Accueil / Blog / Métier / 2014 / Why CMS will not die

Why CMS will not die

Par Eric Brehault publié 09/10/2014
A little insight into the future of CMS
Why CMS will not die

It is always difficult to step into the future in a changing area such as the web. However it is an interesting exercise, and while I am at it I'll get rid of the conditional tense, which is a style I do not particularly like. Here is a series of statements on which I cannot offer any guarantees of course.

All CMS are old

Typo3: 1997

Drupal: 2001

Plone: 2002

Wordpress: 2003

Joomla!: 2005 (forked from Mambo: 2000)

We work in a field where everything tends to age badly, yet the most used CMS appear to be ten-year-old veterans.


Because CMS might be old things, these are old things that work well. Of course, you must stay up-to-date, the web world bringing every day innovations that must be added to the CMS to avoid (1) disappointing users and (2) frustrating developers.

But the main use cases keep quite stable: the user expectations about their CMS do not change much. Even if today we are able to manage a CMS-less site by writing Markdown contents and producing a static web site on GitHub pages (and this approach can be applied to build quite complex solutions), still, in lot of cases, we really need a CMS.

And in those cases, you just cannot stay half-way: at the beginning, you probably just want to edit few pages, but then you realize you will also need to manage medias, a validation workflow would be nice (hence we have to manage access rights), newsletter is a must-have, plus a document sharing restricted area, multilingual, subsites which are all identical but 2 or 3 things, etc.

You will quickly go through this well-known features list, and the conclusion is obvious: you need a CMS. And guess what, lucky you, there are excellent CMS (see the list above), and they are just doing all of that for a long time.

So we just stick to them.

... and it will not change

It will not change, because there will be no shiny newcomer conquiring the entire market in a year.

Because building a CMS is an enormous task.

First of all, it requires to know the standard use cases (see above), there are many of them, and they are quite rigid (you come with a different scenario, and the users will not follow you). And the best (the only?) way to build and maintain such a strong CMS culture is by having an experienced community.

Secondly, the technical side is huge. The basic infrastructure able to provide the standard CMS behaviors is very large, but in addition, you should be able to integrate rapidly and safely the permanent flow of new features / new technologies appearing on the web.

Moreover, a CMS is a laaaaaarge project, larger than regular ones, so you need specific practices to manage it. For instance, a pretty standard Plone site will easily involve about 350 packages (Python eggs), which have dependencies between them. That is why Plone uses Buildout intensively (Drush does pretty much the same kind of work for Drupal), and Django people often make fun about Plone because of Buildout (they prefer pip which is much simpler of course), but clearly, we do not have the same needs here.

Dependencies management and deployment is just an example, but there are many other aspects that make a CMS a very unusual project, and those who would try to build one without the necessary tools and practices might have problem.

So, to make a CMS viable, you need a solid existing base plus a (highly) productive community.

In short, it will be hard for newcomers.

WordPress will not conquire the entire world

In the CMS world, WordPress is special: more than 60% of the websites using a CMS use WordPress. The other ones are all under 3%.

So we could imagine WordPress will progressively take the entire market.

I doubt that, because even if (technically) any CMS can build any website, each CMS has its own identity which makes it more efficient and more attractive in a given context.

And I do not think the natural initial WordPress positionning (=an excellent blog CMS) can be blurred enough to be able to attract the entire market to itself.

So nothing will ever change?

Of course, things will change. Things always change.

I think the major CMS solutions will maintain their position, as argued above. But today, things are changing tremendously fast on the front-end side, and it will necessarily impact the CMS world.

My opinion is CMS will split even more radically than today their back-end and their front-end. The front-end will take over many more tasks than it used to: content edition UI (Create.js is an interesting example), form rendering, user input validation, etc.; and the back-end will become something like a big REST API engine, but will not be just a remote persistence service, because the use cases complexity will not decrease, and it necessarily implies the back-end is able to address it.

That is the approach proposed by Gizra on Drupal and I think it will spread everywhere soon.

Voir aussi
Résolution de problèmes Drupal : construction de site (2/4) Résolution de problèmes Drupal : construction de site (2/4) 09/08/2013

Dans cette série d'articles, nous tentons de vous aider à vous sortir seuls de situations ...

Gérer sa newsletter avec Drupal Gérer sa newsletter avec Drupal 10/02/2014

Drupal offre plusieurs plusieurs possibilités pour mettre en place une newsletter.

Résolution de problèmes Drupal : Installation (1/4) Résolution de problèmes Drupal : Installation (1/4) 02/08/2013

Dans cette série d'articles, nous tentons de vous aider à vous sortir seuls de situations ...

Drupal & SEO : améliorer le référencement naturel de votre site Drupal & SEO : améliorer le référencement naturel de votre site 20/06/2016

Une vision subjective et argumentée des modules à utiliser pour améliorer le référencement ...

Mon Top 30 des modules Drupal 8 Mon Top 30 des modules Drupal 8 16/02/2019

Transcription d'une conférence donnée au Drupalcamp Paris 2019