Slashdot Mirror
The Newton O.S. Creeps Toward New Hardware
GraWil writes "As previously reported, the Apple Newton refuses to die! The Worldwide Newton Conference 2004 has wrapped up (photos) and, thanks to Paul Guyot, there is real hope for an emulator. His talk, titled 'Newton never dies, It only gets new hardware,' describes and shows the Einstein Emulator, that will eventually allow the Newton OS to be built and run on top of Unix. Will your next Linux PDA boot Newton OS next year?"
Not your fault, they just changed the title in the last few seconds. It was originally "The Apple Newton Gets New Hardware".
Karma: Segmentation fault (tried to dereference a null post)
The Newton used a 16-25 MHz or so ARM, and even then it lagged quite a bit. The final models (before Steve killed it) had 166 MHz or so CPUs. The Palm has a 16 MHz 68000, so there's no chance there. On the other hand, modern PDAs (PocketPC, Palm ARM, Zaurus) use 200+ MHz ARM CPUs, so they ought to run the Newton OS in an emulator environment with no trouble at all. The important part is the total lack of need for CPU emulation.
#naabhaprzrag, #sverubfr-000, #agi-fcbafberq, negvpyr[pynff*=' negvpyr-ary-'] { qvfcynl: abar !vzcbegnag; }
http://wwnc.newtontalk.net/program/paulguyot/slide s-paulguyot.pdf
Turns out to be quite the interesting talk.
more from the conference:
http://wwnc.newtontalk.net/program/
You have to catch up on your Apple folklore.
The reason, wich is widely regarded as truth, that Jobs killed the newton is pure retaliation against Scully.
John Scully invented the concept and drove the outcome as the Newton shortly after he had fired Jobs in their power feud of mid 80s. Scully had killed the Lisa and Jobs took over the Mac not to be empty-handed. When Jobs was back at the helm of Apple, he was just pleased at destroying the Newton rather than building on it. To this day, Jobs keep dismissing PDAs altogether while telling everyone that phones will inherit the futur. What does he do next? A frickin' music player.
Jobs has done a lot of good stuff for Apple since his come-back. But the Newton murder wasn't one of them. Marry Newton OS and the iPod and then you start having something interesting. But ego makes this product impossible. Or highly improbable.
NO - Einstien is NOT open source, and most likely will not be open source at any point. There are legal issues regarding licensing of the OS, the HWR input, etc. that the developers need to work out before they can truly finish a product. As such they can't realease the code, what would Microsoft do if you released all of winXP source code on the net? Probably a similar thing to what Apple would (and may still do) to these guys.
Also - they've seen other open source developments for the newton go south - too many hands, not enough direction. They intend to keep things to a core of dedicated developers, to decide a direction to take things and to FOCUS on a goal.
Soups are just relational databases. Their relational aspect is what made them so useful on the Newton. If you stored an entry in your contacts it wouldn't end up steveha.vcf or something of that sort, instead the data would be added atomically to the Names soup. Later when you go to send an e-mail or fax to the contact (yourself for the sake of explanation) you would simply pick the steveha entry in your names to send it to and all of the appropriate information would be filled in because it would all be related in the database to the steveha entry. A note you wrote would be associated with its creation date and if you decided to file it in a group letting you not worry where it was and what it was called.
This system is incredibly powerful because all sorts of data ends up linked to other sorts of data. It is possible to find all of the e-mails that have been sent to you by a particular person or a bit of text stored in a note you got passed by someone. The Newton through its soups had content searches far before things like Sherlock or Spotlight.
I'm a loner Dottie, a Rebel.
Well if you dropped the budget requirement you would seem to be a perfect candidate for an OQO ultra compact PC. It's 4.9*3.4 inches, has a 800*480 display, has a full keyboard, thumb mouse, pen input, scroll wheel, etc. It has Firewire, Bluetooth, 802.11b, and USB built in. It lasts aprox 3 hours on battery. It has a 20GB HDD for storage. The biggest problem for you would be the price, I believe the first generation are around $1800.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
This is out of hand. Newton is 10 year old hardware that has an adamant user base that consistantly reaches over backwards to keep alive. Yet this hardware device is one that Jobs is staunchly against and has consistantly given the middle finger to.
