Julien Boyreau

Months ago, I wrote about what I was seeing as the most radical redistribution of roles in the computer industry since the emergence of personal computers at the beginning of the 80s.

The past months came up with so many announcements and news that it's hard not to be sunk.

However, here are the most interesting questions for me.

1) The Form Factor Battle: Netbooks are a mirage

There is not a week without one of the major companies in this sector trying to come up with a new moniker for this "bizarre" no man's land of 5 to 9'' screen computers.

First Intel reinvented the term Netbook to name them.
Then, Microsoft and nVidia started to try to deconsider this new term with the most simple "low cost small" laptops.
Recently, Qualcomm and Freescale started promoting the "Smartbook" name.

I am more on the same camp as Microsoft and nVidia on this: I don't think that Netbooks can survive as a standalone category for a long time.
As defended at many places on the Web, these computers are sieged between pocketable mobile phone computers that are much more portable and laptops that are much more versatile.
In the 2 to 3 years coming, pocketable will become more and more powerful and able to connect to massive screens and keyboards. In the same time, laptops will become less expensive with Moore's law.
So Netbooks will join the "nice attempt" graveyard of network computer, smart terminal and Co.
There are definitely no room for this cross over: they will die.

2) The Processor Battle: Intel vs. ARM

After exiting the mobile phone ARM-based industry years ago, Intel has come back with an AMD-promoted "x86 everywhere strategy", launching the Atom Brand towards sub-laptop territories, keeping power and compatibility but working on power savviness and integration.
I am quite sure that Intel's focus on Netbooks with the current generation of Atom is just a transitory step towards the main end game.
With the Medfield chipset, due for end of 2010, Intel will be able to target eventually mobile phone and will be in an face to face battle with the ARM's ecosystem, led by Qualcomm, TI and nVidia.
Sure, we'll see mobile phones with Medfield in 2011, probably from obscure Taiwanese companies without enough brand power to really make a difference.
The real deal for Intel would so to sign a deal either with one of the Top Tier mobile phone companies (Nokia, Samsung, LG, Sony Ericsson, Motorola, Palm, RIM, HTC and Apple) or with one of the Top Tier desk/laptop companies, that will eventually go to pocketable computing (Dell, HP, Acer, Asus, Toshiba, Sony...).
I think that whatever the deals, Intel will need to find another angle of attack than "x86 compatibility".
Medfield will need to provide a real breakthrough in term of power vs. battery life to have a chance to win.
We'll see.

3) The OS Battle

In the 2 to 3 years coming, we may be on the verge of a radical change in OS dominants players.
Mobile phone OS suppliers (Palm WebOS, Symbian OS, Mac OS X, Android...) will move up to laptops with cleaner, more web oriented, touch oriented interfaces, able to rival Windows.
And once the Netbookmania has blown off, all the interesting work done (by Jolicloud, gOS Cloud, Intel's Moblin or my friend Olivier Seres' iFrame ) will migrate both to the simplified "laptop segment" and to the pocketable computing world.
I can't know yet who will win and who will loose, whether the 25 year old OS / Hardware unbundling will survive but this is one of the most interesting question of all.
Will the "slimmed down Linux exposing only a browser to the Cloud" work and mark the decline of computing power democratization we've seen for the next 50 years?
Will a dramatically new kind of kernel emerge as we're celebrating the 40th anniversary of Unix?

Finally many questions, few answers, so stay tuned!

9 months ago, in this article, I wrote about that many companies were trying to beat Google and detailed 3 questions towards this goal.

Today, I read about a new website coming in May, made by Stephen Wolfram (who made Mathematica), namely Wolfram Alpha.

Wolfram himself announced and introduced at this blog , and Nova Spivack, who made Twine, had chance to spend two hours with Wolfram being demoed the website, and then commented back about it.
Right after, Jon Stokes, who I cherish the damn insightful articles around chips at Arstechnica, expressed quite abruptely his doubts about the site.

Let's look at this site through my 3 questions:

1) Doing better
On this point, Wolfram Alpha (WA) seems to be on a right track. Following exactly the path I laid out (coincidence obviously ;) ), WA will be to give ANSWERS to requests, NOT just PAGES with the words told.
You will type questions in a typical Google-like search bar and WA will answer you.
You might so not cobble with the toundra of pages from Google to find the answer you're (implicitely) looking for when telling words to google.com.

