this chart was made by a reddit user in 2018:
Each day, he opened this document, and entered his word-count into an excel spreadsheet. It's my favourite illustration of how 1) software is useful and 2) software is shit. I promise I am not a whiney, or negative person. I came here to say that this is a amazing. I want so much more of this in the world. so bear with me,


consider this image:
Think about creating this image - which tool would you use?
Sketch? Photoshop? MSpaint? It's pretty doable. what if you wanted to make the lettuce thicker? Chances are, after resizing it, you'd have to move each of the other layers up. I know there are some sketch plugins, and other exceptions - but this is primarily how we are doing design, in 2019. This seems like a basic thing. Suppose you're very clever, and made the hamburger in flexbox, or CSS grid.
Consider this:
(Happy 10th birthday to this Micorosft vision video)
not this:
it's this:
If you think about it, this is a tricky thing to do. Each item is placed depending on the others. ...but then what it depends-upon changes.
To do this example, i just hard-coded it with absolute-positioning. It took me a good while.

I'm sure, 10 years ago, the people at microsoft did exactly the same thing.
Here's a clip from the Oculus Quest demo-reel:
we are still doing this, the faking of layout.
showing a list is arguably a central point of information technology.
and somehow we don't do this very well.


here's a google all-hands meeting, in 1999:
imagine you're pushing pixels at a startup.

Your mockups are beautiful,
but the results are always fidgety, and compromised.
'One more iteration' they always say, but then you're old.

This has surprised me, how we have not gotten better, over the decades.

you gotta feel sometimes like something is wrong -

indeed, you hear stories about army-generals trying to manage software projects and giving-up.
Developers are stuck trying to fake things, with increasingly-creative fakery.

There's probably no industry with a more dense history of loopy, compromising tricks-on-tricks-on-tricks.
perhaps the most-designed website ever:
when is it officially broken?


Computer latency, 1977-2017 - figures by Dan Luu
18 months is a inscrutable about of time. that's the trick. The truth is that computers got rapidly faster in the 90's and are now measurably getting slower. We don't have a cute name for it, because it's not a fun thing to TED-talk about. Computers are getting slower, and our software is more polemic and competitive.
The best time in history for latency was 1983. Computers had a spec, and you could just engineer under it. Early videogames had essentially no lag, loading, buffering, or time-outs.
My favourite description of software-latency is by Bruce Dawson, where he densely analyzes typing latency as he's writing an email.
It's amazing and I recommend reading it.
There's no finger-pointing, or conclusion.
People have been working hard, for a lot of years on software.
The more software we write, the slower it gets.
A recent analysis showed that most clickbait-articles are larger than the all the major works of russian literature.
Red-dead redemption 2 is larger than two uncompressed wikipedias.


big data is a solved problem. the real problem is doing your taxes. you know,
moving your bank-information to a spreadsheet,

then to the printer....

then paying your babysitter

then playing a song at a wedding

the problem is not lack of engineering. the problem is in our solutions:
  • make a start-up for church sound-systems.

  • i was neck-deep into the semantic web - ' web 3.0' as it was called.
    When it failed it hurt and made sense. It was pretty pedantic, jargony.
    What I didn't see happening, is the loss of web 2.0, the light-weight sharing of content between pages.
    This one has really bothered me.
    mozilla's ubiquity extension.
    we were given a perfect information processor
    a real-time network
    and machines that look sexier than any sci-fi concept.
    and we use it to communicate using screenshots

    we may never be able to interpret plaintext in our lifetimes
    we are creating unmanagable trash, with our perfect computers.
    it's not better than paper.
    it's getting worse.
    it's not about blame
    There are plenty of phenomena, like the tragedy of the commons, which don't require detailed finger-pointing, they just say 'it doesn't end up working', and we nod
    Software may not end up working.
    Has your computer gotten better in 10 years?
    Are you in more control, or less?
    My computer's slower. The keyboard's better. The monitor's flat.
    the software is a bit worse.
    - like goth-music, it may be the case that software development was new,
    - people rolled it forward -
    and then it clearly ran into it's natural end.
    Then people just stopped doing it.