Then and Now: How Little Things Have Changed (2016) | Hacker News Reader

Then and Now: How Little Things Have Changed (2016)

kennethfriedman.org

> We are primarily using the same methods of input and the same applications for “productivity” as we did 20 years ago.

Longer than that. We've been using keyboards for over a century. The basic idea of a document is centuries older still. Maybe there's nothing really better.

For some reason the article focuses on the solved problems and then complains that they are still solved with the same good old solutions instead of being completely replaced by newer solutions that might be worse at the old problems but good at newer problems.

Keyboards have solved the problem of efficient text entry. PDFs have solved the problem of displaying documents exactly the same way on every computer. Written text has solved the problem of efficiently recording and passing on information thousands of years ago.

The new solutions solve the same but slightly different problems by trading off one aspect for another.

The touch screen is not the most efficient way of text entry and selecting 2D points but it's the most portable. But this comes at a cost. When you use a touchscreen your fingers obscure the screen and virtual keyboards take screen realestate away. This is an acceptable tradeoff when you're not constantly interacting with the computer.

The bandwidth of a human to a computer is very limited. Even very fast typist can only transmit two words per minute or 10 bytes per second. It's not worth it to waste 50% of you r already very limited productivity on something just because it's new.

There are a few improvements in UI and interaction. Something akin to the touchscreen already existed 10 years ago - it was the Wacom. On 2006 era laptops like ThinkPads, you had a Wacom behind a screen. Some of these laptops had also a touchscreen on top of the Wacom. So not even the touchscreen is really new.

But there have been a few genuine innovations. Personally, my favorite is the notebook - whether from Rstudio, Mathematica, Jupyer, etc. especially when you add dynamic widgets like D3 or maplets.

It allows not just literate programming, but quick prototyping and verification, by integrating visualizations. Sometimes you can even embed your datasets for replication.

I would love to see scientific publication move to notebook only - as it, I want to download the whole thing and do some tweaks to see if I can poke any obvious hole, or if it is genuinely interesting.

When I say the whole thing", I mean it. I do not want just the authors verbiage and some screenshots of their tables and graphs. I want their data, their formulas, their visualization, already implemented and ready for tweaking without imposing any extra software dependency upon me.

OP here. Definitely some improvements... but (at least in my eyes) incremental.

Notebooks are pretty sweet, but in the hands of very few. They're definitely growing in popularity, which is great.

But even the "notebook" concept is a very limited "low pass filter" version of the visions of Engelbart, AKay, TN, etc.

The video that AKay has been sharing around in the post than @dang linked to[0] is a great example of what's not-even-close to possible on today's systems.

[0]https://www.youtube.com/watch?v=AnrlSqtpOkw&feature=youtu.be...

(sorry on my first sentence I wanted to write "very few" - I gave the example of the Wacom to show that touchscreen were nothing new, and in fact quite incremental changes)

I agree the notebook concept is still quite limited. I just see it as a step in the right direction.

And great video BTW, thanks a lot for the link!

I don't see why this should necessarily be surprising or viewed as a bad thing. We've been using essentially the same controls to operate motor vehicles for over 100 years. Sometimes the market evolves these things to a point where they work as well as possible, or at least better than anything else we can envision.

Since the topics are related, maybe this is a good place to note that there are some excellent discussions with Alan Kay happening right now at https://news.ycombinator.com/item?id=15261691. Fine Saturday reading abounds.

Microsoft Word supports dynamic content. That's what Object Linking and Embedding is all about, and that dates back to the 1990s. Unfortunately, that only works within the Microsoft world.

What's changed since 1998? Media consumption. That works much better. You can watch just about anything, from live sports to porn, on any reasonably capable device, if you pay the bills. That covers the use case for about 80% of the population.

When you read something like Programming Pearls by Jon Bently, it's startling how similar the ye old problems were to today. For example, his article might mention, "So, we built a system that had to store geographical data by latitude and longitude, and we only had blah blah of memory....".

The machines have changed but, remarkably, the root problems are the same.

But how have little things changed?

This article sounds poorly researched to me - examples:

  * He doesn’t have any interest in learning to program, so he doesn’t attempt to create new things with computers.

