Slashdot Mirror
Skunkworks At Apple -- The Graphing Calculator Story
avitzur writes with a link to the story behind the Macintosh Graphing Calculator. An excerpt from this strange account: "It's midnight. I've been working sixteen hours a day, seven days a week. I'm not being paid. In fact, my project was canceled six months ago, so I'm evading security, sneaking into Apple Computer's main offices in the heart of Silicon Valley, doing clandestine volunteer work for an eight-billion-dollar corporation."
Wow. This story really really amazed me. It made me think of dedication. I can think of people *cough* EA employees *cough* that work those long hours, and that finish a project, but that's because they're forced to... I really wonder if this type of dedication for just the love of the work is existant anymore... I, for one, wish it was a lot more frequent.
- dshaw
Recently there have been a number of slashdot postings related to the conditions of working for EA (can't recall the exact URL, but summary best described as "slave-labour like"). I wonder what those folks think of this level of dedication?
On another note, it was a nice holiday feel-good read for the techno-geek developer. Also inspires me to finish the damn project that I am on right now so that I can "be home for Christmas".
Happy Holidays!
just a web application developer and instructor in Toronto, ON Canada
Now if they could only find someone that'd work night and day to invent the 2-button mouse they'd have it made.
Actually there is only one person preventing a multibutton mouse, unfortunately no one outranks him. He won't even allow a build-to-order option when you are ordering online.
You can't legally volunteer to help a for-profit corporation. And for IT staff, there is a minimum amount you have to pay them (well above minimum wage; don't worry).
-russ
p.s. R0ML says that this is why he couldn't get a carrier-grade accounting system turned into open source.
Don't piss off The Angry Economist
The story he describes occurred in the early 1990's, when Apple was beginning to hit its skids. Projects would be raised with a flurry of energy, then cancelled, and there was a general sense of chaos. That was either in the latter part of the John Sculley era or the beginnings of the Michael Spindler, which were NOT good years (eg., the failed Newton, the failed Copland system, and merger talks with Sun Microsystems, etc.) Scully, Spindler, and Amelio were all shoved out of their CEO positions due to unsatisfactory performance.
The problem with fairy tale workplaces are exactly that: They are fairy tales that don't last long in reality.
>I hope we don't hear from this person's significant other soon...
I was dating a high school math teacher at the time, but, unsurprisingly, the relationship did not survive the events of the story.
Does anyone remember the demo Ron gave at the World Wide Developer's Conference? Was it May 1993...?
Anyway, I remember it was supposed to be a lecture about pen computing, and Apple had Ron come out and show the equation solving interface of the proto-graphing calculator. He threw a bunch o' X and Ys on the screen with some sins and coss for good measure. "Now if you want to solve for X"... and he tapped an X, dragged it to one side of the equals sign, and the equation solved itself.
We were floored. There was this deep silence for a couple of millisenconds and then everyone broke out in thunderous applause. He did more tricks with the equation interface and people hooted and hollered. It was a geek wet dream. After he finished he got a standing ovation and there was a long line of people who wanted to shake his hand.
Good times.
I have this theory that programmers who write software should have to do in person tech support for that demographic for at least a year or so. It really opens your eyes as to what users are actually doing, why they're doing it (if you can get them to be frank with you), what they like, what they don't, what works, and what doesn't.
It makes some decisions about how to do things a whole lot easier...
Everything I need to know I learned by killing smart people and eating their brains.
The classic silicon valley hacker/enthusiast vs. big corporate culture. It says alot (in a positive note) on the type of people who worked there and helped these guys along.
I've worked in a big company like Apple in the past and with the right people this just shows how far someone can really go in the most ideal situation. (not really needing a job in the short term)
Good Job Ron!
Absolutely -- if Steve had been working there at the time, he would have been designing the calculator himself!
What I'm listening to now on Pandora...
This is - beyond a doubt - the most amazing piece of software I have ever seen. I never knew this gem was sitting quietly on my hard drive.
:)
:)
At first, I was unimpressed. However, as soon as I saw it animate I was blown away. Of course, when I saw the plane intercept of a 3D function animated, I was visibly giddy.
I so wish I had this while in my vector calculus course. In fact, I think I might stop by former professor's office when school is back in session and show him.
As soon as your site recovers from this merciless slashdotting, I think I might pick up version 3.
