Of course sometimes it's a mistake to use a CMS if you need a lot of custom functionality.

The whole point is that 100-X% (where x is small for some definition of small) of websites on the internet don't need anything more than a plug and play CMS that can run on a $5 a month shared hosting platform and therefore rolling your own takes longer, costs more money and likely has no measurable upside.

but the same holds true when you build it yourself. except with a CMS you've saved yourself 90% of the development time

Can you give me an example of a type of site you would use a CMS for? Make sure to list out each feature you want, the type of content it will contain and how you plan to edit/manage the site.

This will be a fun experiment.

I have a small martial arts dojo that I want to make a website for. The features it needs are:

- a landing page with some basic information about dojo
- pages outlining the classes available at my dojo
- profile pages of my instructors
- contact page including directions and a map to the studio (using google maps)
- photo gallery that i can add videos and photos to of our tournaments and practices
- calendar of events page that lists upcoming and past events

I need to be able to edit the site myself using an online editor as the class descriptions change fairly often, we have teachers coming and going etc.

Do you happen to be a martial arts expert who owns his own dojo and is also technically capable of buying a domain, figuring out what hosting you need and have a lot of experience working with your CMS of choice because that doesn't really seem like the norm.

Most people who want a web site and aren't developers themselves have no idea how to do any of that. That is why they are hiring you. I also feel like if you just threw a random person at the WP admin panel and told them here's how you update your site they would be lost even with an hour or 2 of training, they will constantly be calling or e-mailing you asking how to do stuff.

I have given back ends for some clients that were dead simple. At least on par with the difficulty of sending a plain text e-mail. They asked for the feature of being able to update their site and even with it being so easy to update they still don't do it. You think those people will want to touch WP's admin panel or another CMS?

How long would it take you to get that site up and running with a CMS with all of the features working and looking exactly how you want them to?

Assume that you already have a host setup and you're working with a blank slate here. You need a design, a way to implement it into a theme or are you just going to use a free theme that 750,000 other people are using and finally you need to get all of the features in.

Are you restricting any feature functionality or how things look based on what the CMS provides? What if you were given just a PSD of the design and were asked to supply all of your features into a working product, how much time does this add?

Will any of your choices allow you to create similar but not identical sites quicker? If so, list them along with the reason(s).

No of course not...that is why I'm hiring you.

My experience is you are wrong. I have put up a number of sites for clients that are nearly identical to the one I described above. I give a bit of initial training and some documentation and then handle the few (a couple a year) questions myself.

If they can't deal with the complexity of sending a plain text email then I don't know what to tell you. Obviously they should be hiring someone to update the site for them. I'm not even sure what point you're making here.

Getting the CMS/features set up and live on a host is anything from an hour to a day of work depending on exact requirements.


For the design if we can find a template they like (free or purchased, purchasing usually works better ime) then that is easy. Otherwise I have them work with a designer to get something they like the look of...that can take an widely variable amount of time and money.

why would I? Don't really understand what you mean.

It depends

Of course? Over time you methods of doing this effectively, snippets of code etc?

I, along with many others, have happy clients that use CMS's for their sites and they work great. What point are you trying to make exactly?

urgh. can't believe I'm actually arguing with you. You seem to want to fit everything into a single box and ignore any direct evidence contrary to your "I can do everything, bigger, faster, stronger, myself" attitude.

yeah, guy has his mind made up already

So, instead of giving answers to your cherry picked best case CMS example you'd rather just drop it and answer them in a way that's ridiculously condescending, hostile and incomplete (you don't even know the answers to questions you've supplied...).

What was the point in even replying then? I'm sure you must be a joy to work with haha. Also who said we're arguing? I'm not arguing. I asked a few innocent questions, that's it.

You try to do the same thing to me in every thread. I'll make a comment and then you go into attack mode and try to prove me wrong but fail (example: the guy asking if he should become a web dev thread). I mean, rather than read the OP's post and then my reply, it was almost like you saw that I posted in it and instantly read my reply instead of the OP's question and weren't even aware of the question I tried to answer.

How about we make this really simple. If you have some weird unprovoked problem with me, just stop replying to my posts. It'll save us both a lot of typing.

You asked for an example of what a good use case for a CMS is and I gave you one.

not sure what you're talking about here.

RE the other paragraphs, we just have different attitudes towards development and I often feel compelled to respond.

Trying to convince someone that thinks jquery is too bloated and it's better to write his own JS library to use a CMS seems like a futile battle.

haha i forgot about that

Anyone else seeing a couple of downsides to Stripe recently?

