Jeddah International College — a bilingual (English/Arabic) CMS-driven marketing site for a Saudi private college, migrated from Next.js 14 to 15 while live. — now migrated to Next.js 15 and documented end-to-end via a 16-part, file-and-line-cited architecture knowledge base.

While fixing the Next.js 15 generateStaticParams bug, the trade-off was returning an explicit empty array rather than using the migration as an opportunity to wire up real build-time static params for the [locale]/[slug] tree. This preserves the exact production behavior already running (pages render on first request, then get cached for 3 days via ISR) instead of introducing new pre-rendering edge cases in the same change that was already touching every route in the tree.

Rather than converting the whole single-education/Index.js Server Component to a Client Component to satisfy Next 15's new restriction on ssr:false dynamic imports, the trade-off was extracting just the two swiper widgets that needed it into a new sibling "use client" file. This keeps the page's CMS data-fetching server-side and isolates the client-only requirement to exactly the two components that actually need it.

The trade-off was removing a hardcoded 1500ms setTimeout before Aos.init() and initializing immediately instead. AOS already hides [data-aos] elements via CSS the moment the page loads, so the delay wasn't smoothing anything — it was only leaving above-the-fold content blank for longer before animations could fire.

The trade-off was hand-copying validation/submit/reset state five times inside one 2,107-line component (states1..5, errors1..5, handleSubmit1..5) for the help-and-support page's five form variants, instead of adopting the generic src/form/useForm.js hook that already exists in the repo. It kept each variant's quirks locally editable without a shared abstraction leaking cross-form assumptions, at the cost of the same email-regex validation copy-pasted five times and useForm.js becoming dead code with zero importers.