If you care to check, you will find on most IT vendors websites or marketing documentation the same empty meaningless phrases and unsubstantiated claims, over and over repeated with little to no practical information. That is strange for information technology companies. Most websites are full of intentional misinformation. I think it is actually appalling. Sorry! This article could be developed into a standup comedian act as some people call my articles ‚rants‘. I sometimes worry that I look like a modern Don Quichote and my windmills-becoming-dragons are the ills and spills of information technology. On the other hand I really don‘t care what anyone thinks of me so let me give you a few examples:
Every company is ‘A Leading Company‘. Leading whom in what? Leading a pack of liars in the most broken marketing promises? Hardly ever is specified exactly how a company is leading. Well, some do specify it, but then the outcomes of statistics are what they are chosen to be. And what if a company truly is leading the market with the biggest revenue? Is that really a good sign? In this case Microsoft would be the best IT company there is and it certainly isn‘t. Is stock market share value really an indicator of what this company can do for you? No. It is trader expectations and nothing more. Is a large user base really everything you need? No, because the more users there are, the more requests for changes the vendor has to juggle.
Buying name brand software seems a safe bet for the buyer but is not safe for the business. Name recognition and brands are important but they come at a hefty price tag and have nothing to do with the functionality of the product. With public companies it is easy to see how much money a company is spending for cost of sales. You will find that it is mostly above forty percent of revenue and therefore when you buy a product from this company you obviously pay for its effort to market and sell it. IBM spent billions in 2000 and 2001 to become recognized as a leading e-business solution provider, when they had no product at the time but simply renamed a few old products to the newly created WebSphere brand.
What does actually the term Best-Of-Breed tell you? We used it for a little while too until I felt that it did not describe a functional benefit. It is in fact an admission that all the software of all the vendors of one breed is the same. No wonder that you decide to buy the cheapest.
Open, standard, non-propriatory and plug-and-play software is not desirable, because these properties are of no benefit to you. Open means that you can interface with it, which means a lot of work in coding dependencies. Standard means the software is years old and the vendor had no ideas of his own to sell. Non-proprietary means that everyone else they interface with has the same lack of new ideas. Plug-and-play means that the software can not be adapted to your needs.
Companies that boast huge service departments and consultancy teams are of no benefit to you. These departments are there because you need them and means the software cannot be used without a huge service expenditure. If a software company starts to grow in services more than in software revenue then their software is at the end of its life-cycle.
A long list of partners is a reason to be cautious. Companies that partner with every Tom, Dick and Harry can not stand on their own. Their software is so limited in its solution scope that it brings no benefit if you don‘t buy all the other stuff as well.
The claim of SOA compliant software – don’t get me started on that. It is such a poorly defined term that anyone can claim to be SOA. The idea is good but see my other articles on SOA drawbacks.
The word AGILE should be prohibited when even a single line of Java code has to be used. If you involve a programmer there is no agility.
XML standards are always claimed but a pain in the neck. What is so great about two triangle brackets and a slash? XML is a huge problem because of data size and processing time. I was told by some companies that more than 70% of server time is used for XML validation alone. XML is a monster because of its lack of a metadata concept. You do need many other definition structures to make XML work and once you do that the flexibility that seems so cool when you can read the file goes out the window. If you manage metadata the right way, then message formats are irrelevant.
WebServices is right up there with SOA …
Easily customized, pre-configured and out-of-box is pretty good if it is true. That should mean that the software can be installed and set up in a few days. Don‘t believe it but try it. Once installed ask for certain aspects to be changed to your needs. If that requires programmers to come in then get rid of it as fast as you can.
I have now tried for years to truly understand what IT sells with On-Demand exactly. Sometimes it is a flexible licensing model or charging per use rather than per install. Sometimes it means a flexible deployment model or an outsourced ASP approach. Sometimes it means nothing.
Full Life-cycle is an often misused term as well. It should only be used in terms of application change management. A full life-cycle must provide a complete roundtrip from analysis, modeling, coding, documentation, version controlled deployment, tuning and maintenance without loosing control in intermediate stages or being unable to feed later changes back into the model stage. I have not yet seen a full life-cycle tool that uses Java and XML where intermediate coding is fully reloaded into the model stage and managed there. The same applies to most BPM tools. The models have to be encoded with a lot of Java code that remains separate.
The misuse of the words knowledge and intelligence in IT is appalling: Proactive intelligence alerts, enterprise knowledge databases, and real-time predictive intelligence are no more than pure BS. Knowledge systems are text databases and Business Intelligence is no more than statistical data collection. Knowledge still means nothing else than knowing what-to-do in a given situation and intelligence is clearly defined as the ability to learn by itself. Maybe the intelligence agencies had to reduce that meaning for good reason …
‚We have a repository to store all documents.‘ A repository is not an archive. A repository is an application development tool that more than anything else has to version-control metadata and related source code, ideally over the full life-cycle. A registry on the other hand is just that. A queriable list of entries without further functionality.
I think that this is enough for now. I propose for vendors to be more accurate in what they claim and for buyers to be more prudent. They call me jaded, but I am still standing and fighting.