I've seen people implementing it badly which is dangerous. For example, I saw a site that used Stripe but when you disable JS the credit card numbers are posted to their own server which afaik is illegal. Website owners might also be inadventently storing the CC numbers submitted to their own servers in places such as extended web logs. Forms set as GET would also pass the CC number in the URL which is horrible.

This is made even worse when the page isn't on HTTPs, which I've also seen real world examples of.

I also don't like the way it's being adopted by lots of new websites, ones where I would highly advise people to not enter their CC number on as for all intents and purposes they are unknown. Stripe is amazing, but it's not solving the trust problem that buyers have - Paypal solves this. Stripe is not suitable for every new website yet it seems to be adopted as though it is a suitable solution for every website. A website with a payment form, it would be absolutely trivial for unsavoury people to capture and store the entered CC numbers.

I'm starting to see third party solutions as well that suggest you use their service to record the CC number, another bad idea.

People implementing Stripe badly/wrongly is very dangerous and it seems to be an emerging trend to me. I love the new technology and the hackers that experiment with it, but people are starting to get very careless with it. I for one will never enter my CC number on any website where I don't know or trust them to a certain level. A lot of other customers will not be as savvy as me in this regard and put a lot at risk when they enter their CC on an unknown site.

You cant really blame Stripe for bad implementation. Its sad to see that by creating such a good service, they have inadvertently shot themselves in the foot by allowing any 2 pence site to implement a payment service, but I really think that its up to the end user to decide whether they feel safe entering their CC details on the site or not.

If sites are storing CC info, they are breaking the law and should be reprimanded. Stripe obviously should not condone this practice, and should shut down any sites who are caught, but its not their job to actively police every site who uses their service.

Bad payment gateway implementation is super dangerous, but its been going on long before Stripe came along.

Do you not think that Stripe has some responsibility in ensuring that people use their services safely? I'm not talking about 100% responsibility but I'd be in favour of them reviewing implementations and suspending service if it's not safely implemented.

I would think the CC companies would think Stripe holds some responsibilities, if they see a ton of fraud and security problems coming from Stripe implementations they could easily justify a suspension of Stripes services.

In regards to CMS:

There is one company out here in Los Angeles that has a highly convincing client list that uses only WordPress, Magento (open-source and enterprise), and Drupal. Yes, they have companies every person in the first world has heard of.

One of their clients is a very popular movie theater. The interface shows what seats are available. Now consider a big opening: 1000 people log on at the same time, plus there's people lined up at the B&M cashier, and all of these people have to be able to point their mouse at the available seats and order tickets to whatever movie. This is most certainly a concurrency issue, where it is highly likely that 20 people are going to mouse over the middle seats at the same time. Who wins? The site is written in Magento, tied to the B&M, and of course there has to be a way to add/subtract movie times, upload trailers, etc etc etc.

No, I am not writing a marketing website, and while I agree that it could be done on a CMS, I would like to assert that a CMS is most likely the better solution if you are marketing LOL-SEO or other "Solutions," but if you are, for example, running a data-intensive site that implements multi-variable analysis with linear regression and other math-intensive applications, I would question the sanity of anyone who thinks a CMS is better than Django + Numpy or something similar.

To be fair, I am NOT asserting that my site is better done without a CMS. I am asserting that I am not knowledgeable enough with a CMS to effectively create anything, and I feel that my particular circumstances is too involved with future unknowns, is highly database-driven, and that the unknowns are tied too tightly to the database, and I'd rather be able to implement changes very fast. Even despite not knowing anything about the language I am using before I started, and changing my mind on the UI, thus doing 70% of the site two times, I only spend about 20 to 30 hours total on what I have right now. And that 70% change only took about three hours, and most of my time has really been focused on stuff that, if I was better at this, would have only taken about 10 hours to create what I have now.

For me, the learning curve is worth the effort, as I get to think about the design of the system from the bottom up and top down, and that if I am looking to implement something more complex in the future, I at least have a foundation to start on. I really try to be as non-religious as I can about any of this stuff, but at the end of the day, it really breaks down to what makes me feel good and what I think has a better chance of being more beneficial to me in my growth. If I have to choose between a CMS or Clojure, I think the answer to this is obvious.

I don't see why you shouldn't use CMS for many website projects. Talking to the customers and getting it right is the main thing not some technical mumbo-jumbo.
At the end of the day what people want in most website building cases is a simple to use and pretty site. Usually no bells and whistles required.

The more the site is tied into their business processes instead of being mostly representative the more I#d lean towards something custom build with Django and the likes.