2) Doing it different
On this point, WA seems to be exactly as Google, having to build a (guessed) big database and (guessed) big web sites to centralize all the logic to find answers and serve them to people.
Even worse, as explained by Spivack, the tool has to be bootstrapped and fueled by an army of persons entering tons of information by hand only for the sake of WA and maybe a ever growing total paycheck.

3) Monetizing different
On this point, no information is given.
Jon Stokes made his best point on this: even if the process is (sometimes) painful to find an answer with Google, we humans have quite been mastering the process and are able to go through it quite easily.
So, beyond the possible engeeniring masterpiece, what can WA be useful for?
As I explained, the biggest limit of Google is not for humans to find answer but that computers are completely unabled to use it in order to act for humans.

So what may be the most valuable part of WA is NOT THE HUMAN NATURAL LANGUAGE INTERFACE but the COMPUTER ABSTRACTED INTERFACE.
This could be opened and monetized to any other computer in order to get the answers and act upon them.
To take an example back from Spivack, it's pretty easy for an human to get the price of a share from Google with google.com but pretty impossible for a computer provided that a human read through the cumbersome functions written in XML to tell the computer to do so SPECIFICALLY for this question.
As far as I understand WA, it can contain a kind of a way for a computer to ask GENERALLY, making computers as empowered as we are with search/answer engines.

I've just come back from 3 weeks of holidays and a good discussion with my parents, especially on my philosophy and libertarians.
That's a recurrent process: when I talk with a person about that I am a libertarian, the person is amused at first.
Then when that person is a socialist, I am accused of being sold to big corporations, money or whatever.
When that person is more right wing, I am disdained to be just an idealist because my view can not work.

The biggest hurdle to this kind of discussion is that most often the person is orienting the talk by bombarding me with questions on how to solve this or that "problem" without forcing people with a state.
I previously used to be trapped in this by trying to draw how people without a state would make what some want for any case such as security, education, insurance, roads or whatever.
This always leaded to a kind of brainstorm ping pong on actually designing in details such an answer to a problem currently dealt by the state.

After reading and thinking about it, I now try to make my opponent shift to a deeper and more subtle understanding of being a libertarian.
For me, the main difficulty to compare libertarians and "democrats" (not the US party but the ones that love, trust and enforce democracy) is due to that they are not aiming at the same questions.

Democrats and democracy is a question of WHAT to do WHATEVER the means, FORCE INCLUDED: the way is to let each person in the place vote and do what the most want.
More precisely, in today representative democracy, persons are voting for WHO is going to tell what to do.
That is the rule of the strongest: the state is the strongest and it is making the laws.
Once this has been done, the government has a limited period to do what it wants with any means.

Libertarianism is a question of HOW to do WHATEVER the aims: the way is to let each person in the place do what he wants except forcing people, be it as an end (like a serial killer) or as a mean (like a state).

So as a libertarian, when I want the state to stop acting, this does not mean that I don't want it. It means that for me NOTHING should justify to force/racket people via taxes and rules, even for apparently "heart minded reason".

Introduction
2008 marks the 10th anniversary of one among the most recent successful companies, Google.

Looking at the WWW today, one can see that Google is everywhere, nearing Microsoft like monopoly, in many domains such as revenue growth, brand recognition, management practices and computers' network expertise.

Facing this kind of power, people tend to have one of two kind of reactions.
Being angry and trying to beat the player in front of states' court.
Being fascinating and trying to beat the player in front of people, offering a better solution to them.
I am definitely at my humble level in the second category.

In this text, I will expose some of my thoughts about the way a group may beat Google in front of customers.

1) Beating Google by doing better
Even if the company has been digging in many territories in recent days as Microsoft's been moving around Windows+Office, the main role of Google is to provide what we call "search".
Search for me is an unnecessary complex word for a more simple activity: what's Google actually doing is answering questions from people.
When you're typing words to tell to Google, you're not actually "searching" but morelike asking a question.
Except for a very small subset, the way Google's answering is very primitive.
It's just getting from its giant memory the pages in which the words are: the main value in this process is so not in "what" Google is getting but how it is ordering these pages.
The main limits of Google answers is so:
You always needs to read pages to get the "real" answer.
Sometimes, the fact that Google just see "words" and not "proper nouns", "common nouns", verb or adjectives is that it misses completely the point and you're wasting time either refining your question or reading more and more pages.
Google does not read between the content words and actually encourages people not to type them.
Google is working well to answer to a person, but not at all for a computer.

