Sent by Rickard_Öberg on 26 November 2003 12:12
Serge Huber2 wrote:
> Hey Rickard, I thought you'd resist :)
'twas too difficult. :-)
> However, they are also a LOT of OSS CMS' that can fill that need. And so
> instead of designing the one millionth small CMS, I would recommend to
> spend time learning what's out there, and maybe adapting a project that
> fits your needs reasonably well.
From the closed-source perspective (which is my own), there's now also
the option of buying a proprietary solution (such as ours) and do
customizations using JSR168 portlets.
In the (near) future, it seems to me that writing JSR168 portlets which
use JSR170 to access the content should be a good way for customers to
add new behaviour to a CMS/portal.
> But the main question is : why ?
> Well simply because unless you are part of a major corporation that has
> a huge IT department and can devote lots of resources to projects such
> as CMS' you will inevitably run into a wall at some point. CMS' systems
> grow over time quite substiantially, and it's always amazing to see site
> grow from 5-10 pages to over 10'000 pages in a much shorter time than
> anticipated. CMS' allow people to manage their data in a better way, so
> if the system fits the bill, it will grow.
> So back to the OSS systems : instead of doing it on your own, you can
> benefit from others working on the same code, sharing problems,
> solutions, etc.. Also, one thing I usually look at is what happens after
> developers have left a project. Let's say you've developed your own CMS
> using a team of 2-3 people. Later on the company decides they can no
> longer afford developers on this project and reassigns them. The
> developers leave the company. Later on (this *always* happens in
> software !), the company wants to make changes to the still running
> system. In closed source systems this is very difficult and
> time-consuming because the code (if still available) will be old and you
> will have to work wonders to make the slightest change, where as in an
> open source project you will have more chances to find somebody that can
> help you out with the system.
That is the theory anyway ;-) As an OpenSource developer (I am cofounder
of JBoss,WebWork,XDoclet projects in the Java-sphere, for those who
don't know me) I know that getting people to work on your particular
project can be VERY difficult. There needs to be incentive, and just
"being OpenSource" isn't enough these days. It's a very competitive
market out there, but instead of competing for bucks projects are
competing for mindshare and brain cycles.
But if it works, it can work out the way you say, absolutely.
> I'm not saying OSS is the only solution, far from there. OSS has it's
> problems too : no guaranteed support, the developers of the project
> might also all leave (but leaving the source behind them), the project
> might go into another direction that what you anticipated. A lot of OSS
> projects loose steam after a while, because the original contributors
> have left the project, and the project might be "dead" for a while. Also
> another downside of OSS systems is that they often (but not always) lack
> the finish of commercial solutions.
The last one is more than often a big problem for OpenSource systems.
They are typically made by programmers for programmers, which presents a
problem when your average secretary (or any other non-techie people)
should do content updates. Getting the UI right is very difficult.
> So instead of develop or buy, I would suggest : assume all costs or
> share costs.
My alternative approach to what you describe is: you might as well get a
proprietary solution, and do the customizations using JSR168/JSR170.
That way you get the best of both worlds: a maintained infrastructure,
and freedom to implement the exact features you want dealing with the
problems you are interested in.
Rickard, Java CMS developer
please trim your posts.