Drupal 7: Worth the Three-Year Wait?

Drupal 7.0 shipped January 5, roughly three years after work began. One of open-source’s longer sagas, complete with the soul-searching such waits produce.

The question clients have asked us all week is the fair one: worth it? Our answer, after a launch week of building against the release and a year of tracking its betas: mostly, yes. The ‘mostly’ is worth itemizing, because the two changes that matter most both address complaints we have relayed for years, while the caveats that remain are exactly the ones no release can fix. Here is the working review, from the shop that has deployed every Drupal since 4.7.

The headline fix: an admin experience that doesn’t require apology

Let’s start with the confession behind the celebration. We have lost platform evaluations because of Drupal’s editorial experience, and deserved to. The capability argument always won on paper. Then the content team sat down at the demo, met the administration screens, and chose the friendlier tool. Fairly.

Drupal 7 is the first version built to win that moment. The redesign came through a formal usability project. Actual lab testing at the University of Minnesota. Actual findings, actual changes. The research-driven process our design practice keeps pushing for, applied to the software itself. It shows in the daily screens.

A persistent admin toolbar. An overlay that keeps editing in context. A clean Seven administration theme. Shortcuts for the tasks editors repeat. None dramatic in isolation. Collectively, the difference between a system that tolerates editors and one designed for them.

Our first training sessions on 7 have run noticeably smoother than any Drupal training we have conducted. That sentence is worth more than most feature lists.

The structural fix: fields in core

The deeper change, for the content-modeling readers: the field system, the machinery for defining structured content types with custom fields, moves from the contributed CCK module into Drupal core.

The practical differences. Stability: the architecture Drupal’s superpower rests on is no longer an add-on holding the building up. Consistency: fields attach to everything now. Content, users, and taxonomy terms, one system throughout. And signal: the platform has formally declared what it is for. Structured, connected content, modeled without custom code.

Image handling joins core too. Image fields, styles, and derivatives without the module scavenger hunt that made ‘add a photo properly’ a configuration project.

Longtime readers know our platform shorthand. Describe your content as types and relationships, and if the description strains against posts-and-pages, Drupal belongs on the shortlist. Version 7 makes the shortlist case cleaner than it has ever been, and makes the eventual build faster, because the foundation pieces now arrive pre-assembled.

What persists, candidly

• The learning curve moved. It didn’t flatten. The entry ramp is friendlier. But the concepts that make Drupal powerful still demand real learning. Nodes, and now entities. The render system. The theme layer. The distance between installed and well-built is still measured in expertise. The ‘power tool, rewards training’ verdict from our Drupal 5 review stands, third version running.

• The ecosystem needs its season. The upgrade path’s gating factor, as at every major version: contributed modules. The essential stack is largely ready or close, since the community learned from the 6.0 lag. But the long tail will take months. Any site’s upgrade schedule starts with its module inventory, not with core’s release date.

• Performance still wants configuration. Out of the box, 7 does more per page than it should. Richer defaults cost cycles. The caching setup our checklist prescribes (page cache on, aggregation on, the ladder as needed) remains the difference between sluggish and brisk. Nothing new. Newly worth repeating while launch enthusiasm runs high.

Upgrade advice, by situation

For Drupal 6 sites: no urgency and no deadline panic. Version 6 remains supported under the community’s policy. Support ends when 8 ships, which is years out. The sensible schedule follows module readiness and your own redesign calendar. Upgrading for its own sake buys the admin improvements. For editor-heavy teams, that alone may justify the project.

For Drupal 5 sites: the firmer word. Version 7’s release ends 5’s security support as of this month. Unsupported means unpatched, and running an unpatched CMS in 2011 is a decision with a clock on it. Plan the move this quarter.

For new builds: 7 is the version, full stop, with the usual early-adopter attention to module maturity per feature.

And for anyone weighing the newly interesting question, Drupal 7 versus WordPress 3.0, both now credible in the contested middle: the worksheet from our July coverage governs. Content-model complexity. Workflow needs. Multilingual needs, where Drupal’s title is undisputed. And the editorial-experience weighting that 7 just made a much fairer fight. We are having that conversation weekly now. For the first time, it is a real conversation in both directions.

Field notes from launch-week builds

Three practical observations from our first 7 builds, for the implementers.

The entity groundwork under the fields system pays immediately in consistency. Content, users, and terms now share one architectural idea. One field type, attached anywhere, behaving the same. It also hints at where the platform is heading. Developers who absorb the entity concept now will read 8’s roadmap fluently later.

The image-styles system deserves early configuration attention. Define the site’s derivative sizes deliberately at build time. Thumbnail, card, hero. Named for use, not dimensions. The templates stay clean while editors stop hand-resizing photography. That is a workflow win training sessions surface within the hour.

And the database layer’s rewrite means custom query code from 6-era modules wants a real review rather than a mechanical port. Budget the review. The new layer rewards it with portability the old hand-rolled queries never had.

None of this changes the buyer-level verdict above. All of it belongs in the implementer’s first-month notebook, and ours is already annotated.

The wait, in retrospect

Three years is a long development cycle, and the community’s own reviews are already asking the process questions. From the outside, though, note what the wait purchased. A release that moved the platform’s two weakest fronts, editor experience and architectural coherence, rather than piling features onto old foundations. That is the harder release to ship and the better one to receive.

Whether Drupal can shorten its cycles without losing that discipline is the question version 8 will answer. The early signals, initiative-based development and much public discussion, suggest the community knows it.

For now: the three-year wait produced the best Drupal yet by a comfortable margin, and our picks have been updated to match. The platform wars of 2011 promise to keep us busy, with this release, WordPress’s momentum, and the enterprise tier reshuffling under Adobe. Worth the wait. Now let’s see the follow-through.