[thelist] CMS: opensource or hand-roll?

Zachary Kent zachary.kent at gmail.com
Wed Sep 17 22:50:26 CDT 2008

> Off the Shelf (OTS), open source or otherwise,  can be fine for some very
> defined situations and a horror story in other situations.  Generally
> speaking I've found that starting off with an OTS version can be OK for
> prototyping but then you run into a decision point where their processes
> can't be supported by the OTS software.

I think you are exactly right on this issue.  Every time I start to
think about using OTS I start to imagine the horror story and it is
basically what I am hoping to avoid.

> Then there's what I think of as the "potato chip" issue.  Clients think they
> only need a system to do X, and the next thing you know they all want it to
> do XYZ, then S, then AB.  You have to be able to know your client better
> than they know themselves to be able to predict what they really need now,
> and within the foreseeable future.

almost every time.

> There's also the issues of security and support. A couple of years ago I
> deeply and widely explored OTS systems, both fee/free, both PHP and .Net,
> for a couple of months and I found an appalling lack of attention to
> security, spaghetti code and sites that were more focused on marketing than
> support.

That is good to hear.  I am glad I can benefit from that experience.
I have oftentimes found a great piece of code only to find out later
that something about the core design is bad or makes the app hard to
service or just plain unsecure.

> During the past few years I've either had to custom build or figure out how
> to add custom modules to OTS systems.  Client TCO, so far,  has been less
> for custom build because of the extra charges to recustomize for version
> changes and open source software tends to have a lot of version changes.

Very interesting.  I was wondering about this and figured that I would
only figure this out by trying out several systems.  I hated to go
that route b/c it can be so difficult to move to a new system if it
doesn't work out for the client.

> If it's a small job, with limited editors and permissions, I still consider
> going OTS first and then make the final decision after deeply understanding
> the client's current processes and improved processes.

Thanks for the advice.  I agree with you.  You have helped me decide
that the world needs yet another CMS.   I have been leaning toward a
template based CMS that generates standard XHTML so that the final
site is not locked into the CMS' database.  Of course it will be
modular so I can sell the modules separately and maintain them

I once created a massive online registration system for satellite
teleconferences and I learned a lot.  As I try new products, I am
shocked at times at the problems at the core of most apps.  At least
this way the core mistakes will be mine and I wont be fussing about
someone else's code.


More information about the thelist mailing list