Slashdot Mirror
Computer Folklore, Circa 1984
savetz writes "The full text of the classic 1984 computer book Digital Deli, The Comprehensive, User-Lovable Menu of Computer Lore, Culture, Lifestyles and Fancy, is now on the Web. (Autstralian mirror) A wonderful look at technology culture in the golden age of the microcomputer. 20 other old computer books are at the site, too."
De Re Atari.
Holy shit.
That was one of the first "serious" computer books I ever got -- I won a copy as a door prize at an Atari user's group meeting when I was about 12 years old. By the time I was done figuring out what all that crap in the back of Compute's Gazette was doing, my copy of DRA was so dog eared and broken spined that it couldn't sit flat on my desk.
Good memories. Glad to see it's still around somewhere.
--saint
I love how many books are becoming available in their entirety online. It gives an open-source advocate a warm fuzzy feeling.
http://textfiles.com/ is another fantastic, wonderful resource and window into computer-ages long gone. Check out the top 100 - especially the Captain Midnight story. My kids will be getting this read to them before bedtime some day.
The Secret Guide to Computers by Russ Walter was my personal old computer book favorite, I remember checking it out of the library, it had tons of great info about all different kinds of computers. Great writing sytle, kept your attention and was funny! I recall he had his home phone number in it too...
I wonder if it's still published... off to Google!
... if that's your best, your best won't do... - Twisted Sister
I have an old (~1994?) Introduction to Networking (QUE) text in which it says TCP/IP is a standard that will more or less fade because the DOD insists that future protocols comply with GOSIP (Government OSI Profile). Nice call QUE!
The book contains a short piece by Bill Gates (here: http://www.atariarchives.org/deli/soft.php) and reproduced below. It's an interesting read because I still hear him talk about similar themes even today.
------------------
Today's software is too hard. Usually designed to work well for any and all potential buyers, a few years and hundreds of hours of interaction later a software package will still interface with you exactly as it did at the time of purchase. Your special use may make some uncommon program command the one most often employed, but you'll have to punch any number of extra keys every time you invoke it. Today's software fails to remold itself to express a history of use, and this can lead to incredible inefficiency.
There are programs that allow the advanced user to adjust default values, which are those responses the programmer decided would be most typical for users of a specific application when the software was first booted up. There are also programs that can store a series of often invoked keystrokes and can tell the machine to take the sequence you've named and perform it again. These keyboard macros, the most trivial form of softer software, force you to go through a special set of operations to enter and record changes to the program.
Why shouldn't software automatically adapt to your needs, e.g., learn from experience to change the interpretation of a command, when this is done on a human level all the time? In-human-to-human communication, we adapt our terminology and our method of understanding to our previous history of interaction with each individual. There's no reason computer software should not be as flexible.
"Softer software" is the term I invented to avoid using the poorly understood term "artificial intelligence." In fact, it is a form of artificial intelligence, though not like speech recognition or the expert data base systems that are based on specific algorithms and do not really learn dynamically. Softer software is capable of getting better and better because it has advanced pattern recognition capabilities and can change its performance accordingly.
In general, making software softer requires storing information about a user's history of program commands and analyzing its patterns. This is a form of learning, since the software can build expectations of what the user may do later. Individual characteristics of users, what they're good at and what they're not good at, can be used to establish a reasonably unique dialogue with the computer.
A data management program, for example, could recognize that you always query its files by employee name rather than by an individual's address or hair color. Taking advantage of this pattern and predicting what will be your most common operations on data, the program could customize its query file structure to put information within easier reach. Or maybe it could learn to be forgiving of your most common keyboard mistakes by ignoring misspellings.
Software softness becomes very difficult when recognizing semantics rather than specific operations is required. Say you go into a document, move the mouse to bring the cursor to a certain position and make a word boldface, then go to another position and do it again. Instead of storing up the exact positions where this takes place and trying to match them to later entries pixel by pixel, you may want your software to draw the general conclusion that you boldface the first word in a paragraph and to position the cursor appropriately. Matching things, recording and playing them back at the semantic level: this is the hard part of softening software.
It is possible to say that we have certain types of softness built into software today and that over time we will see a clear progression as programs record a greater number of user events, recognizing more general patterns and building up the dialogue throughout the computer's history. Truly softer software is still some years away, but we are on an evolutionary path where at som
He sure was dead on about the future. Quote Below:
They call us pirates and worse. They lock up their programs behind hardware and software schemes. They set the minions of the law upon us. And still we flourish by our wiles.
Ahoy, ye microlubbers: to pirate a program is not to steal, but to liberate knowledge. We don't take money or goods from anyone; we merely free up information. Most of us don't profit from our buccaneering activities; instead, we share the wealth with our fellow computer users.
The software moguls have only themselves to blame for our cracking open the bars to their programs. If they didn't charge a king's ransom for disks that cost a pittance to duplicate, there would be little incentive for us to practice our skills. There would be no need for them to protect their programs if software were no more expensive than what you and I can afford to pay.
We are no longer in the Dark Ages of personal software, when so few people used computers that program development costs had to be defrayed by high unit prices. Now so many microcomputers are in use that a program should cost no more than a lightweight paperback novel. Instead, we are paying illuminated manuscript prices.
Maybe someday the software publishers will understand how they're killing off the golden goose. But until that time, be warned: there will be many a pirate's flag on the software horizon.
JOLLY ROGER
If you wanna get rich, you know that payback is a bitch