Discussion on:

Architecting a CMS in ASP.NET: One server or a bunch of servers?

As a professional consultant on content management systems, i was immediately interested in the article title only to find a very generic featurette on standard 3 tier architectures used for nearly every web-enabled system around that had nothing atall to do with CMS architectures. It doesnt answer its original question and doesnt supply any help or discussion points at all for anyone interested in building a CMS.

It needed a good discussion on the difference between push and pull CMS architecures which will heavily affect your hardware physical architecture choices. A push system only needs to scale for CMS users while a pull system will need to scale for users of the content delivery system (your website)

I agree with User Deleted. Not only is this "n-tier 101," it doesn't even ask the right questions about what you're out there doing in the first place.

The most important question to ask is, "How do I build something once and adapt it for every need my never-good-enough customers/co-workers have?" The answer is an extensible enough system that any widget you create is easily added by the customers (by dynamically loaded user controls, set via simple UI) and a base DHTML control that does allof your editing for you.

There are so many CMS products in the market there is little need to build your own. Granted that some of these packages are out of any sane person's price range, there are some cutting edge products which aren't.

For example, Metaverse Content Server is built completely off the Microsoft.NET platform, and provides an API, a set of developers tools, and a developers network (similiar to MSDN) for building dynamic content sites with Metaverse. It allows business users to author documents inMicrosoft Word, which are then converted to XML and then to HTML via XSLT.

Metaverse is offered as a hosted subscription service, Small Business Edition starting at $65/mo.

It offers a way to get up to speed with Web Services and .NET without having to build your own system from scratch, and providing a whole lot more functionality than if you had.

All that talk and I forgot to give the URL

www.metaverse.cc

It doesn't seem like an article series on ASP.NET enabled CMS's. I'd at least like to see some mention of ASP.NET. Just a stupid 3-tier explanation I've seen hundreds of times.

Metaverse (and all other products) look great. Challenge is when you hit point you need it to do X and X is fundamentally impractical given the information architecture.

No matter how well designed the CMS, biggest issue is ability to customize to specific needs. (I know as I've designed a few - both custom ones and a commercial product back in '98).

There are many ways to offer this (built in features, API, links to separate non-CMS code, etc.) but often non-optimal (I want to make query that dynamically pulls content sorted against a full text search of documents relating to a subset of products where the product data is in a legacy system, there are usually bumps when trying to do this).

I was interested in the separate topic of open source CMS's and believe there's definitely a place for full CMS systems and simpler extensions to programming languages that allow easier implementation of core cms capabilities as well as well documented patterns to shorten learning curve on best way to spec a security scheme or workflow status stack, but to just offer an one of the box product is not always enough. Sometimes learning curve is greater than time to replicate the key functionality required in the app!

As i pointed out in the first reply, it would have been nice to have an article on some of the base points of CMS architectures.

Ive designed and built CMSs from scratch and done extensive implementations of commercial market-leader CMS and one thing is the same whether it is a "roll-yer-own" or ?250,000 commercial product. Every business needs are different all the time and constantly changing. Nearly any CMS needs to have a set level of base functionality (ie, versioning, templating, delivery, workflow) and then this needs to be customised to fufill the business requirements.

Ive seen it too many times when organisations will buy an off-the-self product and expect to plug it in and hey-we-got-content-management arrives. Then the seem amazed when they have to bring in consultants to make it work.

From my experience a content management product is only as good as the things surrounding it.

Content Management
Content Delivery
Editorial Process

There is no point in having a kick-ass content management system (with massive workflows, custom UIs, integrated into all legacy back-end systems, SAP and CRM) without having a decent performant, personalised content delivery engine. And neither of these work withoutthe business adapting to a editorial process behind it that works for the business and system.

All of these things make a decent Content Managed Solution.

List of CMS products with URLs can be found here:

http://www.cmswatch.com/ContentManagement/Products/