Standards conversations in enterprise geospatial procurement tend to get technical quickly, which is a shame, because the argument is commercial. The reason a utility or a port authority should insist on open interfaces is not elegance. It is that proprietary data formats and closed service endpoints convert a software decision into a permanent one.
The lock-in mechanism
Lock-in in this sector rarely arrives through the application. It arrives through the data. A network model held in a proprietary geodatabase with vendor-specific topology rules can be exported, technically, but the export loses the connectivity semantics that took three years to build. The migration cost is not the file conversion; it is rebuilding and revalidating relationships that were encoded in a format only one vendor reads.
Once that is true, renewal negotiations have a predictable shape.
What to write into the contract
- Data must be exportable in full, including topology and relationship semantics, in a documented open format — and the export must be demonstrated during acceptance testing, not promised.
- Services must be published through OGC interfaces so that any compliant client can consume them without a vendor-specific connector.
- The schema must be documented and owned by the buyer, with the vendor implementing against it rather than the reverse.
- Test the exit at least annually. An export capability that is never exercised is an export capability that has quietly broken.
Ask any vendor to demonstrate how you would leave. The quality of the answer tells you most of what you need to know.
The counter-argument, fairly stated
Proprietary formats are not arbitrary. Vendors optimise storage and processing in ways that standards bodies, moving slowly and by consensus, do not always match. Insisting on a lowest-common-denominator interface can cost performance, and for very large raster or point-cloud workloads that cost is real.
The reasonable position is a boundary rather than a rule: use whatever internal representation performs, and require that everything crossing the boundary — into storage of record, out to clients, across to other systems — does so through open, documented interfaces. Internal optimisation is the vendor's business. The boundary is yours.
Why we argue against ourselves here
Kaiinos builds on open standards throughout, and we say this to prospective clients directly: the architecture we propose is one you could migrate away from. That is not modesty. Buyers who know they can leave commit faster and expand further, and we would rather compete on whether the platform is good than on whether escape is feasible.
