“How easy testing used to be, no fuss about privacy, just use all kinds of live data and real links and so on you went…”
This is something I sometimes hear people of my generation remember (yes, sorry, IT grandpa speaks). I have been using synthetic data for decades whenever possible (Micky Mouse was very popular for a while). But there was regular discussion about “why should we make it difficult” and “what risks do we really run”. I won’t bore you here with the answers I have gathered based on years of experience of what can go wrong.
Fortunately, today there are clear rules about privacy of test data. With the rise of Generative AI (GenAI), its application is becoming even more important. So, do not use production data and examples from your live system!
But what should you do?
At this point, GenAI is both a curse and a blessing. A curse because if you want to be sure that sensitive data is leaked, you just have to enter it in ChatGPT and the like (so never do that!). But also a blessing, because the same Large Language Models are quite good and helpful in generating synthetic data. Write a nice prompt and you will have an abundance of meaningful and compliant data.
But sometimes you also need other things. For example a website. Recently I spoke with some people in a client project about the need to arrange a URL ourselves (register a domain name) in order to have a website for testing. Our intern then intervened (talking about generations), with the, for him, very simple question: ‘But why don’t you use www.example.com?’. Huh, what… It’s nice that an IT student has to explain to me, as an IT professional with over four decades of experience, that websites for this purpose have existed for a long time. Did you know this?
A quick inventory showed me that you have example.com, example.net, example.org and various country-extensions such as example.nl. All free to use, for example in your tests. And you can use it for email addresses too! If you send an email there, you will receive a different error message than with a non-existent domain name or with an email address that does not appear in an existing domain. The error messages are:
- Non-existent address on existing domain: Recipient address rejected: User unknown;
- Non-existent domain: Failed – DNS error: Permanent DNS error resolving XDOMAIN;
- Email on example.com domain: Failed – DNS error: Recipient address has null MX.
So, this is very useful for dummy email addresses. If you already knew this, great, keep it up. And otherwise, you have an extra example to use and make your tests even better.
I wish you a happy, healthy and high-quality 2024!