The last point is the most important: the limits of Google for answering to a person is a waste of time.
But the limits of it for answering to a computer is a dead end. Because of that, one of the main aim of a computer, that is to do something for a person is lost.
When you want for example to know in which cinemas you can see a film, you'll type the nouns, Google will answer probably the pages of cinemas companies first, you will read these pages and get it.
But if you want to tell a computer to tell you that when one of the cinema is in the city where you are, there is no chance it will ever be able to do it for you, as from the words to Google, it won't read the good information.
Thus there are many room today to beat Google by doing better and actually many companies trying to do so.
One of the main problem of them, beside technology is the second way of beating Google.

2) Beating Google by doing it different.
The real reason of monopoly is rarely the undefinite leadership of a company at the product level.
It's more often a particular core pre requisite that's almost impossible to defeat.
Intel and x86 has been there for 30 years because backward compatibility kills any chance to win with a better set of instructions.
Microsoft has been the endless leader of software because of a killer business model, selling to computer makers and so spreading any new Windows version with the PCs growth.
Google is a very solid leader not because of what they do, but because of what they have to make what they do, that is the most powerful, big and efficient mainframe ever made.
It always made me laugh to see that Google's always been pride of telling that you can know anything from Google. Actually, one of the stuff you can not know is about Google's computers. Simple but customized machines, numerous (500000+), dispatched all over the world, this system is fascinating.
However it promotes a model of computing that was popular in the 60's, the mainframe, centralizing all the information and computing resource under a single group.
The main reason why all the Powerset, Twine and al will fail is not because of their product.
It's because they're following the same architecture idea that's to make a super mainframe, and at this game, no one could out beat Google.
In the same way that Microsoft defeated IBM, to beat Google is to do it differently by pushing forward personal computing with P2P and not backward to mainframes.
To do that, there's even not a need to invent complex new protocols, but to come back to what the first vision of Tim Berners Lee for the WWW was: a person's computer that is able to GET and POST as a 'browser' but also to BE GOT and POSTED as a "server".
Promoting this approach, you may decentralize the Google system and gain more privacy, more scalability and more freedom to publish your information to anyone.
This will also put back personal computers in light to rediscover a role lost that's the third way to beat Google.

3) Beating Google by monetizing different
The main thing that's people are seeing from Google is software.
But as all the successful software makers, Google is not selling it to users.
Microsoft is selling to computers makers, Apple is selling the computers and Google is selling spacetime for ads.
As told before, a computer is aimed both for a person to do something AND to do something for a person.
The first is the main pushed by Google because in this one, there will be an interface for the person to do on which you can put ads.
Thus Google's business model is pushing a person to do more and more where a computer should also be a way for a person to do less and less (unpleasant) things.
Automatisation is thus really the way to attack Google business model not by trying to grab a share of it but by repurposing computers on tasks that limits its scale.

Conclusion
Many companies are today looking indepently at these 3 axis to compete with Google.
I think that it's in the combination of both 3 that a new leader would emerge.
This will come and after the US government, IBM, Microsoft and Google, will hightlight a new leader in the 5th generation of computing.

Try a little test : google the questions words in English.
On top of the list, you will find the usual suspects, what, who, why and how.
One will be quite less used than the others, whose.

However, the whose question has been the most important, controversial of all.
It has been the cause of so many wars, murders, conflicts and tears.
It's the essence of what makes humans different.

For anything, you can always ask : "whose is it?".
When this is a person, the answer has been agreed by blood and sanctuarized by a famous text : a person is self's.

But for anything but a person, the question remains: "whose is it?'.
Knowing this is the most essential, because we've been (quite) agreeing that the one whose it is can do what he wants with it and especially gives it to someone else.
To this question are only two answers: it is either a person's or persons'.

When it's a person's, it can change faster than when it's persons'.
Indeed, a person has just to agree with himself to do something to it.
When it's persons', they have to agree on what to do with it.

