Hypercard – Way Too Early

As an example of how far from the 'card' metaphor you could go using Hypercard: Myst was built in it.

Watching this demo of what Hypercard could do in 1987, it seems like there are a lot of things that would be great if they were translated into a modern context right now.

http://www.youtube.com/watch?v=EMFscTOazS0

Hypercard was never too early. I remember some 'dot com' guys building a 'business management suite' in hypercard, maid fortune which evaporated after two years. Hypercard was a mere joke, like applescript.

I've thought a lot about HyperCard.

In 1994, when I was in fifth grade, I used HyperCard to make a demo of a program that would allow you to fax a grocery order to a store using a modem after you checked off the items you wanted to buy. Then the store would deliver your order to you. I didn't actually know how to make use of hardware devices like modems, but it was possible to make the user interface work reasonably well on a Macintosh LC II. I was of course bummed that there was no color. (HyperCard was black and white only unless you used proprietary extensions, which could sometimes make things run very slowly.)

The author of this post seems to think that HyperCard was "before its time," but it wasn't. It defined its time, and ours. It was a huge part of what made the Mac so appealing to so many people--the ability to sculpt new technology in a custom manner. There was no comparably easy-to-use tool on the PC--Visual Basic didn't come close--and there never would be. The result was that all kinds of Mac owners developed their own stacks that solved exactly whatever problem they thought needed solving; PC owners had to hope that a software company would make something for them.

What's sad is that now our computers are so much faster than the LC II I used in fifth grade, but even if they're a hundred times faster, they're certainly not a hundred times easier to develop for. Most fifth graders, even smart ones, probably couldn't write an app in Objective-C that does today the same things a HyperCard stack did in 1994. HTML5 is the next best option, but it's still not as easy.

I worked on a prototype of a modern-era web-based HyperCard with a friend, but we stopped working on it because there was no VC interest and it wasn't clear if there would be a good business model. It's a shame, because HyperCard is one of the things that most excited me about computers when I was a kid. I still miss it.

I think what HyperCard got right was easy structured form based editing. You could insert fields into cards and use those as a template. When inputting a field you actually edited a property of an object directly. I can't remember if fields allowed rich text editing nor programmatic access to it, so replacing Word was probably not possible. Creating complex forms and interfaces for data was certainly easier than with web browser. You could start your own solution by simply copying a suitable card. One knowledgeable user could create a powerful app tailored to company's needs and others could build on that.

In the browser the presentation, runtime and persistence layers (otherwise known as DOM, JavaScript and the web/LocalStorage) are completely separate. So, you need loads of code to bootstrap your particular version of editable web that HyperCard gave you for free, without any save buttons, serialization protocols or web server and database installs. This seems so basic thing, yet it's so hard to get there in the modern browser. Forms have in their own, peculiar data structure completely unusable for all but simple cases without a clever encoding system and JS data binding. Rich text editing is so unreliable across implementations that you need to get a library like CodeMirror to turn textarea into transformation pipeline simulating the process of input. Saddest part is that the priorities of developers are elsewhere, in building the browser into best consuming platform and a fastest tool to send little snippets of text to social network silos.

One of these days I might turn LigthTable into the past version of HyperCard, with the future build in. Since it comes with a browser, server and the JVM there should be more than enough code to replicate functionality of few megabytes of 68000 assembler ;)