AppleScript has been around for a while now - and since OS X you have a full UNIX system with terminal too. Johnny may just be a consumer, but Doris Developer has a mac too.

  * The desktop paradigm has been around in the 1984 Mac — it’s time to re-think interfaces at just about every level.

It's not like this hadn't been tried (GNOME 3, Ubuntu, Windows Vista/8/10, iPad "home"/"apps" screen, google's mobile-optmised interface guidelines). I would hazard a guess that for productive work, the keyboard/mouse/desktop interface is actually pretty good.

  * But there’s nothing inherent about the precision of a mouse that makes it better for content creation.

For programming, I've yet to see any input device that beats a keyboard. For tasks from photo/image editing where sometimes I want 1-pixel precision to games (mostly strategy) to 3D modelling, my experience is that a good mouse is miles better than any touch interface I've seen so far, precisely because of its precision.

  * [Word] allows people to create single document (that can’t connect to other documents).

I'm pretty sure "master documents" have been around for ages, templates even longer and you can certainly manage a bibliography database shared between different documents in recent versions.

  * The documents must be static, you can’t have any data-driven content

I will grant that it's not the easiest thing to use, but various forms of OLE and ODBC have been around since about the Windows 3.1 era. Embedding videos and sounds in powerpoint presentations is really common too.

  * It’s clear: people see the need for new interfaces, and lots of research is being done. Why isn’t any of it hitting the mainstream? I have no idea. Let’s find out, let’s make it happen.

I don't see the need for new interfaces just for the sake of it and a lot of the experiments I've seen over the past years haven't exactly been improvements. It's got to the point where there's a free program called "classic start menu" that gets you the older menu back on newer versions of Windows, or Linux Mint the USP of which is at least for me "like ubuntu but WITHOUT the new interface".

The change from toolbars to ribbons in office also "hit the mainstream" (I think on balance this was a positive change, but I miss the old Control-Shift-V for "paste without formatting".)

Deja Ubuntu still Kinda SUcks Again

I fry my AMD black edition 965 x4 and motherboard 64 bit.Was content staying with Ubuntu 15.10 since it allowed my obsolete AMD video card to work while subsequent editions didn't.

Relegated to my HP mini 1116NR with two MB ram I got the resident XP 32 bit running pretty zippy. But I wanted my Ubuntu back so I went and installed Ubuntu 17 and then 16 but neither of them made running my onboard Broadcom modem possible.

I try 15.10 and the modem works flawlessly at installation and thereafter.WTF? Why am I screwing with modems years after that was no longer a problem.

Fine. But now Ubuntu 32 bit is slogging along barely working and after digging it appears to be the if you plug in USB stuff IRQ/9-acpi goes nuts on the processor and virtual processor on the N270 Intel chip and takes up %100 of the cpu so nothing works. No apparent work around works shutting off acpi and others so I unplug the USB hub...

But still performance sucks. It turns out that COMPIZ is eating to much cpu as well. Support the new desktop and the crappy guess what you want search for applications and other crap menu shit thing that has all thosefilters and tie ins to services...which I ignore and install classic menu instead all the time anyway...

So then I install the Xubunt Desktop log out log into it and that goes flawlessly. My machine humble 32 bit mini is able to function fairly well...although consigned to obsolete browsers since my platform is only 32 bit it takes some tweaks to get Firefox and CHromium to perform adequately.

So obliged to head back to a small ram 32 bit machine on Ubuntu if you move past 15.10 you are screwed. And old bugs aren't fixed. And Modems start not to work all over again. And Compiz Desktop can't funtion without being a hog on the CPU and who really needs it anyway with that crap guess what you need search menu thing.

And systemd or not Ubuntu has gone the wrong way and as a desktop has no backward compatibility and past 15.10 it can't run a broadcom modem in 32 bit.

And deja vu all over again it doesn't work out of the box and one is jumping through hoops to make Ubuntu work because the latest versions are for people that have the latest hardware i.e. money and its not desktop software but some sorta everything pseudo phone software that is not even forward compatible with a damn Broadcom modem.

So things haven't changed when things that once worked don't anymore.

Reply on news.ycombinator.com