tech-trend-watch

Long-Term Technology Watch

Data Sources

Source	Cadence	URL	What it measures
State of JS	Annual (autumn)	https://stateofjs.com	JS libraries, frameworks, tools — usage + satisfaction
State of CSS	Annual (autumn)	https://stateofcss.com	CSS tools, frameworks, features — usage + satisfaction
Thoughtworks Tech Radar	Quarterly	https://www.thoughtworks.com/radar	Broad tech (langs, platforms, tools, techniques) — ring assignments

Fetch the latest edition of each before starting the review. State of JS/CSS also publish a public JSON API:

https://assets.devographics.com/surveys/js<YYYY>/en-US/results.json
https://assets.devographics.com/surveys/css<YYYY>/en-US/results.json

Step 1 — Inventory current stack

List all significant dependencies from package.json:

jq -r '(.dependencies // {}) + (.devDependencies // {}) | keys[]' package.json | sort

Focus on: frameworks, bundlers, test runners, type checkers, linters, CSS tooling, state management, routing, data fetching, animation. Exclude low-level utilities and polyfills from the review.

Step 2 — State of JS/CSS mapping

For each technology in the current stack, locate it in the State of JS/CSS results.

Satisfaction × Usage matrix

                  Low usage        High usage
                ┌────────────────┬────────────────┐
High            │   EMERGING     │    STABLE      │
satisfaction    │  (consider     │  (keep, likely │
                │   adopting)    │   healthy)     │
                ├────────────────┼────────────────┤
Low             │   AVOID        │    LEGACY      │
satisfaction    │  (don't start) │  (plan exit)   │
                └────────────────┴────────────────┘

Thresholds (State of JS survey): high ≥ 70%, low ≤ 55%. Values between 55–70% are borderline — check year-over-year trend to break the tie.

STABLE (high usage + high satisfaction): No action. Healthy ecosystem.

EMERGING (low usage + high satisfaction): Evaluate for adoption. Check: maturity, ecosystem completeness, migration cost from current choice.

LEGACY (high usage + low satisfaction): Schedule migration. High usage means inertia is the only reason it's still there. Check: what are people migrating to?

AVOID (low usage + low satisfaction): Do not start new projects on this. If already used: plan removal.

Additional signals

• Year-over-year trend: Is satisfaction rising or falling? A falling trend in STABLE is an early LEGACY signal.
• Retention vs interest: High retention (users stick with it) + high interest (non-users want it) = safe bet.
• "Would not use again" spike: Single strongest LEGACY signal.

Step 3 — Technology Radar mapping

Check the Thoughtworks Tech Radar for each item. Rings:

Ring	Meaning	Action
ADOPT	Proven, low risk, recommended for wide use	Use as default choice
TRIAL	Worth pursuing, proven for specific cases	Try in real projects, gather experience
ASSESS	Promising, explore impact	Spike / PoC only, not production
HOLD	Proceed with caution, don't start new	No new projects; plan exit if already used

If an item appears on the radar, note the ring and the date of first appearance. Items that have been HOLD for 2+ years without movement are candidates for active removal.

Items not on the radar are not necessarily bad — the radar skews toward enterprise/consultancy concerns and lags emerging JS ecosystem tools by 1–2 years.

Step 4 — Cross-reference and prioritize

Combine signals into a 4-level action priority:

Priority	State of JS signal	Tech Radar signal	Action
P0 — Migrate now	LEGACY	HOLD	Create migration PR this sprint
P1 — Plan migration	LEGACY or falling	ASSESS or HOLD	Schedule in next quarter
P2 — Watch	Declining retention	Not on radar	Monitor next annual cycle
P3 — No action	STABLE	ADOPT or TRIAL	Keep as-is

Newly EMERGING tech: do not migrate unless the current tool is P0/P1. Switching costs > marginal improvement unless forced.

Step 5 — Migration candidate evaluation

For each P0/P1 item, estimate:

1. Migration effort: grep for usage count, estimate hours
2. Target alternative: what did State of JS "movers" migrate to?
3. Coexistence period: can old and new run side-by-side during migration?
4. Risk: does it touch runtime (high risk) or build tooling only (lower risk)?

Example evaluation template:

### jest → vitest

- Current usage: 47 test files (grep -r "from 'jest'" src/)
- State of JS signal: LEGACY (satisfaction dropped from 85% → 61% 2021→2024)
- Tech Radar: HOLD (Vol.30)
- Target: vitest (STABLE, Tech Radar ADOPT Vol.31)
- Effort: ~4h (config migration + minor API diffs, @testing-library compatible)
- Coexistence: no (different config, but can migrate per-file)
- Risk: build tooling only → low
- Recommendation: migrate in one PR

Step 6 — Output

Write a review doc with:

## Tech Watch — <YYYY-QN>

### Data source versions
- State of JS: <year>
- State of CSS: <year>
- Tech Radar: Vol.<N> (<date>)

### Stack health summary
| Technology | Category | Signal | Priority | Notes |
|---|---|---|---|---|
| vitest | test runner | STABLE / ADOPT | P3 | — |
| tailwindcss | CSS | STABLE / ADOPT | P3 | v4 migration due |
| jest | test runner | LEGACY / HOLD | P0 | migrate to vitest |
| ... | | | | |

### P0 — Migrate now
- [ ] <package>: <brief rationale> → <target>

### P1 — Plan migration (next quarter)
- [ ] <package>: <brief rationale>

### P2 — Watch (next annual cycle)
- <package>: <signal to monitor>

### Newly emerged (consider adopting)
- <technology>: EMERGING, TRIAL — evaluate for <use case>

Cadence

• Annual: Run full review aligned with State of JS/CSS release (typically November–January).
• Quarterly: Check Tech Radar new volume for changes to existing stack items.
• On-demand (new adoption): Run Step 3 (Radar only) when evaluating whether to add a new library with no existing incumbent.
• On-demand (replacement decision): Run Steps 2–5 for both the current tool and the candidate. Omit Step 1 stack inventory and Step 6 full output doc — produce a single migration evaluation instead. Use this path when the question has an incumbent ("should we switch from X to Y?" or "should we add Y or keep X?").

Boundaries

• This skill covers strategic decisions, not implementation. For executing updates, use dep-lib-review.
• Do not use Tech Radar alone — it lags the JS ecosystem. Always cross-reference with State of JS/CSS.
• "Everyone is migrating to X" is not sufficient reason to migrate. Evaluate migration cost vs. actual pain with the current tool.

Related

• dep-lib-review — executes the actual update (patch/minor/major batching)
• frontend-review-deps — operational dependency audit with CVE triage and trend-watch scripts