Improving MRDCL productivity by using functions and subroutines
How to automate things that are boring to script …Read more
I have always admired Quantum as a market research analysis program. It is still powerful enough to process most survey tabulations, but after almost 20 years without development, its position has slipped. Whilst still functionally strong, modern releases of Windows do not support Quantum. The alternatives are few with MRDCL, Merlin and Uncle being the main choices.
Quantum’s lack of development means that it cannot read everyday tools like Excel workbooks. So much information is conveyed in workbooks – brand lists, non-survey data, coded data, lookups, code lists etc. – it’s a necessity, in our view, that tabulation software can read Excel workbooks without having to paste the lists into text files. Merlin and Uncle have not added such important tools.
If our sales office is anything to go by, we are experiencing a rapid increase in the number of enquiries we have about switching from Quantum to MRDCL. Old machines running old Windows versions have been kept running, but at some point they need to be replaced.
Switching from any serious software system in any business is not a trivial issue. Legacy systems mean that caution is needed and there is a natural temptation to postpone the day when change takes place because there will be some hardship. The three key questions are: -
This can often be the single biggest reason for delaying the inevitable move that you will have to make from Quantum. And, there’s good reason to fear this move. Transferring any complex programs or project to another system can be a time consuming process. Quantum scripts can contain complex logic which may be laborious to transfer to another system. Further, it may be easy to make an error when converting; errors which may prove difficult and time-consuming to track down.
There are some key tools at hand to help you. There is a free Quantum to Triple-S conversion program available on the Triple-S website at http://www.triple-s.org/software/utilities/. This allows you to convert a project to Triple-S so that you can import it using a free utility into MRDCL script. MRDCL’s unique design means that you can read old data, Quantum variables in this case, alongside new data, which may be in a different format. The advantage of this approach is that you can export Quantum variables exactly as they are without worrying about complex logic, data recoding or any other programmatic manipulation of data.
If you decide not to make the conversion via Triple-S, you can use raw Quantum data within MRDCL exactly as it stands. Quantum supports its own binary data format as well as reading ASCII data. If you want to read raw data from Quantum, the only thing you will need to do is give the files a suitable file extension as MRDCL recognises file types by their file extension (it does detect most errors if you give the wrong file extension). The disadvantage of using the raw data is that any complex logic will need to be transferred to MRDCL script, but this may not be a major consideration for many projects.
Quantum binary data needs a .cbe file extension and ASCII data needs a .asc file extension. Strictly, MRDCL will treat any unrecognised file type as ASCII, but it will ask you to confirm that it is ASCII each time you run with the data file, so it is better to rename it with a .asc file extension.
Anyone who can master Quantum can master MRDCL. They are two scripting languages that both have good set of tools for dealing with more or less any market research data and tabulation needs. Understanding what is needed and the disciplines of being a good market research analyst are the most important attributes.
However, there are really two parts that you should consider when staff convert from Quantum to MRDCL. 1) How long will it take to become competent in using MRDCL? 2) How long will it take to make the most of MRDCL? These are very different questions.
To answer the first question, it shouldn’t take long. There will be a brief frustrating period when someone converting to MRDCL will make minor errors due to the specific syntax of MRDCL, but regular use soon brings familiarisation. It’s the second part of the question that is arguably more important. I say this because I have seen some Quantum converts to MRDCL learn how to do things more or less the same way as you do things in Quantum. This might be a good way start, but it means that the experience is likely to be less than satisfactory.
Let’s say, for example, that MRDCL has 50 better ways of doing things than Quantum and Quantum has 10 better ways of doing things better than MRDCL. It can mean that someone who does not explore more advanced and modern or different approaches offered by MRDCL will just see negatives. It’s important to embrace the modern features that a product with almost 20 years’ extra development, responding to user demands, can offer. MRDCL users switching from Quantum can expect to find some significant productivity gains.
One of the main, but less obvious, benefits of MRDCL is that it enables team work and project sharing. 20 years ago, MRDCL was like Quantum. It used to frustrate me to see highly skilled data processing staff spend much of their time entering simple script. MRDCL’s modern approach means that templates can be built so that less skilled staff or colleagues can share projects. Costs are reduced substantially as less skilled staff can handle the easier tasks within a project, leaving the higher paid skilled staff to spend most of their doing skilled work. It can mean a change of approach within an organisation but one that offers good financial benefits.
For simpler ad hoc projects, switching to MRDCL will be fairly easy. For more complex projects, learning the best approaches in MRDCL will be highly beneficial. MRDCL offers more solutions than Quantum in general, which means learning to pick the right approach rather than the most obvious approach can be important. We try to make this easier for new users by critiquing their scripts so that better approaches can be learnt. The biggest problem is moving large multi-country or tracking studies where there is complex logic in place. I have always been happy to advise on the best approach for such projects as there are a number of potential routes that can be taken.
Commitment is the key to switching systems. MRDCL will offer a lot of advantages, improving productivity and offering tools that mean that teams can work on projects rather than individuals.