Nothing really comes free
Selecting a software solution primarily based on cost might appear beneficial in the short-term but will almost certainly cause issues later. While using open-source/in-house software may seem compelling and, in some cases justifiable, factors other than the purchase price will affect the financial outcome and so need to be considered before a software platform is selected. The "real" cost of software is influenced a number of factors:
• the training and implementation costs
• product development time and time-to-market
• complexity of servicing and maintaining the software (how much debugging and who does it?)
Providers of proprietary software are more likely to cushion the financial impact of all the above.

The down-side of using open-source/in-house Vision software
Using open-source software delays product development, evaluation, testing and documentation, affecting both the total cost and the time-to-market of the project. On the other hand, proprietary software typically allows fast prototyping and comes with the most complete library, both providing quick reaction time and shorter development time and, therefore, contributes to cost-savings.

There is no guarantee that open-source software supports latest technologies; nor is there any way quick way of determining if the tools provided are fast enough. Also to be considered is the effect of a change in hardware/operating and whether or not open-source can be carried across smoothly to another system.

Being dependent on your own in-house or contracted open-software developers brings with it significant problems. Open source developers that understand both the industry and the software are not only difficult to find but potentially difficult to keep. If the developer is no longer available, how long will it take to find a substitute? And if one is found, how easily can another developer make use of the existing code, particularly if there is no good documentation of the open source code to guarantee a seamless and quick transition?

The main users of open-source Machine Vision software are academic users who use it for research and do not have a commercial imperative. There are very few industrial users, as few commercial companies risk their investment with open-source software. In fact those commercial companies that take the leap with open source almost all inevitably revert to proprietary software over time once they experience the shortcomings of open source.

What to consider

What is the life-cycle cost?
Undeniably, the initial cost of buying proprietary software is greater than using an open-source version. However, the calculation of the "real" cost of software includes a far broader range of factors, aside from the initial acquisition cost, and when these factors are considered, using proprietary software becomes the best, cheapest and more risk-free option. Life-cycle costs, additional to initial purchase, to be considered are:

Training
Training engineers to learn software can be significant especially if there is time pressure. Good documentation makes it easier and reduces the time and cost. Documentation of proprietary software will be professional and detailed and the vendor being the best port-of-call when difficulties arise. Their intimate knowledge of, and familiarity with the product will provide a faster reaction time, enquiries often being pre-empted with FAQs. It is also possible to receive customer-specific training and advice on what tools to use for an application. Images can be sent to the vendor and tested. Advice is provided free-of-charge.

Machine Vision software companies offer up a great number of tools to prevent integrators having to reinvent the wheel. These tools are being continually improved and updated with the latest algorithms. Additionally, software companies offer free application evaluation by providing a prototype solution which can save significant time. This is proprietary software's biggest advantage over in-house development.

Proprietary software will generally be of higher quality, more user-friendly, more convenient, sustainable in the long term and easier to implement. It will allow quicker time to market and algorithms used will be researched and complete. With some software, interfaces can be created by non-programmers, using point-and-click, that are fully-customised to the needs of that particular user resulting in more money saving and a quicker turnaround. Upgrades/fixes are published and downloadable as are high-quality, fully-documented manuals. With proprietary software, extensive documentation allows new developers quick access to a comprehensive tool library.

How much could have been saved or earned with a better decision earlier on? A tried-and-tested system will permit very fast prototyping using a complete library of tools and with all the research for new algorithms already performed by the proprietary software company. Time-to-market is decreased.

Often specialised equipment needs specialised software. What is the cost if the project fails or takes too long due to the use of inferior software tools?

After the outlay for the initial purchase, further projects with proprietary software become comparatively low cost.