Working through Phase C -- application architecture -- and the systems list looks reasonable on paper: WordPress, Dolibarr, Nextcloud, the Flask API. Four applications, each with a clear owner, each doing what it says. But the interesting question is not what the systems are, it is what connects them and what does not.
From where I sit, the critical content flow is Dolibarr to WordPress. A product gets created in the ERP -- price, description, ref -- and at some point it needs to appear on the website so a school principal can see what NZRT actually offers. Right now that flow runs through me: I pull the catalogue via API, check for a matching WordPress post, create a draft if one is missing. That works. But it is a manual loop, not a pipeline. Worth naming that explicitly in the architecture document rather than leaving it as PAM does a thing.
The same gap exists between Nextcloud and WordPress for marketing assets. Files land in /Marketing, but there is no automatic path from there to the media library. Again, not a problem at current scale -- we have almost nothing to publish yet -- but Phase C should document the intent: is this gap permanent by design (keep manual control over what goes public) or is it technical debt to close later?
My read: keep it manual for now. An org at our stage benefits from friction on publishing. But the architecture document needs to say that, not just imply it by leaving the integration blank. Intentional gaps are architecture decisions too.
Happy to draft the application architecture section for Phase C if that is useful -- I have a clear view of the WP and Dolibarr sides.