The most used rule for one that's persons' today is the one of the most.
What to do with it is what the most wants.
Most often, this is implemented by voting.
People are most often voting in two cases : for what to do with a state, where a vote is a person and for what to do with a company, where a vote is a share from the company.

The biggest difference between the two are :
For a company, only people that have shares can vote. These people BOUGHT these shares so took a risk to care about the company. They can SELL these shares when they don't care anymore.
For a state, only people that are citizens for this state can vote, so:
People living in the country where a state is can not vote.
People "investing" a lot in the state are as important as the others.
Citizens can not escape from the state to give away both obligations and rights got from the state.

Whatever the case, persons' is most often slower than a person.

So, while I am hating the system for states, I do think that anytime we can have one that is a person's, we should make it.

As I wrote before, we have been witnessing since the beginning of WWW a tacit fight between two philosophical camps in the computing microcosm:

In the left corner sit the promoters of "cloud" computing, the same that tried to sell the Network Is The Computer propaganda ten years ago.
As I explained before, "cloud" is just a more cloudy term to refer to mainframe computing, the popular architecture of the...sixties!!!
The few have computers and the many have screens and keyboards to tap in these huge and shared computing resources.
In this model, you are depending on the big guys for the right to do what you want with THEIR computers.
The main forces in this Empire are the biggest on WWW such as Google, that built the biggest mainframe ever with hundreds of thousands computers, and the makers of servers like Sun and Oracle.

In the right corner sit the more hidden promoters of "even more personal" computing, building on the shoulders of computing democratizers of the 70s, especially with Peer 2 Peer (P2P) computing.
As I explained before, "P2P" is just a fancy term to refer to the next step of personal computing, where more and more is done by a person's computer, sucking the control from the mainframes.
The main forces among these Jedis are the biggest makers of personal computers, from Nokia to Apple.

The latest interesting nascent trend from the Jedis is the idea of Pocket Web Servers.
This was started by Nokia with their Web Server for s60.
This has been worked by the iJetty project, targeting the Android stack from Google.
This SHOULD be the most important feature to add to Firefox to make it a real competitor to the Facebooks of the world.

A Web Server in the Pocket could make person to person communications much simpler by enabling a direct and standardized HTTP messages between handsets.

What's lacking for this?
Number 1, the collaboration of carriers to allow then optimize incoming requests to handsets on their network.
Number 2, a more spread will of persons to buy, thus own their name on WWW to be fully independent from service providers.

I can thus dream of the day where I may have Julienboyreau.com on my iPhone, letting who I want GET FROM it and POST TO it in real time, punching back all the Twitters and Facebooks of the world.

In the short term, 4 key companies may have to choose a camp :
Google may have to choose between Adwords (The Empire side) and Adsense (The Jedi side).
Intel may have to choose between Xeon (The Empire side) and Atom (The Jedi side).
Nokia may have to choose between Ovi (The Empire side) and Web Server, s60, NSeries (The Jedi side).
Microsoft may have to choose between Live (The Empire side) and Windows (The Jedi side).

Any person on Earth knows a language.

This may be English, French, Mandarin, Spanish or whatever, but any person knows at least one among these.

Any text between persons are in one of this language.

We've been writing and speaking with each other in these languages since a long time and they all became quite robust and powerful to tell what you want.

Then since the 50s, we've been writing and speaking not only to persons but to computers.

And we've kept on "making" new languages to talk with computers, from assemblers to Javascript, XHTML and CSS by C, C++, Lisp, Java, Smalltalk...

All of these languages for computers seem to me to share false starting points.

1) They're choosing an "orientation" be it "imperative", "declarative", "object", "functional", whereas languages for persons are multi orientation.
2) They have thus a specific grammar, focused on a single concept such as "object", "function", "query", "tag", "class", "resource", "triple"...completely different from the grammar for languages for persons, that ALL exhibits COMPLEMENTARY ideas such as "noun", "verb", "adjectives", "adpositions", "affirmation", "question"...
3) They let a person choose whatever words to reinvent "identifiers" from words picked up in a language for persons, either by concatening them ("PersonsWithASalary") or abbreviating ("iframe" or "char"...).

All of these draw a line between persons that took the time to learn one of this language and the others.
With time, the so called "programmers" group has been growing but in 2008, it's still a very small one compared to the rest.

