The path to a smoothly run business may be the yellow brick road of Oz. Ian Scales takes a stroll.

Oz doesn't use a single worksheet of rows and columns, but separately displayed files of cells for individual entities - sales, wages, office expenses, etc. All these fit into a hierarchical relational structure based on parent/offspring/sibling relationships between departments within an organisation.

Oz provides a coherent structure to monitor closely and understand a company's particulars as well as the board sweep of its performance. You can make notes on variances in figures to accompany an entry (e.g. 'dock strike incurred delviery penalties' could accompany a variance in sales profits in a certain month).

Overall performance can be determined by consolidating data into profit and expense totals. Data can be analysed by variance and comparison and the results displayed by graph. Reports can be generated and stored and formats recalled. Communications facilities are included to allow the importation of files from Multi-Plan, Lotus 1-2-3 and others.

I had thought that the days of makeshift documentation were over but, unfortunately, Fox and Geller seems to be a software traditionalist. The first batch of programs is being sent out with photocopied instruction manuals and a correction sheet with no fewer than 18 updates although typeset versions are being prepared as I write.

Sadly, Oz is let down by a single, rather confused manuscript which tries to cover everything in the 'do this when you're doing that and if that happens then do this except if this has already happened then you're in trouble and you have to do this' style. A tutorial on disk with supporting guide has been included as an attempt at balance, but it would be nice to see as much information on the application itself as on the nuts and bolts of the program.

Oz has the normal help facilities and on-screen prompts and the whole thing (documentation aside) functions very efficiently.

You are greeted with a series of menus on your way into the system which tell you the model is created from the top down. The lines files section displays the entities for revenues and expenses - products, salaries, rent, etc. - and further computations are expressed in this display to identify net and gross profit lines, for example, or total expenses.

An organisation chart is provided and defined by the user. At the base of the chart are the elementary organisations (the departments with no sub-departments) and each of these has an accompanying data file to hold the various facts and figures by which the organisation's performance is measured - for example, the sales department will have a string of monthly sales figures (budgeted, forecast and actual) attached to it.

The non-elementary departments are the parents to the elementaries - these consolidate data from their elementary child dependents so overall or regional performance figures can be intercepted by the user on their way to the 'bottom line' at the top of chart - the net profit line in the lines file.

The consolidation, reporting and analysis functions tend to be very disk-active and are not to be entered into lightly.

The above description illustrates the complexity of the package, but the real issue in evaluating an application as different as Ox is not to ask "how well does it do its job?" but "how good a solution is the job itself in a typical business situation?"

Oz is really about structuring, in one program, a whole range of functions which would normally have to be 'tailored' by the user, perhaps from a spreadsheet or database (low-level but flexible), or an integrated software package.

However, Oz provides a solution for one problem rather than a tool for many and I suspect that companies not dealing in neat, statistically-valid volumes of sales over monthly periods (especially enterprises involved in providing high-value, low-volume goods or services) are going to have a few problems shaping the system into a usable form. It's so structured there seems little room to manoeuvre.

Oz is a very ambitious program. It's obvious that it could find a happy home in the middle on an integrated program suit and it's not beyond the bounds of possibility that somebody will write something very like it for Lotus' Symphony.

The worrying thing about Oz is the necessary degree of rigidity in the package. As I mentioned before, Oz has to make certain assumptions - the existence of a predefined company structure, the likelihood that it is going to be selling items, hiring employees, comparing projected, actual and budgeted figures, and so on. In many situations Oz is undoubtedly going to be a great asset. It's very powerful and well thought out and its processes make great sense for the textbook sales-based company. For companies whose operations happen to be less standard I think it's worth a careful look - with the emphasis on 'careful'.

Features 5/5
Documentation 1/5
Performance 3/5
Usability 2/5
Reliability 3/5
Overall Value 3/5