ISO 19650 in practice: naming and attributes people actually fill in

The most perfect ISO 19650 naming convention in the world is useless if people do not fill it in.

I configured the CDE for a large scale airport infrastructure program on Forma: folder structure, permissions, attributes, review workflows and the container naming convention under ISO 19650. And the most valuable lesson of the whole project was not technical. It was behavioral.

Why naming conventions die

The pattern repeats in almost every project: a committee designs an impeccable convention, with twelve fields and every possible combination, and publishes it in a PDF. Three weeks later, half the files carry invented values and the other half have empty fields. The standard exists on paper and nowhere else.

The problem is not the people. It is that the convention was designed for the auditor, not for the person uploading twenty documents on a Friday at 6 pm.

The principles that actually work

Fewer fields, better filled. Every extra attribute is one more reason to invent a value or leave it blank. A seven segment convention that gets respected beats a twelve segment one that gets ignored.

Closed lists over free text. If the field accepts anything, sooner or later it will contain anything. Every segment of the name and every attribute should come from a defined list: disciplines, document types, statuses.

Validate at the door, do not fix afterwards. A badly named container that already entered the CDE has polluted searches, views and review workflows. Validation has to happen at upload time, not in a quarterly audit.

The attributes do the heavy lifting. The name identifies; the metadata filters. Fighting to cram all the information into the file name belongs to the last century: that is what CDE attributes are for, feeding views and filters without overloading the name.

Where the standard lives

ISO 19650 is not a manual you print: it is an operating agreement that lives in the CDE configuration. On Forma that means the naming standard configured in the project, the defined attribute lists, the folder structure with its permissions and the review workflows that reject non compliant containers at the door.

And like any critical configuration, it is governed: changes are tried in a test hub, validated and only then replicated to the production template. A standard that every project interprets its own way stops being a standard.

The measure of success

The metric is not the elegance of the convention: it is what percentage of the project's containers can be found by filtering on attributes. If the standard only exists in a PDF, it does not exist.