6 ms·

Tactile pocket computing - iPhone, etc - was a huge paradigm shift for users. The UX as a whole is completely different to the desktop experience. In terms of

Tactile pocket computing - iPhone, etc - was a huge paradigm shift for users. The UX as a whole is completely different to the desktop experience. In terms of software itself - it's not about typing code or about text UIs vs graphic UIs. It's about programming being a fiddly and tedious pastime with explicit handling of every possible scenario, because the machine has no contextual background knowledge and isn't intelligent enough to interpolate or intuit the required details for itself. So you get modern software which is basically just a warehouse full of virtual forms glued together with if/then/else - or simple pattern matching if you're really lucky - managed by relatively simple automation pipelines. Plus a UI/API glued on top. And underneath that is TinkerLand, full of lovely text-langs which are different-but-increasingly-similar and command line environments which are like little puzzles which are not really hard - in the sense of Physics PhD hard - but are hard enough to give people who enjoy tinkering a sense of satisfaction when they (more or less) solve one. You can do a lot with this - like ordering goods and services, selling advertising, collecting creepy amounts of data about users, and keeping people entertained. But it's basically just automated virtual bureaucracy. It has no independent goal seeking skills, it doesn't improvise, invent, or adapt, and it's ridiculously brittle and not particularly secure. Visual programming won't change this. Nor will ML (although it may contribute.) It needs a complete change of metaphor, and there doesn't seem to be much evidence of the research required to make that happen.

fsloth 6y ago
I think you managed to succinctly verbalize and create concrete form to what I've felt. Especially the "faux-value-add" quality of many software systems due to their puzzle like nature rings very true. Perhaps software cannot be nothing more. But there is lots that _can_ be done with virtual bureaucracy. For example 3D modelling tools do enhance the human capability to rapidly design complex forms. Text processing tools beat hands down typewriter in several tasks. Etc.
Roritharr 6y ago
"automated virtual bureaucracy" Oh boy you hit the nail so hard on the head that my spine is still tingling. In the past 15 years of my career writing business applications, websites etc. I always had this nagging doubt that this is just me being a very active part in building moloch, adding to all the bureaucracy instead of reducing/simplifying it. Simply because wherever we built something to make things more efficient/easier, those gains are usually swallowed and overshadowed by the demand for more complexity that arises from latent potential that has been opened up. "we've made taxes easy!" "...could you also make tax evasion easy?" "yes, but it would be more complicated" "More complicated than before you were there?" This is just an made-up example but that's how it so often goes and it really makes it feel like I should have become a hermit and wrote indie-games while living on cupnoodles instead.
- dustingetz 6y ago
  yes the purpose of automation is scale, note scale is what enables big salaries
alisonkisk 6y ago
UI will always look simplistic because it has to communicate with users. But beneath that tiny surface is a huge depth of software, usually on the server. Think Google Now/Assistant, notifications, recommendation systems,and all the semi-automonous server management stuff like Kubernetes and EC2. Also non-consumer stuff like protein folding.

fsloth 6y ago

I think you managed to succinctly verbalize and create concrete form to what I've felt. Especially the "faux-value-add" quality of many software systems due to their puzzle like nature rings very true. Perhaps software cannot be nothing more. But there is lots that _can_ be done with virtual bureaucracy. For example 3D modelling tools do enhance the human capability to rapidly design complex forms. Text processing tools beat hands down typewriter in several tasks. Etc.

Roritharr 6y ago

"automated virtual bureaucracy" Oh boy you hit the nail so hard on the head that my spine is still tingling. In the past 15 years of my career writing business applications, websites etc. I always had this nagging doubt that this is just me being a very active part in building moloch, adding to all the bureaucracy instead of reducing/simplifying it. Simply because wherever we built something to make things more efficient/easier, those gains are usually swallowed and overshadowed by the demand for more complexity that arises from latent potential that has been opened up. "we've made taxes easy!" "...could you also make tax evasion easy?" "yes, but it would be more complicated" "More complicated than before you were there?" This is just an made-up example but that's how it so often goes and it really makes it feel like I should have become a hermit and wrote indie-games while living on cupnoodles instead.

dustingetz 6y ago
yes the purpose of automation is scale, note scale is what enables big salaries

alisonkisk 6y ago

UI will always look simplistic because it has to communicate with users. But beneath that tiny surface is a huge depth of software, usually on the server. Think Google Now/Assistant, notifications, recommendation systems,and all the semi-automonous server management stuff like Kubernetes and EC2. Also non-consumer stuff like protein folding.