That is just untrue. Steve has said that he could have saved Newton, but that he didn't have the management talent to do it. I believe that was just part of the story. The Newton group was working on StrongArm based products before things were killed. The StrongArm was a part of DEC that was acquired by Intel (When they picked up the Alpha technology & Engineers), at that point in time, Intel wasn't really sure what they wanted to do with it. It would have been insane for Apple to spend time rebuilding a business when they didn't know if it's major supplier was going to keep manufacturing. I was the last person hired into the Newton team.
So at this point is it "economically unfeasable" for Apple to Open Source the entire Newton OS and not just the toolkit???
Well, let's suppose for a moment that you were an executive at Apple: how much budget would you allocate to 1) finding the code, 2) determining whether it embodies any patents that Apple licenses from other parties, 3) seeing if it builds, 4) documenting it?
Apple doesn't use open source as a dumping ground. Darwin is live code. It's maintained, because Apple is using it today.
The thing is, it's just not as simple to give code away as many people think it is. Certainly not for a large, publicly-traded corporation.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
NewtonScript has a Pascalish interface, an unusual Prototype-style OOP model, first-class functions, closures, and Lisp-like s ymbols.
:= func (n) func(i) n := n + i;
In NewtonScript, objects are dictionaries which hash arbitrary things, each keyed by a Lisp-like symbol. The symbols are the slot names in the object. Functions stored in the objects, when called as methods on the object, automatically have access to a variable called this which refers to the object itself, and their scope automatically includes the object. A particular symbol, _proto, is used to key to an object called the "prototype" or "proto" (or nil if there is none), and another symbol, _parent is used to key to a prototype called the "parent" (or nil if there is none).
When a method refers to a slot in the object, here's how lookup is done. First, we look in the object. If it's not there, we look in its proto. If it's not there, we look in the proto's proto, and so on up the "proto chain". But it doesn't stop there. If we haven't found it yet, we look in the parent. If not there, then the parent's proto, then the parent's proto's proto, and so on. Then the parent's parent, the parent's parent's proto, the parent's parent's proto's proto, and so forth. So the lookup path resembles a comb.
Yes, bizarre. Why two ancestry pointers (proto and parent)? In NewtonScript, if you set a value in an object, it's set in the object: if the value was set in any protos, it's not changed there. But if an object has a parent, and you set a value in the object, the value in the parent is set as well (if it exists in the parent). Thus proto inheritance allows for polymorphism and sharing of defaults which can be overridden without hurting the ancestor; but parent inheritance allows for sharing of variables. Clever, but convoluted. In reality Newtons rarely used parent inheritance (widgets had parents, but that was somewhat of a different thing).
Because methods are first-class objects, they can be created at any time and can have closures, exactly like in Lisp. Thus Paul Graham's Accumulator Generator in Lisp is written as
(defun foo (n) (lambda (i) (incf n i)))
And in scheme it is
(define (foo n) (lambda (i) (set! n (+ n i)) n))
But in NewtonScript it's particularly pretty -- indeed it's the prettiest thing on his web page (I know: I provided it!)...
foo
Functions (which is all a method is) are all anonymous in NewtonScript. They're created with the func declaration which looks like func (_args_) BLOCK where BLOCK is either _statement_; or begin _statement_;* end The last statement in the function's block returns the function's value (unless there's a premature return statement).
Thus the above code says "set foo to a function taking an argument n. That function returns another function which takes an argument i, adds it into n, and returns the current value of n".
Pretty indeed. Essentially identical to the lisp code.
Actually, the HWR in OS X *IS* the very same Rosetta HWR engine that was in the MP 2100 (The printed/mixed engine). It even has the same easter eggs as the Newton does. Just write "Rosetta! Rosetta! Rosetta!" and it will recognize it as "Rosetta! Rosetta! Hey that's me!"
Unlike the cursive recognizer, which was developed in Russia, the Rosetta engine was written and developed in-house by Apple. If you do a search, I think that you can still find the ACM papers written by the guys who developed the engine. It's an interesting mix of Neural Nets, traditional HWR, and dictionary based guessing of the words.
...si hoc legere nimium eruditionis habes...