Cloud Computing GIS and Standards (OGC/ISO)

There is a discussion going on on LinkekIn about GIS and cloud computing.

I responded to the posts and include my comments (which should be read in light of the threads themselves) as follows. The reason is because I have been thinking about the failure (or silence) about whether any update functions (eg the variously named, non-standard, standard functions – ST_UpdateVertext, ST_DeleteVertex, ST_InsertVertex) may ever arrive from the evolution of the Simple Features Specification standards for SQL from the OGC or ISO.

Don’t assume anything. I get the impression from the posts here that many think that OGC standards have done wonderful things and solve all problems….. to an extent, yes, but to a larger extent, no.

GIS/OGC/ISO standards are built external (remotely) to real users and real deployment environments with little feed back processes ie “transparency”.

Wrt the first, how many non-academic, non-commercial representatives are there on OGC/ISO standards bodies? Why does the ISO restrict the availability of its standards documents to those who can pay (gee, a perfect to tell everyone you don’t understand the WWW – Creative Commons standards anyone?)? If you listen to users of OGC standards “compliant” software you quickly discover what is missing in the standard. Since the bodies have no way of talking with these people, many who discover standards API shortfalls conclude that the standard was written to dumb down what is possible by those with a lot to gain.

Some of the truly good work in the GIS world has been done by open source projects and people: yet they don’t appear to be listened to (nor their users). Do we need a new concept in standards development? The idea of open sourced standards? Crowd-sourced standards built by inclusive consensus models of development rather than a “pay to enter” model that is restricts the process to vendors and large bodies like our CSIRO (in Australia)?

Wrt to David’s [Sonnen] comment on database spatial. It is not in the slightest a lovely, warm, fuzzy, integrated, homogeneous world out there. Some vendors demand storage as WKB (not required by any OGC/ISO standard or Relational Database Theory) and it seems to have traction only because the type systems of the different databases are such a mess….. Anyway, cross-database spatial is a triviality next to the fundamental architectures of the host database products!

No update functions for Simple Features in SQL…… Huh!

The topology “standard” from the ISO/OGC appears to be universally unacceptable.

The lack of a raster processing SQL standard sees proprietary implementations… 3D Solids/Tins anyone?

It goes on.

Wrt the second issue, it is probably as it should be in one sense eg the TCP/IP standard didn’t need a business model underpinning it or making it “relevant”. But not in another. Which is the very great reality that SOA is not a product. In only a very tiny number of businesses does one find strong standards (or homogeneous technology) alignment. What one finds is evolution based on what is practical, experiential and affordable. A decision to move to the cloud may have a spatial component but it is so minuscule that one should not overly give it too much importance (which those who work for the OGC etc I guess can’t afford). Ensuring flexibility in the cloud is where OGC may become a critical resource but since OGC works through consensus, the geospatial user community (which includes the far wider and more complex IT community) needs to be deeply involved in OGC processes.

I can’t see it happening.

In summary, the OGC and ISO Spatial standards groups need to become more open to real users. It also needs to understand that “spatial isn’t special” which will drive open-ness to IT communities and “ground” their understanding of what can be achieved. Finally, it needs to work on stuff that is far less sexy: the maturation of existing standards that need a lot more work and urgently to become better and more indispensable.

Just my 2c worth.