Legends of IT: Why Newton failed and whether Russian recognition technology is to blame
There is in the history of " Paragraph " - the first startup from Russia who conquered the world - one paradox.
The global success of the Russian company brought an innovative recognition technology that Apple bought to use in its legendary Newton device . At the same time, Newton eventually failed, and many blamed the recognition technology created by Russian scientists for the failure.
The company's founder, Stepan Pachikov, is still remembered by Newton’s work, which is far from perfect. Well, when I told on “Habré” that I was writing a book about “Paragraph” , this topic, of course, immediately surfaced .
While working on the book, I certainly tried to figure it all out - and now I want to share with you some considerations, which are partly set out in the manuscript, and partly not.
In the book, after all, it is not always appropriate to retreat strongly from the main narration and fill it with too much detail and reasoning. Meanwhile, the collapse of Newton, like any big failure, and after so many years can give good food for thought.
Not one or two errors led the project to failure. And it seems to me that their catalog will become an excellent guide to action “from the opposite” for all those who are still trying to invent something truly worthwhile and revolutionary.
So I decided to compile a catalog of errors. And that's what happened.
When connected to a MessagePad modem - the device itself was called that - it could send faxes and even, though not without the help of additional lotions, receive and read pager messages. This shows that the Newton developers saw their offspring not so much with an electronic notepad as with a pocket computer, but in addition with a communication device. Just initially, the Newton team, led by Larry Tesler, was working on a laptop-sized computer with pen input. Deciding to greatly reduce the size of the device, the creators of MessagePad could not really rebuild. They followed the same logic: the more functions, the better. This has set too many almost intractable tasks for developers. Later, the creators of Palm will do better, greatly simplifying the functionality of their device.
One of the direct consequences of the redundant functionality of Newton was the high price. MessagePad cost $ 900, and this is without additional gadgets. In modern prices - more than 1500 dollars. Serious money, especially for the device, which in essence was an electronic notebook. At the same time, many additional options Newton, which cost money and should have made the small MessagePad a competitor to a large computer, had very limited functionality or actually worked poorly. The machine, for example, could send a fax, but could not receive it.
Even today, we usually open a laptop to write a letter or work with a table. Smartphones do not replace computers, but complement them, allowing you to continue working on the go - sometimes you have to respond briefly to your email or correct a couple of digits in the budget. However, the creators of Newton, trying to make a handheld computer, underestimated the value of the bundle MessagePad with the desktop . The first users faced serious difficulties even when they tried to synchronize their calendars or address books created in computers. That is, the Newton team developed many unnecessary functions and missed one of the critical ones.
The situation was further complicated by the good advertising that Apple made in its development. In January 1992, company CEO John Scully spoke at the CES trade show with a keynote speech that predicted the emergence of a new category of electronics - "personal digital assistants." This was followed by two years of spectacular speculation. By the release of Newton, market expectations were warmed to the limit - and far exceeded the real capabilities of the device, even if they were implemented without flaws.
It was more difficult to make such a cleverly conceived MessagePad than Apple had expected. However, after Scully’s speech, the company couldn’t feed the public with breakfast for a long time - the release was already delayed. As a result, the Newton team worked in terrible hurry and stress (one of the employees even killed himself). Closer to the premiere, the work schedule did not take into account the real situation - the product was moved forward, despite the flaws. As a result, the operating system came out with a lot of bugs. One of them turned out to be critical: the first MessagePad was periodically filled with RAM, which completely paralyzed the work of the recognizer, who also used it. It helped only reboot the device.
To sustain an unrealistic launch schedule, Apple in the last year before the premiere began to seriously build up the team that worked with Newton. In a sense, this only aggravated the problem: each new engineer had to get in the know so that he could understand the device architecture and 750 thousand lines of code. The more the team grew, the less manageable it was, the less effective each engineer became individually.
When Apple licensed the Paragraph Recognizer, no one knew how Newton would turn out. The project was in its infancy, both in terms of hardware and software. And the recognizer itself was just a promising prototype. Therefore, at the start, no one could even assess whether there is enough speed and memory for the full work of the recognizer - and yet, according to the authors, this was the key interface of the device. In recognition, the Paragraph algorithm relied on the built-in dictionary. But the first MessagePad could only hold seven thousand words. The processor speed also did not allow complicating the algorithm - recognition did not happen so instantly.
Apple considered the Paragraph only as suppliers of the technology they needed - but no more. By virtue of the company's inherent secrecy policy, at first Russian scientists did not even know on which device their algorithm would work. About making them “co-authors” of the final product, there was no question at all. But recognition technology is closely related to the interface in which it is used. A “paragraph” could solve many of the problems of the first Newton if it became a full participant in the process : for example, make a letter-by-letter mode in which it would be easier for users to correct the error and enter a word that is not in the dictionary.
