Drupal 5 Fixes the First Twenty Minutes

Every platform has a reputation, and Drupal earned its own: powerful, flexible, and about as welcoming as a server closet. The capability was never in question. But the first twenty minutes, when a would-be site builder met a wall of configuration, cost the platform adoptions it deserved. Drupal 5.0, released January 15 after nine months of work, aims squarely at that wall.

What’s new

The headline items read as hospitality. A web-based installer replaces the old ritual of hand-editing config files and running database commands. You upload, you click, you have a site.

Garland, the new default theme, replaces the old engineering-department look with something a stakeholder can see without wincing, and you can change its colors from the admin screen. The admin pages have been reorganized by task instead of by whim. And jQuery now ships in core.

That last one deserves a note. We standardized on jQuery last fall, so core adoption by a major CMS is welcome news. In practice it means Drupal’s interface behaviors get consistent, lightweight scripting, and module developers share one foundation instead of each shipping their own library. Expect module interfaces to improve through the year.

The quiet story: structured content grows up

The bigger long-term development is not in core at all. The Content Construction Kit, or CCK, lets you define custom content types with custom fields through the admin interface. It has matured alongside Views, its companion for building filtered, sorted lists of that content without code.

Together they add up to the thing organizations used to buy six-figure licenses for: a content model you design yourself, displayed however your site needs.

Concretely, a university can define a Program type with fields for school, degree level, and deadlines. A Faculty type linked to programs. An Event type with dates and places. Then the program finder, the faculty directory, and the events calendar are all built as Views over that structure. No custom application code. Upgradeable, and maintainable by whoever comes next. Version 5 makes the assembly faster and the result far easier for content staff to run.

What the release doesn’t fix

Candor section. Drupal 5 lowers the entry ramp. It does not flatten the learning curve behind it. The concepts that make the platform powerful still demand real orientation. Nodes, taxonomy, blocks, and the theme layer all take time to learn. The distance between installed and well-built is still measured in expertise.

The module ecosystem remains both superpower and junk drawer. Thousands of contributions of wildly varying quality, and knowing which hundred your site actually depends on is exactly the knowledge you hire when you hire Drupal experience.

Performance keeps its standing footnote too. An unconfigured Drupal site does more database work per page than it should. The difference between sluggish and snappy is caching settings that do not happen by default. None of this is disqualifying. All of it belongs in the estimate.

Upgrading from 4.7: a project, planned like one

For sites already on Drupal, the core upgrade itself is documented and manageable. The schedule risk lives in contributed modules, which each need a 5.x release before you can move, and the long tail of small modules lags core by months.

The sequence that works: inventory every module in use. Check 5.x readiness for each. Then decide per straggler: wait, replace it, or discover that nobody remembers why it is installed and retire it. Every upgrade we have run has shed modules this way. Version boundaries are the natural audit.

Custom code gets the same scrutiny with higher stakes. Themes need updating for 5’s theme-layer changes, and anything hand-built against 4.7’s internals needs developer time in the budget. Then the routine. Full backup. Staging rehearsal. Content freeze. Production move. Verification. Days for a modest site. Weeks for a complex one.

Sites still on 4.6 or earlier should hear a different message. Unsupported versions stop receiving security fixes, which turns ‘we’ll upgrade eventually’ into an exposure with a clock on it.

Two settings that pay immediately

Two switches separate sluggish Drupal 5 sites from snappy ones, and both ship in the box. Enable CSS aggregation, so many stylesheets become few requests. And turn on page caching for anonymous visitors, the single most powerful performance switch in the product. We find it off in a startling share of the sites we audit.

Neither needs a developer. Both belong in the launch checklist and in the handoff notes your implementer leaves behind. Ask for those notes, by the way. Their existence is a vendor-quality signal all by itself.

Fit, put plainly

Drupal merits its complexity when content is truly structured and connected. Multiple content types. Relationships between them. Filtered views, detailed permissions, and community features. That describes more institutional websites than their owners realize.

For straightforward publishing, lighter tools launch faster and leave less to maintain. News, pages, a contact form, one or two editors: that profile does not need Drupal. WordPress keeps getting better at exactly that profile. Pretending one platform serves every need is how agencies turn tools into hammers and clients into nails.

The evaluation question for anyone platform-shopping this spring: describe your content as types and relationships, on one page. If the description strains against ‘posts and pages,’ Drupal 5 belongs on your shortlist. And for the first time, the demo will not need an apology before the login screen.

The trajectory worth noting

Step back and the release says something about open source generally. The knock on community-built software was always polish. Drupal 5 is evidence of a project sanding its own rough edges, prioritizing the experience of people who are not yet community members.

That is product thinking. Its arrival in open-source CMS land should worry the commercial vendors whose remaining pitch was ‘ours looks finished.’ The capability gap closed a while ago. The finish gap is closing now. We will keep reporting as it does.