One of the refreshing idea of the past 10 years has been XML : this is not a language per se, but symbols or syntax to talk to computers.

The problem of XML are however serious by design.
1) The "tag" and "attribute" in XML are not mapped to any natural language, so two persons can equivocally do what they want with them.
2) Even if it can be made clear what a tag and an attribute are for, two relationships are not enough to tell everything you can tell in a natural language.
3) You can make whatever tag or attribute you want, exposing people to the drama of "nominalization" ("creation"), "concatenation" ("RectangleWithRoundedCorners") and abbreviation ("char").

Point 2 is particularly serious.
For example, 'animal' => 'dog' => 'Malik', 'USA' => 'San Fransisco' => 'Malik', 'Clara' =>'dog' => 'Malik', or 'Malika' => 'Malik' are 4 different meanings, that is, Malik is a dog, Malik is at San Fransico, Malik is Clara's dog and Malik is from Malika.
In English, it's easy to tell ; in XML, it's just equivocal or impossible to do.

For me, what's needed is thus not a new "language" but symbols much better than XML, but one that is :
a) Richer than the 2 "tag" and "attribute" thing.
b) Nearer from the grammar of a language, with way to see a noun, verb, adjectives...
c) Univocal to tell what a symbol is for.
d) More restrictive on words, preventing a person to reinvent a noun, verb or adjective already in a language.

I have been working for 1 year on such symbols and making recently progresses to make them into a working example.
The coolest right now is that I can tell it in plain French, let it get what I told from French for this to be told in English and for this to be thought about from English.

In the next 3 to 4 months, I hope I will show it to more people than myself ;).

Stay in touch if you're interested to overcome "programming" languages and moving to more natural interfaces.

I bought an iPhone and upgraded to a plan for iPhones 40 days ago.
My previous phone was a 3410 from Nokia with a GSM only plan (not even SMS!) so the upgrade was quite huge ;)
Everything's been told about an iPhone but I will tell you how I feel with it.

I) The present

1) It's so beautiful.
I am keen on gadgets mixing usability and art and I must say my iPhone is really the most beautiful I've ever had.
In the first few days, I was keeping on just looking at it and felt pleased.
The lines are so perfect, stripped down, "pure" that I don't know how one can be more beautiful.

2) being pervasive and unlimited mitigates how slow EDGE is.
Yes, EDGE is not the fastest and having Wifi on my iPhone is reinforcing this fact.
But coupling how pervasive it is and that it's unmetered makes me feel the biggest value similar to the transition from dial up to DSL : you don't even think about connectivity, you don't hesitate to grab some news or a map when you want it, it's just there.

3) music and videos is great.

4) the internet eventually.
Since 2000, carriers has been stuck into the idea that you should need a language (WML) and a protocol (WMP) for what's via the Web on your mobile.
Surfing the Internet on my iPhone definitely trumps this idea off !
The navigation is really simple, double tapping is smooth.

II) The future

When you're working in this industry and once you tasted an iPhone, the tough question came around : how to make this better ?

1) Short term.
At least 4 better are obvious and easier predict.
i) Faster cellular : HSPA then LTE...
ii) Better camera : 3 then 5 then 8...
iii) More memory : 16GB then 32GB the...
iv) More precise location : cell triangulation then GPS then...

2) Mid term
It became there harder to say.
v) NFC : to pay or pass gates with your iPhone.
vi) Automatic wireless connection to a bigger screen or keyboard : this will come but may be dangerous for Apple because it will put an iPhone in a competition with laptops and desktops for persons in the mainstream.
vii) Automatic wireless backup to a bigger and safer memory : would it be at home or at Apple's ?

3) Long term.
Somewhere in 2009-2010, an iPhone with the previous will come out ; it will be engined by a chipset as Moorestown from Intel, fast enough to do most of what the mainstream want.
It will be the ultimate monolythic-and-carried-in-a-pocket computer, to get only smally better.
Here will come the need for a new era, more than personal, not monolythic, not carried in a pocket, the first symbiotic computer.

60 years ago, John Von Neumann made a revolution when he figured out his famous architecture for computers.
Computers at this time were told what to do as any other machine ; by plugging cables once to do only one thing.
Von Neumann introduced first the idea of a computer with a memory to remember.
Some persons used this memory idea to remember as digits numbers, colors, texts.
Von Neumann then introduced a second more important idea : a computer able to remember an instruction aka "what to do".

