[Mapbender-users] FOSS GIS Web Service Client-Side products: MapBender

[Arnulf Christl]

John Roberts wrote:
> Dear All
> 
> I am currently studying for a Master's degree in GIS with UNIGIS. I am
> currently doing my dissertation which evaluates the main FOSS GIS web
> service products, both server-side (GeoServer, MapServer, Deegree Web
> Services & MapGuide OpenSource) and client-side (Deegree IGeoPortal,
> MapBender, MapBuilder & OpenLayers).
> 
> I would like to compare and contrast the four client-side products using the
> OWS specifications. I have taken the main part of my dissertation and
> created a document which:
> 
> 1. summarises the OWS specifications (chapters 1 to 4).
> 2. summarises the organisational structure and functionality of the four
> client-side products (chapter 5).
> 3. creates a matrix based on the OWS specifications against which each
> product is measured as to whether it is able to generate the OWS request and
> parameter and whether it is able to receive and process the OWS response
> (chapter 6).

Hi,
three of those are already OSGeo projects and you might find some additional information there. A tentative try at categorizing Web Mapping Platforms:
http://wiki.osgeo.org/index.php/Choosing_a_Web_Mapping_Platform

> I have found the exercise of creating a matrix harder with regard to the
> client-side software as opposed to the server-side software. MapBender

Diversity in clients is also a lot higher than on the server side. Server technology can be abstracted a lot more. Perspectives on distributed Web Mapping make the distinction between client and server fuzzy if not arbitrary altogether (MapServer is a client to PostgreSQL, GoogleMaps is a GDAL client...). 

> documentation describes the controls and the functionality that they
> provide. However a particular control and functionality is not related to an
> OWS request and the HTTP method used (i.e. is it the MapFrame control that
> generates the GetMap request and what parameters is it using?).
> 
> I think that this would be useful as it would allow a direct comparison of
> server-side and client-side software compatibility. For example, if one were
> to know exactly what OWS requests and parameters the MapFrame control were
> generating, then the server-side software could be analysed to see if it
> were able to receive this requests and generate a response and hence support
> the MapFrame control.
> It is perhaps that this is not the best method to approach comparing and
> contrasting client-side products and if anybody has any suggestions for
> alternative methods, I would be very grateful.

I think your observation is very precise (see below about about sharing controls). An alternative method to compare the current state of the different products might be to look at the scope. 

* OpenLayers is entirely a client side set of software and implements a variety of specifications, not only OGC
* MapBuilder is mostly client side has some server ties and has comprehensive OGC support
* Mapbender has both server side modules for management and client modules for presentation
The OWS Security Proxy module  is a completely server side module used as a broker for SDI clients like Desktop GIS or other WebGIS clients but it is also part of the software. deegree goes to extremes in implementing all server and client components at the same time. 

At the last FOSS4G MapBuilder, MapBender, ka-map, Cartoweb3, OpenLayers and others met at a BoF to discuss how to make best use of potential synergies. One was seen in providing building blocks (controls) for web mapping software. The most obvious project as a source is OpenLayers. The idea is now to continue OpenLayers development with the goal of using it as *the* map frame for a variety (if all the above) web mapping projects. 

> I have trimmed down the document for MapBender by removing the sections
> pertaining to the other 3 products. I attach the document to this e-mail. I
> would be grateful for any information/comments/criticisms anybody is able to
> provide me with.
> I have highlighted any area I am not sure of in yellow. I am mainly looking
> for input in Chapters 5 & 6.
> I will acknowledge anybody who provides input in the acknowledgements
> section of my dissertation. If the input is straightforward functionality
> then I would use the information and I would not make an explicit reference.
> However, if I wanted to make an explicit reference (for example if I am
> referring to a more subjective comment with regard to the relative pros and
> cons of the 4 client-side products) then I would contact the person again to
> ask his/her permission. I am going to post similar documents for the other 3
> products on their mailing lists. Once everything is complete, I would like
> to post the final document with information on all four products on the four
> mailing list.
> 
> Thank you all very much in advance for any time you are able to give.
> 
> Kind Regards
> 
> John Roberts
> Manchester, UK

It would be great if you would protect your dissertation with an Open and Free license, then OSGeo would be happy to provide a link to that document or the document itself for download on their site (lots of fame and glory for you). 

I read your document, it looks good. You might consider using OpenOffice.org document format until Microsoft Word also supports an open standardized format (as you are writing about both standardization and Open Source :-)

We will have a closer look at the looong table and if it makes sense we will copy it to the Mapbender Wiki and keep it up to date there. Maybe it is a useful perspective on how to do things (what is your time line?).

Yet another comment... The perspective OGC has on spatial data and how it is processed is obviously that of an important standardization body (read: !STANDARDIZATION BODY!) and that of a business club (look at the long list of commercial members). Sometimes this results in rather theoretic and static or mightily proprietary smelling and in some cases this is blunt painful.  Mind me, I say this as avowed Princial Member of the OGC and mother to a strictly OGC following software project. Not all shiny specifications are gold or something like that. 

Best regards, 
Arnulf.