Newsletter

Highlighting some gems in STEM 7.4

16 February 2015

STEM 7.4 was released on 08 December 2014.

The features of this new version have been outlined in earlier articles, but some of the new extensions warrant a closer look now that the actual software is out there.

Much has been made of the ‘solution to the right-click problem’, and our early adopters are now gradually adapting to the simpler and more intuitive approach for accessing time-series parameters when required (rather than all too often by mistake as before). In this article we are going to play a small fanfare for the new Find function, and also take a closer look at some subtle but very satisfying changes to the way that charts are presented in the Results program and on the web.

Navigation in an unfamiliar and/or large model

The models we show in training or demonstration examples are typically small enough to fit on one screen to minimise unnecessary scrolling in a learning context. The reality is that many models accommodate disparate markets and service types, expand to cover multiple network functions, aggregation levels, and sometimes technology variants, and typically evolve to consider installation labour, power, rack and office space across many of the original network resources.

In practice therefore, even the most modest view will outgrow the limited confines of the screen (especially when working away from the desk on a laptop or tablet screen), and for reasons of convenience and modularity, it will make sense to break up a model into separate views for ease of navigation. It is also common to include additional views which allow for different perspectives on some of the same elements; for example:

  • a high-level summary view showing the principal elements and drivers
  • a summary of the scenario structure showing only the affected elements, removed from their original context
  • a collection hierarchy showing how costs are grouped, again separated for clarity from the original dimensioning context.

For all of these reasons, it may not be immediately apparent where to find a particular element you need to review or edit, or even in which view (or views) to look. After some time away, this can easily be an issue with a model you have developed yourself, and it is more frequently an issue if the model was built by a colleague. (The latter context is how you have to imagine our support staff working most of the time!)

So, the principal reason for adding the new Find function is to help you locate and swiftly access an element by name (or sub-string match). The fact that you can also search for text in view names, text boxes, formulae, values and notes is almost an afterthought!

At its simplest, the process to locate an element called DSL access is as follows:

  1. Press <Ctrl+F>. The Find dialog is displayed.
  2. Enter the name, DSL access (or as many characters as required to get a close match), and press <Enter>. The dialog immediately lists every matching element (and view name and text box) and the first match in the current view is highlighted.

Figure 1: the first matching element is selected in the current view

  1. Press the Next button to highlight the next match. To minimise jumping about between views, by default the sequence will progress across and down the current view (and then from the top-left back to where you started), and then in the same manner through other views with matching content in turn. (You can group the list by element if you prefer.)
  2. Alternatively, if you can see what you want in the list already, you can double-click it to go straight to that item in the relevant view.
  3. For audit and review purposes, you can also pick several elements in the list and click Show in a New View to display just these elements in a new view which you may or may not choose to discard after a specific check is complete.

It is easy to get distracted by the wealth of options for how and what to find, and the difference between searching in views or data dialogs (or both), but the principal benefit is to not waste time finding or even scrolling to that elusive DSL access! If you are interested in all the other stuff, then just press <F1> in the Find dialog to access the help system where the options and commands are explained in full.

The obvious extension to support a selective Find and Replace is already in the works; we just have to figure out the best remedy for when a modified element name would be no longer unique, or a modified formula would have broken syntax!

How to label line and column charts

“At the risk of stating the obvious, the x-axis labels on a time-series chart show the periods associated with each of the values shown on the chart. On a column chart, the labels are positioned mid-column; on a line chart (for the overwhelming majority of end-of period results) the labels follow the data points (up to the end of the chart).”

Figure 2: a chart displayed by STEM 7.3

Not any more! (Never be scared to think the unthinkable.)

In earlier versions of STEM, a chart would show vertical gridlines for all periods by default in order to highlight, particularly on a line chart, when results for quarters or months were available. These gridlines were all classed as ‘major’.

When developing equivalent charting functionality for business models on the web with eSTEM, it became apparent that it was unnecessary to display labels for individual quarters and months when these could be inferred readily from the corresponding year labels. The overall presentation is much clearer without the plethora of over-long sub-annual labels like Q1 2015 and Q2 2015 all competing for space!

So, we decided to carry this logic over into the C-STEM Results program and suppress the sub-annual period labels there too. We also re-invented the sub-annual vertical gridlines as ‘minor’ in order to differentiate them from the major gridlines for years. In short, the minor gridlines are drawn with a lighter shade of grey so that, in a non-intrusive way, the years stand out more clearly.

Figure 3: the equivalent chart displayed by STEM 7.4

For a column chart, when you have four columns for successive quarters in a year, it is not hard to see that the year label should appear midway between the four columns, exactly where it will be if you switch to showing consolidated annual values and the four original columns are replaced by a single column for the year.

On a line chart showing only annual values (e.g., end-of-period connections), we would originally have shown the year label under the value, even though you might regard ‘the year’ as being everything between the previous end-of-year value and the next. This becomes more apparent when you add quarterly values and so now we place the year label ‘mid-year’, which will be under the (end of) Q2 value if quarterly values are shown.

This may sound convoluted, but in simpler terms it also means that the year labels will now come between the major grid lines which mark the ends of each year. It also means that Y0 will not be shown on a line chart anymore because the first label will be Y1 under the portion of the chart which shows the evolution of the data in from the end of Y0 to Y1. This, in turn, nicely meshes with the separate decision to suppress Y0 data on column charts by default. (Typically one is only interested in the movements, revenues and costs from the beginning of Y1 onwards, and in Y0 only as a starting position for line charts.)

Note: if you think there must be occasions when it would be helpful to show quarterly or even monthly period labels, then yes, you are right! STEM will display months if a model runs for only one year in months, and quarters if it runs for only two years. The rationale is not more than around ten period labels on the chart, and only eight for a run period longer than one year because then the individual labels have to include the year too!

In fact, we have substantially re-invented much of the presentation logic for charts in the results program, making it far easier to develop and apply a style to an entire view or workspace, and also to work with colours from a custom palette. Most of this is pretty self-explanatory and will be intuitive as you work with it. Again, you can click through to the online help to read more of the details.

STEM 7.4a already

It had to happen. Customers find things we don’t. This is a fact of life, making the first point release for 7.4 as inevitable as a SP1 for any other software product. A number of minor irritations and cosmetic errors have been identified and will be addressed by a STEM 7.4a release in February 2015 which will be notified directly to recent users as soon as it is ready.

       Buy SSL

© Implied Logic Limited