And again, wow.
heh. If there are any PovRay developers reading this, send me an e-mail. I'd like to discuss this. It's on the big list of features for future releases. http://www.PacificT.com/TheList.html
Ok. Done.
I have this theory that programmers who write software should have to do in person tech support for that demographic for at least a year or so.
Years ago when we developed a replacement CRM application for a large telco ISP, we did something unheard of - we integrated the customer service reps into the development process. At first we shadowed them for days to get a feel for how they use the existing application, and interviewed them to see what they liked and disliked. Then we invited at least one rep to every design meeting. During development they were constantly reviewing the work, making sure it was perfect. They almost cried they were so happy.
As an aside: their number one complaint was when they were doing data entry on the very long web form, they constantly had to take their hand off the keyboard, find the cursor, position it over the scroll bar, scroll the page down, then position the cursor over the text field, and resume typing. Tabbing took care of some text field focusing, but wasn't intuitive and predictable enough even when combined with javascript. We broke the data entry into multiple pages with simple navigation. I really miss the old days of character-based terminal applications (so do a lot of end users).
_______
2B1ASK1
Just setting straight some of your inaccuracies
Internal numbers are accessible via the last 5 digits on an internal phone, but not all (or even most?) start with 2. Or maybe you're trying to get your friend in trouble?
If you tailgated in years ago, that may be true. These days, good luck tailgating if you're not known by the person you're following, even if you have a valid badge. Also, while all buildings have a double set of doors (access to the lobby from outside, and access to the inside from the lobby), the outside doors (into the lobby only) are unlocked during business hours. Good luck distracting the secretary (or more likely, secretaries). You'll need more than one accomplice to do that for you (they're really not busy enough for you to bank on random traffic, and even when they are busy they have a clear view of the doors and will stop you from tailgating), at which point you could just get a valid visitor's pass instead.
Cool stuff generally is not just "lying around", unless you want posters and such off of the wall. Everything else is in a locked lab or occupied offices, and in the latter case anything you could easily get away with is personal property. Do you feel good about stealing from people? (ignoring that you're suggesting stealing from a company)
The free sodas are still there.
If that's your goal, you need to have good inside sources. Entertainment items vary from building to building and floor to floor. If your heart is set on Donkey Kong, you'll be disappointed to find only Street Fighter 2 if you didn't do your research (and that's not publicly available, or even easily internally available aside from visiting every building).
Which are not sitting out in plain view, if available at all in that building. If it's software available to all internal employees (for example, connection manager software to connect to the VPN from home), you have to get it from the receptionist. If it's for a product group, it's either locked up in the lab or in the group admin's office (or more likely, not available in CD form, but on an internal share you'll not have access to). Either way, don't expect to find piles of booty just laying around.
I've never seen that, but most buildings are on a timer to shut off lights (not power) after a certain time of night. There are internal overrides if you're still working.
There's a good chance your car would've been towed if you weren't showing a valid parking pass or visitor's parking pass. And if you drove back across the lake to get to I5, you wasted a whole lot of time sitting in traffic on the floating bridges (I90, SR520). If Canada is the goal, better to take I405 up around the lake and meet I5 there.
There will always be an Apple Computer.
"Flyin' in just a sweet place,
Never been known to fail..."
Folks, if you've been paying attention, you will notice that an actual Microsoft employee has been "social engineered" into revealing information about the security in the buildings. This guy says "good luck" plenty of times, without realizing that these amazing ninja-turtle secretaries and others would cough up info with less trouble than he has.
Fascism trolls keeping me up every night. When I starts a preachin', he HITS ME WITH HIS REICH!
This story is guaranteed to be very boring for 99% of readers, but it's probably my only chance to tell it where anybody might be remotely interested.
Back in the 80s I was part of an IT group in a manufacturing dept at Tektronix. Our software involved inventory control, tracking batches of work through assembly steps, that sort of thing. One of the computer operators asked if I could help him solve a problem for the stockroom people. Their job was to hand out parts to assembly workers, receive and store the finished subassemblies and hand them out for additional steps until they left the area as finished goods.
All movement of material was tracked by a giant MRP system on an IBM mainframe in another building. The IBM machine generated stacks of PUNCH CARDS which were delivered to our computer room and loaded into our VAX 11/750. As the stockroom people handed out and received material, they had to manually keep track of what they did, noting shortages and errors. Then they entered the information into the 750, which wrote it nightly to a tape that was hand-carried back to the building where the IBM system was.
