Some Discussion of HPM
PP13 - 18


The Database Approach.  HPM states that most medium and larger firms have been in a transition towards consolidating their databases so that it can be shared and used throughout an organization.  This certainly seems true in my experience.  Information is being used more and more as a competitive weapon.

Some Major Advantages.  Then HPM lists some of the major advantages of what they call the database approach.

  • program data independence - data descriptions are stored in a central repository - this helps make it so that data can change and evolve (within limits) without changing the applications that process the data
    • I consider this an ideal that can often be achieved (within limits)
    • this was much more difficult to achieve when certain data or the nature of the data wasn't very well known to everyone that really could use it
  • planned data redundancy - organizing the data so that it is centrally coordinated and controlled makes it less likely that there are needless redundancies
    • for example - you are less likely to have more than one group collecting the same data
    • planned redundancies can occur for a variety of reasons - but it is desirable to have these occur due to intentional choices rather than lack of organization and awareness of what is going on elsewhere
  • improved data consistency - coordinating the data makes it much less likely that there will be conflicting data within the organization
    • for example - do you really know a customer's current address?
    • for example - if a customer's address changes then it is done with much higher validity and accessibility
  • improved data sharing - obviously if the data that is available can be known throughout the entire organization then there will be much better sharing of data and possibly insights gleaned form the data and what other data should be collected
    • user views - developed so that a user can examine some salient subset of the data - usually done with a query or report
  • improved application development efficiency -
    • if data is already available and well organized then the developers can more easily make use of it and design their applications appropriately
      • if certain data isn't available and is needed they are more likely to be able to start getting it
    • a lot of tools will very likely already be built into the DBMS
      • report generators
      • query tools
      • end user form development
    • if data can be securely accessed through the Internet then this increases availability considerably
      • interface development can also be done using fairly common and well known approaches based on HTML, JavaScript and some sorts of middleware
        • much less likely to need developers for fairly obscure languages
  • enforcement of standards - this is generally a strong advantage - when approaches are no longer fragmented and isolated standards can be more easily set that apply to everyone and work to the entire organization's advantage
  • improved data quality - this is a huge issue if you've ever really worked with data - data quality is never as good as it should be
    • who entered it?
    • were mistakes made?
    • can procedures be implemented to improve the quality?
  • improved data accessibility and responsiveness - this should be apparent from the other discussions
  • reduced program maintenance - this is possibly more theoretically true than really true - but generally these sorts of approaches to database should ensure that at least the data sources and results meet certain standards.  This helps make maintaining programs much easier at least for these aspects.
  • improved decision support - some databases are expressly designed for decision support. 
    • think about how data mining of sales data should be much improved.
    • think about how someone should be able to access all the data they need to analyze the entire supply chain and not just fragments
    • financial analyses should also be improved
      • what are the real costs?
      • what is really making the firm money?
  • organized central source of expertise - this can really allow an organization to have high caliber database administrators and network/security administrators working out of a synergistic single source
    • unfortunately - it does happen that this central source can be overly controlling - and worse - constantly use their specialized knowledge to unfairly impact what projects get done and which do not even get started - I've definitely seen these central sources work very hard to do as little as possible while trying to make sure no other expertise develops elsewhere

When I was a professor at Rider University in NJ, I worked on some analyses to help them predict which students that have applied would actually matriculate in the early 1990s.  I was astounded at how uncoordinated and unreliable most of the data was.  Admissions had its own data sources.  Financial Aid was on a completely different system.  These didn't really even tie into the system that was coordinating course registration, fees and degree progress.  There were plenty of other problems. 

Many of these problems were straightened out by using a single system to organize all of this sort of data within the University.  They chose to go with Datatel.  But unfortunately, this system wasn't all that user friendly and it definitely did not have a Windows interface.  But at least by having one coordinating system they were much more confident that they knew where to find things.  They also took fairly strong measures to help ensure that the data that was entered was reliable.

But I should say more.  I was on fairly good terms with the system administrators at Rider.  They were all truly experts in VAX - VMS systems, and it was easy to respect their intelligence.  So there was one administrative VAX super-mini and two academic VAX super-minis.  Unfortunately, even though this was the 1990s we could not get them to move towards UNIX or Windows.  So the implementation they chose to use of Datatel was really quite primitive in comparison to what they had for other platforms.

Now for another interesting question/issue.  When Rider decided to totally revamp their library systems how much should they be compatible with their existing systems?  Rider Libraries did choose to go with a Windows based system that was largely entirely separate from other academic computing foci.  They fairly cleverly chose to be a test site for one of the first Windows based library information systems.  They got a great deal and were confident they would get great treatment because the company doing the development really wanted to use them and one other school for referrals and examples of how their systems really worked.  This plan appeared to work like a charm to me.

For example, the computing group within the faculty had two of their own Windows servers and a computer lab focused on teaching computing courses.  These Rider administrators were usually quite reasonable to work with when it came to these sorts of needs even though they wouldn't implement these sorts of things in their own central data center.

Some Major Disadvantages.  Most of these disadvantages can also be turned into advantages.

  • new specialized personnel
    • how expensive are they?
    • who will manage them?
    • how much are they actually contributing to the organization's competitiveness and functionality?
  • installation costs and complexity
    • different approaches require higher costs and complexity
  • need for explicit backup and recovery
    • in my mind more of an advantage than a disadvantage
  • organizational conflict
    • probably one of the biggest problems relating to human, or maybe I should say inhuman, quests for selfish power and money
    • who gets the leverage of being in charge and control of the database systems?
    • who might get left behind if they do not want to progress?
    • do the systems really meet the user's needs?

I am leaving out the discussion of traditional file processing systems for a number of reasons.