Gregory Szorc's blog

Basic Expectations of Enterprise Software

In a few of my previous posts, I vented some steam about my disappointment with so-called "enterprise" software. To prove that I am not just a pathological complainer, I am going to back up those posts with a list of what I believe are some basic expectations of enterprise software.

Basic Expectations

• There is an open forum where fellow clients can discuss the product, post issues, code, etc. There must be at least a message board and/or wiki.
• The software can be seemlessly inserted into your existing enterprise solution stack without having to install additional products from the vendor that offer the same functionality as an existing solution
• User authentication should support everything you can throw at it
• A well-documented XML interface is provided. If you support SOAP, you supply the WSDL. The XML interface guarantees you can use nearly any programming language you want to interface with the product.
• The SDK (if provided) is thoroughly documented and is available in more than one programming language (easily done if SDK is just a XML interface)
• None of the claims of the software are half-baked. I often find products saying they do something just so they can get a check box when in fact they only do 5% of something.
• Built on standards. Your product means nothing to me if it breaks a specification. Standards exist for a reason. Improve the standards through the appropriate channels if you want to change something. Until then, keep it.
• Free from a clause that prevents you from benchmarking it. If your product is so good, let us prove it.

Unfortunately, most enterprise applications that I've seen don't meet these expectations. At Case, we have offenders for all these expectations. Blackboard doesn't have sufficient documentation available to freelance developers. Oracle is notorious for requiring their own software be installed so you can install product A. For example, they have a customized LDAP server that is required for the portal and thus people have to maintain two central directories instead of one. Yuck!

As the university moves ahead with deploying its next large-scale enterprise applications, I am concerned what those products will bring. Will the Peoplesoft student information system have an XML interface, or will it lock data in so I have to go through the vendor supplied interface to obtain information? Will I finally be able to get RSS feeds out of Blackboard when it is upgraded? Will people realize Oracle's entire application stack can be replaced with open source software? I don't have those answers. I can tell you that if the enterprise applications of the future meet the basic expectations I have outlined above, then I should be a happy man.

You can ping this entry by using http://blog.case.edu/gps10/mt-tb.cgi/6097 .