The stockroom data entry program was very cumbersome to use. It simply did a one-way scroll through the entire inventory -- thousands and thousands of parts and subassemblies -- and allowed the user enter a code on the few items that mattered. To get to an item near the bottom, the clerks had to hit the Page key dozens of times and wait for the slow page refresh in between. Sometimes they would hold the Page key down for a while and go away until it caught up. If they overshot they had to start over because there was no Back function. The stockroom people spent most of their time doing data entry and were consistently several weeks behind, which forced them to come up with various manual ways of keeping track of things. This affected their ability to hand out parts and was starting to have an impact on manufacturing deadlines, and ultimately profits.
In spite of the importance of the situation, the stockroom was low on the IT priority list. So we had a couple clandestine meetings in which the staff told me how the business end of the system worked and the computer operator explained the behind the scenes parts. Working a couple hours a day on the sly for about 2 weeks, I came up with a new data structure and an editor that let the users search for what they wanted and produced various on-screen reports. I also changed the loading procedures to use a tape instead of the stupid cards, and my operator friend persuaded an IBM sysop to bypass the change control process and generate a tape for us instead of cards.
When the users were satisfied with the way everything worked, we put it into production one afternoon as the swing shift person came on duty. In that one shift she cleaned up their entire 3-week backlog of data entry. When the morning people arrived they were speechless. With the extra time they now had, they set about reorganizing their operation and making improvements that they had wanted to do for months.
It was amazing to see what this change did for the morale of these people. Their jobs had been absolutely miserable when they had to work with the old system. They were so happy they brought me a great big apple pie, and were almost in tears giving it to me. Best award I ever got.
Apple squandered a great opportunity in the 90's. Macs were much faster than many Sun workstations with the kind of work we did (computational fluid dynamics), much cheaper and ran a broader selection of applications. Despite this Apple knew nothing about the scientific market. I remember going to a seminar at MacWorld Boston in 1996 on scientific uses of the Macintosh. None of the presenters talked about how a PowerMac 7500 with a 3rd party 604 accelerator smoked a Sparc 20 for about 33 percent of the price. Instead they talked about how they could use a Mac to model the behavior of a lobster. I felt as if I was in crazy world, here was Apple with this insanely great line of CPUs and they basically ignored a market that would have gone for it lock stock and barrel.
Things have gotten better since then and I have been pleased to see that Apple is targeting bioinformatics applications with the Xserve, but they're going to have a lot of work ahead of them to keep up with Linux's inroads into the market.
cheap labor conservatives - they want to keep you hungry enough to be thankful for minimum wage.
Ron's story points to the reason he and Greg felt compelled to do it at Apple. It was the best environment in the world to accomplish the Graphing Calculator. The resources were there. There was a top-notch research library there at the time. Many of the people who are determining the course of computing at Apple and in other places today were either interning there or working there after graduating from college or had been there for a while. He mentioned the QA people. They were and are true advocates of the users.
I don't know about other places, but working at Apple was--and, I imagine, still is--like playing for the Yankees. The expectations of the fans was everything. The penchant for doing things the right way permeated every nook and cranny of the place (except upper management until Steve Jobs returned). Little bits of fit and finish that weren't even noticeable until you had them pointed out to you (or they were missing from a product) were all-important. The best projects assembled teams of people who were involved from the design phase to product release; the good managers made sure that everyone stayed on the same page.
It is the people that make Apple so special. They care about the customers. They care about the products. They care about each other, for the most part. Reading Ron Avitzur's reiteration of what passes for high praise at Apple ("this doesn't suck"), brought a tear to my eye. Another saying--at least among my group--was "we do good work." I have my own business now, and the things I learned at Apple guide everything I do.
Wonderful article, Slashdot. Thanks.
I was struggling through algebra I not long after this program came out (1995). I just wasn't "getting it". I know the phrase is cliched now, but this program was just so *intuitive* that after a few days of fiddling I understood almost all the math I'd ever take right up to 1st semester calculus on a conceptual level.
For me, at least, seeing things in motion (that nifty little value slider) made the concepts just click. Once they were there, the actual mathematical manipulation was much easier, because I was able to visualize "they way this should work out". My teachers were trying to show it on a static chalkboard, and it just wasn't getting through.
