If you have Progress skills, and want to work in Brisbane.. WE want to hear from you.

See the Employment page

Home arrow Articles
Recent project update

Some recent projects we completed are worthy of comment:

Query tuning and performance optimisation

We identified some very interesting tricks for tuning queries where the existing indexes are less than optimal, and even developed some useful utilities to guarantee success. Applying these concepts cut one customer's billing process from 2.75 hours down to 20 minutes! In other cases we have measured performance improvements of over 200 times!

Openedge (version 10) upgrade

Another project involved the upgrade of a ADM2 GUI application from 9.1D to 10.1B. This project demonstrated that Progress have made dramatic strides in database performance. The following figures were achieved on a 3 year old single processor W2003 server with a 1.3Gb database (38 million records)

 ProcessTime 9.1D
Time 10.1B 
 Full Data Dump - Data Administration tool
16 min
 6 min
 Full Data Dump - Binary proutil
 6 min 3 min
 Full Data Dump - Binary proutil (1)
  1 min
 Full Data Load - Data Administration tool
  210 min
 Full Data Load- Binary (2)
  9 min
 Full Index rebuild (2)
  6 min
 Full Data Load- Binary (1,2)   3.5 min
 Full Index rebuild (1,2)
  3.5 min

    (1) After migration to mirrored and striped disks.
    (2) type II storage areas

Custom systems vs. Packaged systems dilemma (agile 100% fit vs. 80% inflexible compromise at best)

Custom systems vs. Packaged systems dilemma 

Many businesses recognize the need for better information technology systems to serve their business. Often this can be a result of outgrowing the existing systems or just recognizing shortcomings and inefficiencies.

Once the quest has begun to find a new system, many businesses find themselves evaluating all kinds of ERP systems which can come in many flavours from the small to the mega-large.

The choice of an integrated package for a business is a very difficult and challenging one, and it goes without saying that the best solution will not the same for all businesses.

Customising a packaged ERP system

You will often hear horror stories about how projects failed because the users demanded massive amounts of customisation, the customised version of the packaged ERP system can't be upgraded so therefore the only answer is to go with a plain vanilla package. This is a very complex decision, but we can make some observations:

  1. Packaged ERP systems rarely achieve a high degree of fit without some customisation (most need a lot if you are to achieve savings and efficiencies)
  2. Packaged systems CAN be modified or enhanced in such a way that you do not lose any ability to take on later upgrades, but you have to plan and architect the changes "the right way".
  3. Packaged systems are necessarily far more complex than equivalent custom systems, this is a double-edged sword.
  4. Just the implementation costs for large ERP rollouts (100 users+) can often easily exceed the cost of a custom system and still compare poorly on the degree of fit. (this is counterintuitive but is due to the exponential complexity factor with packaged software)
(an interesting case study - we suspect having been burnt by a previous poor customisation effort Hawker may have gone to the opposite extreme and still not found a good result.  By isolating, planning and designing any customisations "the right way" you can have the benefits without the problems)
The problem with spreadsheets

There is no doubt that spreadsheets are a fantastic tool for analysis and automation of certain tasks. They are easy to set up and use and this ease of use can be seductive.

In many cases the unique requirements of a business that is not fulfilled by a standard software package may be handled by a simple spreadsheet or similar. Beware of this approach when several staff are involved and need to rely on this information.

Getting the right ERP fit

The following is an opinion published in Australian ComputerWorld about the ERP Buy vs. Build question. We tend to agree. 

Reproduced here with the permission of the author.