Since then a computer was able to get from memory what to do as instructions.
This made an interface between circuits made with transistors to "do" the instructions and something else to make many instructions triggered by only one instruction from a person.
This gave birth to what we're calling now "software", the capability for a computer to do many things from one thing from a person.

The problem since then has always been the gap between the instructions done by circuits in a computer and the instructions a person can tell to do.

A person is thinking and telling to other persons in a language such as English, French, Mandarin or Spanish whereas a circuit can understand only instructions in a so called "machine language" to move, calculate and compare numbers.

The history of software is the one of narrowing this gap by making stuffs between "natural" languages and "machine" languages.

This history is full of attempts from assembly language to C, C++, LISP, Smalltalk, Java, Python, Ruby, Perl, Javascript, HTML, CSS...

This history is also full of (very) successful companies, from Microsoft to Google.

Any successful company today has to deal with software.

What is the most striking to me is that NONE of the most successful software makers don't sell it to individuals using it!

Yes, Microsoft is selling software but gets 90% of its sales and profits by selling it either to PC makers selling to people or to companies buying PCs for employees or servers for themselves.
Yes, Apple is selling to people but gets 90% of its profits by not selling the software but the machines (iPods and Macs) in which the software is.
Google is not selling software at all but the spacetime people spent using its software to advertisers.
Adobe is selling software not to users but to groups making either software, videos or images.

That is the rule to remember : if you want to get big profits from making software for people, don't sell it to them !!!

For 6 months, I have been reading more and more often texts polluted with this "cloud" meme.
Many among the most brilliant and salient tellers keep on talking about the magic of the cloud, how it will change everything and how good it is.

Be careful ! This is just the latest version of the famous "the network is the computer" propaganda made by Sun Microsystems and Oracle 10 years ago.

To get back on Earth, let's think about the underlined question behind the cloud : whose are computers and how many are they ?

1) In the 50s, computers were few and states'.

The first commercial computer was an UNIVAC and was sold to the US Census Bureau. Around 50 were made later and sold either to groups in states or to the biggest companies.
These had very big problems to solve and as a computer at this time was for 150 000 dollars, only big groups could afford them.

2) In the 60s, computers were few and companies'

The posterchild computer for this time was PDPs sold by DEC to more and more companies as they were for around tens of thousands dollars.
At this time yet, there were mostly one or two computers in a group but more and more persons wanted to use it.
So they had to SHARE it, that lead to persons responsible for deciding who to use it and to use it for.
A person had a screen and a keyboard to tell the group's computer to do something for him.

3) In the 70s, computers were many and companies'

Moore's law dropped the price of transistors at a point where you could make a computer cheap and small enough for companies to buy a computer for more and more employees.
An employee did not need now to share a computer.

4) Since the 80s, some computers were made a person's

The revolution of "personal" computing is less about "microprocessor" or "GUI" or "Apple" than that for the first time, a computer was "really" made a person's.
Since now, a person can decide what to do with his computer by getting software.

5) Will 2010 be like 1951 ?

Here we are : behind the cloud, there is this trend to come back to the 50s where :
There are more and more computers.
There are all in less than 10 gigantic mainframes that are companies', such as Google, Microsoft, Yahoo, eBay and...Facebook ?
A person has only a screen (to touch) ; anything else is Google's.

Some problems may be rooted in this :
The more you tell Google's, the more dependent you will be on one provider.
The more you tell, the harder it will be for you to switch to another provider.
The more many tell, the more painful it will be when there is a failure of Google.
Google will be more and more tempted to sell what it knew from you to advertisers.
There will be a day when you want to do something Google does not want : without your computer, you won't be able to do it.

6) Or will there be a personal computing new champion ?

Let's face it : Microsoft screwed the personal computer, by delivering bad updates (think V...) and forgetting his DNA to seduce corporations and advertisers, running behind Google.
The reaction to it is a belief that the best way to beat Microsoft is to come back to the 50s with mainframes.
We already know the leader of this camp, Google, a company that's morphing into the third giant in computing history, after IBM and Microsoft.
I am still waiting for a company that will move forward the personal computing revolution, sell a computer to a person, investing in P2P technology...

Will this be Apple, Nokia or someone else ?