Ideally, yes, it would be great if Stripe could ensure that everyone was implementing their service correctly. But in practice this is almost impossible. Stripe's selling point is that its super easy to implement and cheap for small businesses. I would guess that a v high % of Stripe's customers are small businesses that generate less than $100/month in revenue for Stripe (around 100 transactions at 2.39% + 30c).

To ensure their system was correctly implemented, they would need to do an audit of every site that wished to use their service. I just cant see how this would be possible from a financial point of view. Also, a big factor in Stripe's success is that its very quick to setup. If you had to get an audit of your system before Stripe would let you online, it would kill most of their business.

I think that Stripe should closely monitor activity though their system, and be quick to suspend accounts which break the rules, but I just dont see a way they can reasonably ensure every site using their service is legit.

(This might actually be a flaw in their business model. By opening up a cheap and easy way to take payments on any website, their system is very open to abuse).

Maybe we just have different types of clients. I don't usually get site jobs where a CMS ever really fits.

Most sites I get are:

1. Very very static sites where at most we're talking about a contact form or paypal buy now buttons. The content itself is put up, updates are very infrequent and when they are done it is done by me for a fee.

2. Mash up sites which take data from x y z places and then does stuff with the data. Usually has admin dashboards for only what they need access to or want access to, very specific to the site.

3. Content-ish sites with calendars, image galleries, etc. but more often than not there is also functionality tied in directly to their business logic. Custom forms, custom db setups, custom validation, etc.. Maybe they also want to add in very specific pages that require a lot of custom programming to get what they want. Admin dashboards for these sites too, but it's trivial to implement.

It seems dumb to try and shoe horn #3 into a CMS when the code base to get the semi-generic CRUD-y stuff can be written once and used anywhere with minor modifications.

I'd much rather further my knowledge of javascript, python, ruby, php or whatever other language I'm using rather than learn to use, modify and theme a CMS. I don't ever plan to work at a place in my entire life that depends on a current open source CMS so the knowledge gained to me is 100% useless.

I don't really do design work too unless it's a minimal-type site, and finding a designer who knows how to properly theme x CMS is a lot more rare than just asking a designer for a PSD. I like having as many options as possible. I don't want to have the burden of having to theme out a CMS.

Wait a minute: You want to use an open-source programming language that was created to solve enterprise issues but you don't want to use a stack that involves an open-source enterprise program? Admittedly, Java is huge, but I beseech you to make a phone call via some telecom giant without it involving Erlang.

I'm not entirely sure if I should continue this thread of thought, but did your research say that Erlang Web compiles to JVM or is there some other thing involving JAVA? I would assert that if it compiles to JVM, there are at least 4 good benefits that may make you consider using JVM.


Yeah, but the "customer" in this situation cares about technical mumbo-jumbo, and of course "easy to use" is relative. I struggle like hell to use 90% of the CMS back-ends I have been exposed to, and I've seen far too many to have a positive attitude on them. Yeah, I guess I'm a weirdo since I find using raw HTML + CSS is easier than dealing with the sometimes-it-works-but-usually-fails editing features I've had to work with.

There have been a few members posting open source software and tools they have created. Off the top of my head:

- Free Poker DB (the one and only, latest and greatest) - http://****************/apps/mediawik...itle=Main_Page (btw, you guys have any plans to move to GitHub? SourceForge really sucks)

- C based poker odds calculator - http://forumserver.twoplustwo.com/19...lator-1233513/

- .NET hand parser - http://forumserver.twoplustwo.com/19...arser-1232028/

I am sure there are others that have been missed, and maybe some which members have not posted here yet, but I think it would be really great to see these in the Sticky Thread so members looking to build poker related tools have some resources to get them started.

clowntable,

If you research Erlang more let me know.

You might be interested in this article too:
http://jlouisramblings.blogspot.com/...to-behave.html

Follow up comments on HN:
http://news.ycombinator.com/item?id=4372378

Ah maybe I wasn't clear. What I meant was that Erlang Web sounds a lot more like Enterprise JavaBeans than Rails. Enterprisy vs Agile. More of a behemoth. Chicago Boss otoh is more Rails like...hackish (in the good sense) I guess.

I have no idea if that is actually true or not but it was mentioned by a couple of people in different articles/posts I read which is why my pre-choice would be pciking Chicago Boss over Erlang Web.

what is the difference between a framework with an admin dashboard and a cms?

The biggest difference to me is you only release what you need. You have a lot more control what the client ends up seeing. You also don't have to worry about them breaking the site or messing with things they shouldn't be touching.

In a lot of cases (maybe not all) I personally think a CMS is very much overkill. If some dude only needs a media gallery and a user system I'd definitely use existing code in rails or whatever language/framework you use vs a CMS.