In drupal you can assign users specific privileges so they can only access certain parts of the backend.

I don't really know enough to pick sides in this discussion but it seems you don't either.

I know enough to know that I haven't ran into a single case where a truly generic solution supplied by a CMS fit the needs of a paying client. The only exception would be a 100% standard blog, but when people ask for that I just send them to tumblr and if needed help them setup a domain/etc..

I also know that I want to spend my time learning something that has value to me.

prepare yourselves for the sickest most massive brag of all time:

sounds like you're too lazy to learn a cms api

the two toughest achievements on SO are

1) answering a .NET question correctly before jon skeet gets to it
2) winning a bounty offered by jon skeet

i think #2 is wholly impossible

Not really. Until I see a real world case where I would potentially need to use one, why would I spend time learning something I don't have a need for?

Fast application development is about being used to what you're doing and also automating as much as possible within reason. I'm very happy with my work flow for creating new sites/apps.

Like I said previously, I like options. I also like being able to pick whatever programming language(s) and database types I want to get the job done in an optimal manner.

Just to clarify (and I'm not trolling, I just can't tell); is your position that there is never a real world use case for a CMS?

how are you arguing that rolling your own is better for "fast application development" than using a CMS

What about rolling your own CMS

CMS vs Bespoke, like most things it just depends on what you're trying to achieve and what your goals are.

I'm saying this as someone who really doesn't like CMS's either for various reasons, but I can definitely see their benefit.

The discussion will never end if you debate a false dichotomy like this!

I actually think there's an argument to be had there, unless you are already an expert in the CMS. I have found (Magento comes to mind in particular, not sure if that qualifies as a CMS but same idea imo) that learning how to configure a ready made solution can be more work than learning a framework from scratch.

I mean, why would you ever hand code html when you can use dreamweaver?

Often the shortcut "easy" methods of prebuilt solutions make edge cases so difficult to solve that all their benefit is lost.

No. You need to read a few posts up to get the context. My position is I have never found a reason to use one.

It's dead simple. You're allowed to re-use the code you write.

If I wrote an image gallery once, why wouldn't I be able to use it again in a different project? I would just copy it over and modify the bits and pieces that need to be modified.

If I were using a CMS the same thing would need to be done except the functionality of the image gallery is limited to whatever image gallery module I decided to pick. To change it would involve digging pretty deep in code written by 1 or more people.

As soon as it really needs custom functionality (something a bit more basic than let's say adding a new field to be output), then I have to deal with learning the entire image gallery code base, and I have to be very comfortable hacking away at the CMS itself in case I need something it doesn't offer.

What happens when the site you've been asked to make is using 8 different modules and 6 of them need to be heavily modified? That seems like an awful position to be in and I would hate life if I had to deal with that on a regular basis.

What I need most is stuff like bootstrap. ****ing layout and design, I suck at that (which is why I'd never be a solo web dev in the first place)

Been there done that. The same problems arise in the end. You end up trying to generalize stuff to the point of silliness and the code+features suffer. It also becomes harder to maintain because it's much more complicated and performs worse because it's doing a ton of extra stuff.

There has to be a balance between performance, code complexity, flexibility and dev speed. Using any type of CMS will not give you all 4 things at an optimal level.

Actually writing code and learning how to glue together the pieces with minor code duplication gives you 4 all things at an optimal level. Once you know how to write decoupled code and how to connect it all together development is really fun and you become really productive.

i suck at layout and design too. another benefit of the using a CMS is that i can just download random themes and try out different styles. you think i'm gonna waste my time writing CSS divs ?

with a CMS, an average computer user can build a website without having to hire a developer. concrete5 is another CMS thats gaining a lot of traction

Fair enough.


I think you're just overestimating the probability of this happening. This just isn't really an issue in my experience as it's usually pretty easy to tell ahead of time whether you're going to paint yourself into a corner.

I use C5 for a couple of clients and they love it fwiw.

edit: Re: design...I think it is a very rare developer who should attempt any sort of design work. We pretty much all suck at it.

I know I am. I use Bootstrap for all of my internal tools now.

yeah bootstrap is awesome for that kind of thing. I'm using it on an intranet app at work and have got a number of comments from the users who are impressed by how "professional" it looks.

Yeah likewise. The only people worse at design than me are my customers, thankfully.

Clearly the best reason for a developer not to use a standard CMS: lock-in.

If you give your client Wordpress then they're not wedded to you for all time.

not being locked in to a developer/client relationship is a bonus imo.

Reoccurring payments are very profitable. People really don't mind paying for updates.

A typical mom and pop shop might have very little employees. They also don't care too much about trying to run the web site. They just want a web site and want to concentrate doing what they do best (running some business).

In situations like that they will be more than happy to throw you $ to occasionally update the site. Maybe every month or every 2-3 months they give you changes to implement and then you get a check in the mail.

Being locked-in is an ideal situation to be in, as long as it's not more often than once or twice a month for updates.

This arrangement doesn't require lock-in though. Most clients are going to continue in this type of arrangement as long as they are happy with your service no matter what backend/framework you are using.

Until you want to change jobs, reduce clients or whatever. It's easier to recommend a replacement for yourself when you are using a widely used solution. You can re-assure your clients that if you are hit by a bus another developer will easily be able to pick up where you left off etc. I really don't want my clients feeling like they are handcuffed to me.

It doesn't, but if you were the original developer and they are happy with your updates they have no reason to just ditch you for someone else.

As for future maintenance. That is reasonable but you need to make sure everyone is clear on the plan beforehand and then it's not close to a problem.

You need to make sure your clients know that the site you're selling them is as-is at the time of payment and any future updates are considered separate transactions and you're not obligated to maintain the site for any amount of time (unless you specifically say you will).

Also having someone dump a CMS site onto someone else who had no relation to the project isn't exactly that great either.

Someone asked me if I could help them update their word press site a while ago. The guy who set it up had like 50 plugins and a lot of them were so garbage, he was also unreachable and I know nothing about word press. Eventually the site owner wanted to move a few things around and it was a massive headache.

I got things changed but it was beyond dirty. I'd rather look at code written by $3/hour coders than deal with that type of mess again. I also think there's more general rails or PHP devs than specific CMS devs.