Who cares about user interfaces? (by alaric)
User Interface development in the 2000s
The turn of the millennium was an exciting time. The Internet was everywhere, and open source software development had exploded in its wake. Mac OS X was about to come out, and the Mac would be reborn. The year of Linux on the Desktop was surely imminent, as the GNOME project - bear in mind that GNOME originally stood for "GNU Object Model Environment" - attempted to build a distributed object system, with a user interface on top of that, allowing for advanced inter-application integration.
But, it all fizzled out. Microsoft had claimed a firm grip on the desktop computer market, and seemed to be pushing out versions of Windows with mainly cosmetic changes to the user interface rather than real attempts to improve it; their attention was much more on fighting the Browser Wars (more on THAT below...); I suspect that, now the windows UI had become so widely deployed that nearly anyone who wanted to use computers had learnt it anyway, they had no motivation to make it easier to learn, easier to use once you'd learnt it, or able to do more; and competing GUIs moved their emphasis towards "Looking exactly like windows" so they could reduce the switching costs of Windows users, rather than trying to make their systems easier for new computer users to learn... or better for experienced users.
Mac OS X was, similarly, fairly similar to classic Mac OS; updated to meet the capabilities of new hardware with fancy transparency effects and lots of pretty stuff, but largely still the same WIMP interface at heart. However, once thing I liked was that core personal information management stuff like contact lists, appointment calendars, and TODOs were integrated through a common API, meaning that applications that touched those things all integrated with each other reasonably nicely.
A lot of the talk at the time was about how user interfaces would develop to allow more inter-application integration. Microsoft's OLE was kind of cool, letting you embed Excel spreadsheets into Word documents and... well, yeah, that was the main thing people did with OLE, although people seemed to be thinking about how it could be used elsewhere. This inspired Apple to create OpenDoc, which was based on CORBA, a standard for inter-application communication (those Distributed Objects that GNOME was originally interested in - and original GNOME was also CORBA-based). I feel the focus on embedding and linking data between documents was a bit of an artefact of particular world of office software; nowadays, embedding images and spreadsheets and charts into documents and presentations is really just a bespoke feature of word processors and presentation software, because there's actually only a few combinations of those things people really want to do. I remember an example of what OpenDoc embedding could do, in the form of an application for managing the loading of cargo into aircraft; a container application provided an outline for the cargo hold and objects representing different types of cargo, provided by applications specialising in that type, could be dragged into position, with the container app enforcing that the packages would physically fit in the space. This seemed a bit contrived; so I'm not surprised OLE and OpenDoc were kind of dead ends, and I'm sad to say that I think the centralised personal information management stuff in OS X was the last bit of real growth in desktop user interface design.
One big reason, perhaps, that the development of desktop UIs stalled is that everyone's attention was drifting away from them - to the Web, and later the Smartphone.
I'll talk about smartphones first, even though that came later, temporally.
Tue 8th Jul 2025 7:39 pm 
Subscribe