His workshop started with some improv games based on the work of Keith Johnstone which, by exposing our awkwardness, showed us that we were conditioned to behave in certain ways, that we have patterns of operation. As testers we want to be free to think, investigate, explore, perform.
A second round of exercises had us giving and receiving imaginary presents to illustrate the notions of offers and blocking. Here, one person creates a context for the other (the offer) which can be accepted or rejected, but both parties must be aware of the ways in which they might constrain that context (the block).
For example, I might mime the shape of something to pass to my partner and then, as their hands reach for it, change the shape. This constitutes a block - I am not collaborating as fully as I might. Blocks come in many varieties; the receiver may block by not accepting the gift or refuting some aspect of the context.
We formed small groups assigned to apply the notion of an offer - with no suggestion about the ways in which we might do it - to testing a task management application. Here's just a few of the thoughts I noted down, pretty much raw out of my notes:
- every interactive component of the application is an offer.
- the user interface, user experience, terminology, documentation and all other aspects of the product are offers to make a judgement about the software, its quality, its value to the user, its function, its domain and so on.
- offers may be implicit or explicit.
- is there a difference between an offer that is recognised as such by the receiver and one which is not?
- some offers are compound; a form has a submit button but also fields that can be filled in. The fields are individually offers, but the whole is also an offer.
- some offers are conditional; a particular field in a form might only be available when other fields are populated.
- it is frustrating when the the relationships at play in a conditional offer are not clear. An offer that appears and is then removed for reasons the reciver doesn't understand is distracting and frustrating. The receiver feels let down.
- when we saw some offer (say, a date field in a form), our first thought was often how "can we accept this offer in a way that violates its likely intent?" (say, a date in the past for the start of a task).
- is the receiver blocking when they accept an offer in a way not intended by the giver?
- an offer that doesn't obviously result in some change is confusing to the receiver; for example, pressing a button but seeing no obvious consequence.
- the likely consequence of accepting some offers is clear, but in others we're taking a leap of faith. The error dialog that says "You tried to do some complex action, but there was a problem. Do you want to continue? Choose OK or Cancel" doesn't help us to understand the consequences of accepting the offer.
- rejecting an offer is not a null action. It still has consequences.
- accepting or rejecting offers can have unintended consequences. When multiple groups were testing the same application we were (probably) changing each others' data, resulting in some confusion (to my group, at least, until we had a hypothesis).
- inconsistency of offers is confusing. Multiple different ways to report form submission failure; different icons for the same functionality; the same functionality under buttons with different icons; use of colour in some places for some data, but not others. The receiver doesn't know what to make of offers that are apparently similar to others in some respects - should they expect the same outcome or something different? This is a kind of block.
- an offer that is taken up (say, a form is submitted) but then results in a block (say, a validation error) is unpleasant for the person who accepted the offer. It is possibly more unpleasant than an offer that is taken up only after all negotiation on the terms of the offer has been done (such as when fields are validated during input).
- offers are always choices. If nothing else the receiver can accept or reject. But they are often more than binary, even in simple cases like an OK/Cancel dialog with two obvious buttons there may be a close button in the title bar, keyboard shortcuts for cancelling (often Escape), different ways to navigate the dialog (e.g. tab, shift-tab, using space or return to select a button, or using the mouse); the dialog might be modal or not and if not, the offer is deferrable.
- offers can be thought of as nodes on a graph of the testing search space. And the reverse: any node on a graph of the search space is an offer, although not necessarily one made by the software, but perhaps made by the data or the tester, or some external context or constraint (such as time or project priorities).
- deferring choices is a kind of blocking - is it important to defer consciously?
- noticing, and accepting, offers is a way of breaking patterns of behaviour. Perhaps I always get to the admin page of some product by opening it, clicking on the Tools menu and selecting Admin. But the product offers me many other ways of getting there - I can create a browser bookmark for that function; I can customise the toolbar of the application; I can launch the application using an Admin-only account. Accepting the offers puts me in a different context, ready to see something different(ly).
- There's a literature on human psychology around giving (also giving up) and receiving. How much of this could be relevant to human-computer interactions?
- I like to give software the chance to demonstrate itself to me. Am I making it an offer?
- what can I do to avoid being overwhelmed by the explosion of offers?
That possibility of a different perspective, a new view, a cleaner vision is incredibly exciting, but until I've used the tool some more, built and broken something with it, uncovered some of its foibles and fortes and put some sweat into its handles, I won't know whether it's a microscope, telescope, a prism, a mirror, a window, rose-tinted spectacles or a blindfold.