Build your own bespoke CMS platform. Dumbing it down - how and why?

[Whitey]

What if you do not have limitless knowledge of the big 3 CMS's [ Drupal , Joomla and Worpress ], or arguably the time to learn; ... and, if you're planning to start simple and scale in an SEO strategy that takes in the likes of multiple sites, excludes "middleware" and "complex backends" or "conversion paths.

Would you consider a bespoke CMS? If so, why and how would you go about it with optimal efficiency?

[ergophobe]

I started out with a bespoke CMS because nothing out there had the features I wanted.

Now there are so many choices for so many use cases, I would only go bespoke in one of three circumstances.

1. I had very special needs that simply can't be met by an off-the-shelf solution.

2. I am or have access to a developer who has turned out dozens of similar apps using Django or Symfony or some similar dev framework that could be leveraged to make something light and fast and simple.

3. I have a ginormous pile of cash that I am just itching to spend reinventing the wheel.

Otherwise, no, I wouldn't.

[Whitey]

Have you seen anything out there in a check list comparison format that details features [ not high level things ]?

What are the best reference points?

[ergophobe]

[cmsmatrix.org...] offers quite a bit of detail.

What do you mean by "reference points"?

In any case, I usually suggest starting with requirements, then looking for a CMS.

[Whitey]

@ergophobe / thanks

CMS Matrix was a really useful reference point - I just hope they can keep it up to date as the last updates are back in early 2012.

So often I've seen requirements drawn up, only to find a few years later a major obstacle blow up, that needs to be catered for. Usually it's something not thought of or anticipated. Is there a planning framework tool out there that can assist in reducing this risk?

[ergophobe]

I don't think so. I'm not sure such a thing would be possible. Consider that the term "responsive design" was first coined by Ethan Marcotte in 2010. By the end of 2011, most people were thinking in terms of responsive design for new sites. By BADCamp in late 2012 everyone was talking about how to make their legacy sites responsive. By BADCamp 2013, nobody was even talking about it at all. It was assumed that you either knew what you needed to know, or you didn't care.

Similarly, if you were to have tried to "future proof" a site and started planning a new site in January 2010, you would have tried to make certain that you could support Flash... but then in April Steve Jobs would have issued this statement: [apple.com...]

So I think you need to project likely future needs and hope you get it reasonably close (will you grow enough to need an editorial workflow, for example?) and look at current offerings and just expect that if you have fundamental changes, you may simply need a new site.

Also, most major open source CMS are highly extensible. So if you do have a new requirement and you've gone the open source route, you'll often be able to throw some money at a developer and have the feature custom built.

[jamie]

hi whitey,

i custom-coded all our sites up until last year. yes it was perfect for what i needed; but... omg talk about spaghetti code (coded over 12 years) and nearly impossible to maintain myself and even more impossible to get someone in who understood it.

switched to drupal last year and whilst it is a heavyweight in terms of learning, we've just launched 3 sites off the same code base quite easily (much more easily than with custom code). furthermore we've had a couple of developers just pitch in and help with some aspects - unthinkable previously - because there is such a large developer base for the big CMS's.

would never custom code a CMS again, unless it was for a mini-site.

i'd like to use symphony or cake because they really appeal to me personally, but drupal has so many features just there. and after a year of painful learning it is finally proving to be a real timesaver ;)

[ergophobe]

jamie, did you know that the next version of Drupal is built using Symfony? So you should be able to leverage both the Drupal ad Symfony dev communities in the future (in theory).

[explorador]

Would you consider a bespoke CMS?

I did 4 times and I'm working on the next one.

I won't write too much to avoid boring you or boring myself, but the first two ones were about: there is no other choice. Learned a lot and boy they were useful. By the 3rd one I built something extremely superior to my previous attempts, fast and adoc to my needs and the company. It was solid, SEO friendly, all that wee need and it worked with high peaks of traffic.

Then the last one... then I noticed my own mistakes and the so many things I could/should have improved. That one still drives my most visited sites.

That one included features I didn't need as a developer but it was more about the clients: multi accounts, multi "databases", etc even a drag and drop module view that was a pain to build.

Over the years I saw others including tech peers battling with WP ha ha falling down, then I converted some of those sites to Drupal because I was not going to be in charge anymore except from my own sites, and I wanted to retain my ownership over the code.


Now that I look at it, my experience is such that I know the code can be better.

If so, why and how would you go about it with optimal efficiency?

reuse code, efficient code, multi purpose code, etc. I keep that in mind and I'm working on my next version but this one will take longer. There are so many things you can do once you have a clear view! and I hate those multi megs solutions when I have fast ones under 3mbs. Not commercial, not for sale etc.

[ergophobe]

Explorador fits into criteria #2 in my first post - as a dev himself, it can make sense to create something that meets your specific needs and that you can reuse for future clients. If you were explorador's client, you would also fit into criteria #2.

A competent developer building a system to meet the specified needs for a given site and not the needs of 1M other users, is going to end up with a lighter, faster site.

But I still think that if you're talking about a DIY site and you are not a developer, some off-the-shelf system is the way to go.

I go back and forth on it though. The off-the-shelf systems get you up and running so fast, but come with a lot of baggage. It can feel like golden handcuffs. Of the "big three" the OP mentions, there are days I hate Drupal and Wordpress. Joomla escapes my ire only because way back when I tested it, I never liked it enough to ever use it on a site, so I've never wrestled with it.

[jamie]

"jamie, did you know that the next version of Drupal is built using Symfony" - now that you say that, i remember. but it's still drupal isn't it ;)

i also have love/hate relationship with drupal.

the one thing i really have learned is that there is not one solution for all problems. you're always going to have to get your hands dirty. at least with drupal, others can help more easily than with custom code (unless you document it well - yeah right! ;)

[engine]

Custom built CMS has worked great for some of the applications I've needed, however, no point in baking your own when there's perfectly good solutions available. I've used a custom built solution for parts of a site that require regular updating. The advantage is that the input forms and formatting are set, and the training for the user is kept to a minimum. Also, they are less likely to make a mess of it. ;)

[Whitey]

Just thought I'd post an update. Our digital development team stepped in and pushed aside the notion that we needed the complexities of a CMS, which they viewed as overkill.

They are ditching the old bespoke .NET CMS - ( actually it's more of a push system, than a CMS ), and have been rebuilding in a different development environment using a wysiwyg html editor in conjunction with a robust database system to publish dynamic and static content.

[ergophobe]

I built a way pared down system like this back around 2000 with just a few variables available to club officers and their emails. It was complex in its way - reservation system for events with email notifications and waiting list, newsletter upload function and a few things like that, but very simple for the main content. To my knowledge it is running unchanged to this day.

Since there are very few SQL queries and very little javascript, it's never really even needed a code update for security (the server platform has been upgraded many times).

It doesn't really meet current needs, but works as designed unchanged for about 12 years. I don't think anyone could say that of site on a Wordpress (1.0 in 2003), Drupal (1.0 in 2001) or Joomla (Mambo 1.0 in 2001).

[timchuma]

I made one by accident when I was fooling around with MS Access and direct file writing. Was not a very good one as it required post-editing. It would store the templates and content separately and generate the files at the click of a button.

I would not recommend going with a custom-built CMS as it locks you into that vendor/programmer who can hold it over your head and you will just require a long and expensive migration to another product later.

There are sites to compare CMS products so you can get the one with features that you actually need.