The biography of Steve Jobs clearly shows that leadership in the development of innovative products plays a critical role. However, before the launch, the Newton project lost two key people on whom it depended. First, Apple left Larry Tesler, who led the Newton team. A month before the premiere due to poor financial results, the company lost his job as CEO John Sculley. He more than anyone believed in the future success of the device - it was with the blessing of the CEO of the Tesler team began to develop a "pocket" gadget. However, Scully could not bring the matter to the end.
* * *
Of course, there is an extremely simple explanation for the failure of Newton - he was too ahead of his time. However, even with all this, his fate could have been quite different.
If the company focused on key features. If it were not in a hurry with the release. If it hadn't overheated the market with excessive promises. And if I trusted my partners more ...
There you have valuable instructions, in case you suddenly decide to do something revolutionary. Focus on the essentials. Soberly assess your strength. Try not to overdo it with advertising. Do not release a product with critical flaws. Work as a team ... in any case, this is what I learned from the whole story.
When it comes to Apple Newton, most often they recall not working recognition on its first models (over the years, almost all the flaws have been eliminated, but it was too late).
Part of the blame for this, of course, lies in the "Paragraph." But in many ways the flaws of the technology were due to circumstances that remained outside the control of Russian scientists.
They created the best recognition technology of their time. But due to circumstances, even this best technology was still not good enough.
Do not agree? Have something to add? Let's discuss.
The global success of the Russian company brought an innovative recognition technology that Apple bought to use in its legendary Newton device . At the same time, Newton eventually failed, and many blamed the recognition technology created by Russian scientists for the failure.
The company's founder, Stepan Pachikov, is still remembered by Newton’s work, which is far from perfect. Well, when I told on “Habré” that I was writing a book about “Paragraph” , this topic, of course, immediately surfaced .
How can advanced design disappoint users so much? Why didn't Newton really take off? And, most importantly, is the team of Russian scientists really to blame for his failure?
While working on the book, I certainly tried to figure it all out - and now I want to share with you some considerations, which are partly set out in the manuscript, and partly not.
In the book, after all, it is not always appropriate to retreat strongly from the main narration and fill it with too much detail and reasoning. Meanwhile, the collapse of Newton, like any big failure, and after so many years can give good food for thought.
Not one or two errors led the project to failure. And it seems to me that their catalog will become an excellent guide to action “from the opposite” for all those who are still trying to invent something truly worthwhile and revolutionary.
So I decided to compile a catalog of errors. And that's what happened.
Too many features
When connected to a MessagePad modem - the device itself was called that - it could send faxes and even, though not without the help of additional lotions, receive and read pager messages. This shows that the Newton developers saw their offspring not so much with an electronic notepad as with a pocket computer, but in addition with a communication device. Just initially, the Newton team, led by Larry Tesler, was working on a laptop-sized computer with pen input. Deciding to greatly reduce the size of the device, the creators of MessagePad could not really rebuild. They followed the same logic: the more functions, the better. This has set too many almost intractable tasks for developers. Later, the creators of Palm will do better, greatly simplifying the functionality of their device.
High price
One of the direct consequences of the redundant functionality of Newton was the high price. MessagePad cost $ 900, and this is without additional gadgets. In modern prices - more than 1500 dollars. Serious money, especially for the device, which in essence was an electronic notebook. At the same time, many additional options Newton, which cost money and should have made the small MessagePad a competitor to a large computer, had very limited functionality or actually worked poorly. The machine, for example, could send a fax, but could not receive it.
Lack of critical features
Even today, we usually open a laptop to write a letter or work with a table. Smartphones do not replace computers, but complement them, allowing you to continue working on the go - sometimes you have to respond briefly to your email or correct a couple of digits in the budget. However, the creators of Newton, trying to make a handheld computer, underestimated the value of the bundle MessagePad with the desktop . The first users faced serious difficulties even when they tried to synchronize their calendars or address books created in computers. That is, the Newton team developed many unnecessary functions and missed one of the critical ones.
High expectations
The situation was further complicated by the good advertising that Apple made in its development. In January 1992, company CEO John Scully spoke at the CES trade show with a keynote speech that predicted the emergence of a new category of electronics - "personal digital assistants." This was followed by two years of spectacular speculation. By the release of Newton, market expectations were warmed to the limit - and far exceeded the real capabilities of the device, even if they were implemented without flaws.
