You are here

Drupal 8

Drupal 8 is shaping up to be another Drupal 7 waterfall without the benefits.

Waterfall or rapid development?

Back in the Gordon Gecko days of the 1980s the big rich companies used the waterfall approach to application development. Everything was big and one way, like the Victoria Falls in Africa. The waterfall approach is widely discredited and almost every other approach is better.

Rapid application development is better. There are many approaches to rapid application development, including some not so rapid approaches that are called RAD. The rapid approaches are of varying use and each approach has different benefits, making them suitable for different types of projects.

Drupal 7 was a waterfall project where everything went through massive changes you cannot reverse out, the total gain was small compared to the effort, and the time between releases was massive, way too long to be useful.

Some consider the time between the release of Drupal 6 and Drupal 7 to be acceptable but Drupal 7 was released at least a year after the point where users where ready for a change, plus Drupal 7 did not become usable for almost a year after release, making Drupal 7 two years too late.

Drupal 8 to be worse

The changes lined up for Drupal 8 are bigger than the changes for Drupal 7. Some minority interests have won their political battles to make changes that will slow down every aspect of Drupal 8 without really improving anything. The winners will then push to change everything to their way.

No upgrade

Based on the change to Drupal 7, there will be no clean upgrade path for most sites.

I upgraded over a hundred Web sites from Drupal 6 to Drupal 7 and have over a hundred to go. The big roadblock is the module upgrade path. Some modules have no upgrade path. Some modules make you throw everything away and start again.

The module upgrade roadblock is caused, in part, by Drupal 7 incorporating changes that are not document or do not have useful documentation. There are a whole lot of changes in Drupal 7 that were never tested, where never tested against documentation, and were never tested for an upgrade path.

The changes are political

Many of the changes are political. Looking back at the change from D6 to D7, many changes were added that had no advantages, created performance problems, and were not tested in the form in which they were added. If the D7 project was professionally managed, some of the new code would be split into smaller chunks and tested one addition at a time. D7 would have arrived a year earlier with fewer changes. The other bits would be back at the drawing board or in a D8 available on about the same date as D7 actually arrived.

If Drupal switched from the watrfall approach to something more modern, the Field system would not be missing important field types, would have a working interface, and you would be able to tell the difference between the several parts of a field. Fields and internationalisation would work together without all the pain.

D8 could arrive with a real miracle, Views working either consistently or reliably with multiple tables.

The performance problems would be either predictable or curable. The band-aids slapped on Drupal sites to cure performance problems only work in some circumstances and are often a distraction from the real problems.

Imagine yourself in a remote area stuck on the side of the road with a flat tyre. Along comes a Drupal consultant. You ask for help. The Drupal consultant tells you to empty the gas out of the tank because that will make the car lighter and faster.

You point out the limited distance you can travel with no gas and remind the consultant of the flat tire. The consultant tells you you should have used a boat, you should have loaded the car on a boat because boats do not have tires and cannot have flat tyres. You look around at the rocky desert and think about the chances of someone finding a consultant's body out here.

Drupal 8.x-dev

April 2013 The Drupal 8 development version packaged on April 23, 2013, contains 9167 files totalling 39.8 MB. most of the changes are minor compared to the previous week. Some files, such as token.inc, disappear and are replaced by classes or Symfony code, with many cases resulting in one Drupal code file replaced by many Drupal/Symfony files.

Conclusion

Drupal 8 is scheduled to be a bigger change with far less benefit, little architectural design, and few changes aimed at the areas creating the biggest productivity losses.