I just got my BS in Physics, and without Graphing Calculator, I doubt I'd be where I am today. To the author, if he reads this:
Thank You.
On a side note, something similar and free already exists for windows:
You can download Powercalc.exe from Microsoft's XP PowerToy page.
While designing Concorde, some engineers started working on their own project without telling anyone, diverting a small part of the huge ressources needed for concorde to this.
;)
At the end of the concorde project, managers discovered with great surprise they also had almost all the plans of a working regular subsonic jet : Airbus was born.
Ironically, the unofficial project actually succeeded far better than the official one.
It doesn't happen only with software companies
Kirinyaga
I'm not sure, but this may be why Google is so great, their culture is hacker-friendly, being formed by two graduate students.
k .html
Also I believe the original creator of AutoCAD, John Walker, wrote something about creating democratic companies where everyone is equal below the one visionary.
It's called the AutoCAD File:
http://www.fourmilab.ch/nav/topics/autodes
Bypass Compulsory Web Registration -- http://bugmenot.com/
That's where I have an issue with how copyright is enforced. Copyright was written in such a way so that after a given period of time, the work passes into the public domain. DRM effectively prevents that. In my opinion, if you want to require a licence/contract to view your work, then your work should no longer be protected by copyright law. In effect, you should have to choose between contract and copyright protection for your product.
I have long considered releasing GC under an open source license. While I have total sympathy and support for the open source movement and philosophy, my analysis is a pragmatic one. My goal is to best serve my users, and adopt a strategy to best accomplish that. (I think the events of the story give proof to that.)
Remember that my software's users are primarily high school students or younger.
An open source release would cause the existing revenue stream to vanish, making it impossible to continue to support existing customers or maintain and develop the code base. Pacific Tech has provided free support for its customers and would like to continue to do so, and continue to maintain the product on Mac OS and on Windows. The reason GC is useful is the ease-of-learning and ease-of-use of its user interface. I do not know of a product which demonstrates the open source community's ability to produce excellent user interfaces.
In a high school classroom where any time spent on software is time taken away from teaching, usability is the most important feature of our product. In these respects, open sourcing the code could prove to be a large disservice to our customers.
This is educational software for high school users. There are remarkably few people in high schools, either students or teachers, with both the skills and the time to contribute to open source development. This removes one of the major motivations for open source development - the "I need this tool for myself" reason for working on something.
Furthermore, the reason this software is so useful to schools is not how powerful it is or how many features it has - it is completely unlike the classical monolithic mathematical applications of yore. It is useful due to the restraint in choosing a minimalist feature set and interface to address teachers' and students' needs with elegance. I fear that as an open source project, the incentive structures would lead down the slippery slope of creeping featuritis, which, while it might create something cool for hackers, will do little to help children learn math and like math.
I would like to find a way to have the best of both worlds. Genuinely open and free software, for all that that implies, and the ability to continue to create great software for people that are not programmers, for people that hate math, and know nothing about computers, and to have it installed at the factory, so that it actually reaches the people that need it where it can do some good.
Our then-novel ideas now turn up not just in math software, but in applications as well as operating systems. User interfaces incorporate live animated feedback instead of dotted outlines, direct interaction instead of dialogs or configuration files, a functional rather than demanding starting point for new users, context-driven help, and in the best cases, minimal preference settings. We didn't patent the ideas or the algorithms; rather, we wrote about our goals and methods, and encouraged people to take the ideas and run with them.
So if the complaint is just that the source code isn't free for anyone to copy, rebuild, and redistribute, then give us a model for doing so. It needs to be a model where we can cover the cost of ongoing development by professionals; the calculator has evolved in the past 10 years, as have operating systems. Paid support isn't a good answer, as that would reward us for making crummy rather than excellent software; we want users to feel empowered, not dependent. And the idea that students or schools could or would pay for support contracts is silly.
Ron has never turned down a reasonable licensing request. Getting students to learn and enjoy math is the goal. But letting other programmers recompile our code isn't interesting, nor would that really move it very far on future platforms. Better that developers learn from our interface designs, deduce our algorithms (or just ask us), then build better software on the next generation of computer platforms.
If you are bothered that you can't recompile our ten year old application yourself to fix a bug, then you really aren't in the target audience we are aiming to reach, nor are you among the people who will deliver the next leap forward in software design.