose which components to install, remove or uninstall. For those who have dealt with a Java runtime engine with the blissful 503 Error will know the utility.
- There are different installers: Windows, Mac OS X, CentOS / RHEL, Fedora, Ubuntu and Application Servers.
- The recent version 4.02 brings PostgreSQL 9.3.1, PostGIS 2.1.1, GeoTools 10, GeoServer 2.4.3 and GeoWebCache 1.5; And supports OpenLayers 3.
- In the start menu you create direct links to stop or start GeoServer and Postgres; Also to raise the user interface of data loading shapefiles to Postgres (shp2psql) and also to access the PostGis database (PgAdmin).
- Also in the start menu there is an access to the localhost, which in this version eliminates the client interface of the 3 version, with a clean control panel towards the services GeoServer, GeoWebCache and GeoExplorer.
- This product, GeoExplorer is an impressive development of Boundles based on GeExt that acts as data viewer for GeoServer, allows the upload of data from a local file or from a data warehouse, being able to configure color, line thickness, transparency, Labeled, including rules and saving directly to the geoServer style file (sld). No one in their right mind works this to pure code and GeoExplorer is an excellent solution -Although it does more things-.
- The installed version of GeoServer includes a link to the import of data, and you can create sources from local shape layers, including PostGis so that you can move data from one base to another included from Localhost to a hosted service; It is interesting that this data upload solves problems OGR2OGR that unless they are done with console line, throw difficulties when you upload a multipolygon layer, because the default is simple polygon.
- In this case, the WPS services appear because in the option to install I decided to integrate them.
- At the time of installation you can add GeoServer add-ons such as CSS Styling, CSW, Cloustering and support for GDAL image libraries. There is also an Add-on for PostGIS that supports point clouds on the database and as a client you can also install GDAL / OGR. For developers there is option to install Webapp SDK and GeoScript.
- Unlike in my version hosted on the server, I see that there are more possible data sources, which can surely be added, but in the case of the version that comes with OpenGeo Suite it brings comma delimited text, H2, H2 JNDI, SQL Server, OGR, Oracle and a handful of possibilities in raster backgrounds.
What about Qgis?
- Of the best, for Qgis they created a great plugin called OpenGeo explorer with which you can interact with the Postgres base and also with GeoServer. From here you can edit the slds, move layers, layer groups, edit names, delete, see workspaces, cached layers, etc.
- If a layer is deleted, the sld is deleted; All this is configurable and in the end gets a job from the client controlling what is up, that synchronization can be using the REST API.
- For now, what it does not have is shp2psql but I'm not surprised that it is integrated into the same panel, maybe as transparent as the Spit plugin, which unlike the UI stores the connections, you can upload several layers in block, the progress bar is more Realistic and most understandable error messages.
With this OpenGeo Suite is not saying that this is the magic recipe. But it will surely move much of the community to this preference, especially because companies that sell courses will prefer to teach this route that guarantees a shorter learning curve.
The combo is compatible with other tools that can be mounted on the server.
What impact comes with OpenGeo Suite
We will see what impact this has on the community, because behind Boundless there are people with a lot of experience in the field, who has been involved in the development of tools and libraries that now make the sector sustainable. But above all with grouting in entrepreneurship and marketing services, which is often wasted from the technical level. To mention at least six:
Eddie Pickle and Ken Bossung, founders of IONIC, which bought ERDAS on 2007 and is now owned by Leica.
Andreas Hocevar and Bart van den Eijnden, who were immersed in the development of OpenLayers 2 and GeoExt.
Victor Olaya, who left us that legacy of SEXTANTE,
Paul Ramsey, of the early initiators of PostGIS.
The other positive impact is in the formality of a large company, which, apart from becoming a monster in the market - which is always a risk - provides a formality for competition against private sector companies in aspects such as support, credibility, security and Quality control over developments.
Boundless's offering of services, ranging from platform migration to annual support services, seem consistent with the corporate and institutional market that gradually understands the difference between having local support and business support. This market should not be easy, but we see with good eyes how institutions mature in thinking, valuing software development and information as an asset, thus managed to move from assigning automotive mechanics to their drivers, to contract specialized insurance and services Of the distribution companies.
In the free code model, there is opportunity for all. So what Boundless offers, it's there, with an opportunity for To be a partner; there the ability of those who wish to enhance their ability to sell services in terms of implementation, training, support or development. The example we find valuable and good lessons to learn and complement the effort that the gvSIG Foundation takes, which we will talk about on another occasion.