Graphic design software has steadily increased in power. At first the focus was on improving the visual expression capability of the software but gradually, with the addition of work-flow tools, productivity has become important. However, today’s software is limited by a direct manipulation mindset.
Direct manipulation interfaces allow designers to tweak individual elements on screen by using a mouse and keyboard (and/or commonly a tablet). This means that the designer’s vision must be transformed into discrete steps of computer interactions - think of it like the visual equivalent of programming. Any adjustments to the “feel” of a design piece are non-trivial and will often involve many interactions.
What if there was software for designers that worked more at the level of thought, message and feel? There are two areas of research that combine to produce the design software of the future: Generative Design and Aesthetic Science.
Generative design systems take parameters then produce a set of candidate designs for a designer to select from. Sophisticated generative design systems can then use those selected items to further iterate and produce more (and hopefully better) candidate designs. Imagine feeding the text and image library for a poster to software and having it come up with thousands of concepts for you within seconds.
Aesthetic Science tests (usually following Psychology modes of testing) for aesthetics that are universal. Once an aesthetic principle has been established as universal then it can be algorithmically modelled. A computer can run the algorithms to judge design pieces on an aesthetic level. Yes, aesthetics is a complicated area but we do not need a complete understanding to produce computer tools that can still be useful.
Examples of existing work in aesthetic algorithms are Donald Knuth and the TEX typesetting system, auto-balance/contrast and colour correction in photographic software and the Text Colour Contrast algorithms used by the W3C. My own research is searching for algorithms for measuring the design principle of unity. Unity has already been established as an aesthetic universal by fMRI brain scanner research in Europe. My research is in the blue-sky stage.
A criticism of generative design is that the generated candidate designs tend to break too many aesthetic universals. The designer ends up weeding out useless chaff rather than thinking at the level of message and feel.
Future design software could use a generative system approach but incorporate aesthetic algorithms to pre-screen the concepts shown to the designer. The designer can then discard designs, keep designs and even use a semantic differential like system for rating the feel of designs. Then, the generative system can iterate more concepts – for as many rounds as the designer deems necessary. Changing the feel of a design is simply a matter of asking for something in emotive terms : e.g. A bit happier and a stronger sense of community. At the end the designer gets a file they can load into direct manipulation software (like InDesign) for refinement. This process could take less than five minutes.
UPDATE 20130401: Complexity Science (eturnerx blog article) could provide insights that help build a generative design system. Conventional science methods might be too limited.
Welcome to a more human and productive design future. There will be more articles in this series.
I invite your comments. Your reactions inspire me. I hope my ideas are worthy of sharing with your posse.
This is the first article in the Future Design Software series. Read the second article: File|Save is stupid. What should replace it?.)
This is a really interesting article, which I have some thoughts about. Firstly, while I think this is great for productivity etc. I feel that turning the design process into something which is based on system software and algorithms etc is quite a broad approach, which would take out the creativity and individualism of each design piece and turn a designer into someone who just sits in front of the computer pressing buttons – so to speak. Allowing software to ‘design’ based on algorithms and set parameters doesn’t allow for designs which have definite concepts and messages, which often hugely effect the design itself, especially if the imagery and type need to convey an idea or humour etc, rather than just an aesthetically pleasing design with a certain look & feel.
ReplyDeleteYou’ve talked about Direct manipulation interfaces which allow designers to tweak individual elements and that the designer’s visions must be transformed into discrete steps. I personally feel that this is a huge part of the design process for me, which I couldn’t do without. While I like to have a concept worked out on paper before hand, deadlines, budgets and time restraints often mean that the concept can often be just a scribble of a basic idea, rather than a well defined design. Therefore once I’m on the computer working on the design piece itself, the steps I go through tweaking the design often become part of the design process.
I do however think there is a place for this kind of software, probably more at the lower end of the scale, local print shops etc where an Artworker or Mac Operator (rather than a Designer) isn’t looking to design a piece of work with a concept behind it, they are more looking to create something with a certain look and feel, with the necessary typographic details and imagery inserted.
I’m looking forward to reading any future articles on this topic, I’d also be interested to hear about your own research as well.
Cheers
Adelle
Thanks for the feedback. Many of the issue you raised I'll write about in more detail as in future articles.
ReplyDeleteThe intent is not to take the creativity away from designers - only to have them co-create with a computer. A designer still maintains ultimate control over the selection of pieces but the process becomes more like Art-Directing rather than Mac Operating.
A tool like this is also better away from the low-end - it really does rely on the conscious visual communication skills of a designer to choose the best candidate designs.
I also remember the "uncreative button pusher" allegation used around twenty years ago to describe the new direct manipulation interfaces that were coming in. I don't think direct manipulation is going away - I just think we can be more productive by inserting a step sooner.