Unrealistic deadlines and raw product
It was more difficult to make such a cleverly conceived MessagePad than Apple had expected. However, after Scully’s speech, the company couldn’t feed the public with breakfast for a long time - the release was already delayed. As a result, the Newton team worked in terrible hurry and stress (one of the employees even killed himself). Closer to the premiere, the work schedule did not take into account the real situation - the product was moved forward, despite the flaws. As a result, the operating system came out with a lot of bugs. One of them turned out to be critical: the first MessagePad was periodically filled with RAM, which completely paralyzed the work of the recognizer, who also used it. It helped only reboot the device.
Uncontrolled growth
To sustain an unrealistic launch schedule, Apple in the last year before the premiere began to seriously build up the team that worked with Newton. In a sense, this only aggravated the problem: each new engineer had to get in the know so that he could understand the device architecture and 750 thousand lines of code. The more the team grew, the less manageable it was, the less effective each engineer became individually.
Too many unknowns
When Apple licensed the Paragraph Recognizer, no one knew how Newton would turn out. The project was in its infancy, both in terms of hardware and software. And the recognizer itself was just a promising prototype. Therefore, at the start, no one could even assess whether there is enough speed and memory for the full work of the recognizer - and yet, according to the authors, this was the key interface of the device. In recognition, the Paragraph algorithm relied on the built-in dictionary. But the first MessagePad could only hold seven thousand words. The processor speed also did not allow complicating the algorithm - recognition did not happen so instantly.
Distrust of partners
Apple considered the Paragraph only as suppliers of the technology they needed - but no more. By virtue of the company's inherent secrecy policy, at first Russian scientists did not even know on which device their algorithm would work. About making them “co-authors” of the final product, there was no question at all. But recognition technology is closely related to the interface in which it is used. A “paragraph” could solve many of the problems of the first Newton if it became a full participant in the process : for example, make a letter-by-letter mode in which it would be easier for users to correct the error and enter a word that is not in the dictionary.
Loss of leadership
The biography of Steve Jobs clearly shows that leadership in the development of innovative products plays a critical role. However, before the launch, the Newton project lost two key people on whom it depended. First, Apple left Larry Tesler, who led the Newton team. A month before the premiere due to poor financial results, the company lost his job as CEO John Sculley. He more than anyone believed in the future success of the device - it was with the blessing of the CEO of the Tesler team began to develop a "pocket" gadget. However, Scully could not bring the matter to the end.
* * *
Of course, there is an extremely simple explanation for the failure of Newton - he was too ahead of his time. However, even with all this, his fate could have been quite different.
If the company focused on key features. If it were not in a hurry with the release. If it hadn't overheated the market with excessive promises. And if I trusted my partners more ...
There you have valuable instructions, in case you suddenly decide to do something revolutionary. Focus on the essentials. Soberly assess your strength. Try not to overdo it with advertising. Do not release a product with critical flaws. Work as a team ... in any case, this is what I learned from the whole story.
When it comes to Apple Newton, most often they recall not working recognition on its first models (over the years, almost all the flaws have been eliminated, but it was too late).
Part of the blame for this, of course, lies in the "Paragraph." But in many ways the flaws of the technology were due to circumstances that remained outside the control of Russian scientists.
They created the best recognition technology of their time. But due to circumstances, even this best technology was still not good enough.
Do not agree? Have something to add? Let's discuss.
Quote from the book : The
press was merciless. “Apple promised too much, but it never managed to create a useful device,” the New York Times reported. “Newton is practically useless,” PC Week magazine summarized.
A bad start, of course, could not but spoil the relations of partners. When at conferences Apple representatives were bombarded with complaints about the poor performance of the recognizer, they sometimes could not resist not to blame the Russian development.
In turn, Stepan Pachikov, defending himself from criticism, could in an interview with a journalist mention the weak computing power of the MessagePad, which severely limited the capabilities of the recognizer.
The difference lay in the fact that the powerful Apple treated the encroachment on its image much more zealously than a small team of Russian engineers and scientists could afford.
The Vice President and Head of the Personal Devices Division even sent a letter to the founder of the Paragraph, politely but unequivocally asking not to mention their joint creation in a negative context.
- The book on "Paragraph" is still in the works - but it can already be read. I published several chapters on Habré. Links - below. You can also subscribe to the project newsletter and get all the finished chapters in one file .
"Pioneers of Silicon Valley". Maxim Kotin's book
Chapter 1. “Wait a minute,” said Mzhavanadze
Chapter 2. Scientist-watchman
Chapter 3. Computer people
Chapter 4. Beta
Chapter 5. "We can solve any problem"
Chapter 6. Wild East
Chapter 7. Eight hrenyatin
Chapter 8. I am a walrus
Chapter 9. Armature into caterpillars
Chapter 10. American
Chapter 11. Language D
... to be continued ...