In 2001, for the vast majority of the inhabitants of the planet Earth Internet and Internet Explorer was almost the same. Attempts to oppose this monopoly something of their own predictably ended in nothing. Pointless to even try.

In June 2001, in an atmosphere of absolute secrecy, Apple started development of its own web browser. To organize and carry out this adventure (see the reporters about this project, “gamble” would be the mildest definition conceived) has assigned Scott Forstall.

The project team, challenging a huge monopoly, consisted of two people. In April, the company Eazel finished the development of the Nautilus 1.0, the shell makes Linux the operating for the casual user of a competitor for Windows and Mac OS X, besides free and on the same day reduced two-thirds of his state. Apple took the best of those who were under this reduction.

Don Melton, who became technical Manager for an adventurous project, in Eazel led the development of the Nautilus project, Ken Kocanda – still the only engineer in the team was one of the programmers in this project. Of these, only don Melton has had some experience in the development of browsers (the transformation of Netscape into the open source Mozilla – he’s been clearing the code from obscene comments).

Common sense tells us that nothing good they could not get…

Nothing good is happening!

That they have to work together, and their first job in the Apple will be writing a browser better than Microsoft Internet Explorer, don and Ken heard by signing a non-disclosure agreement and read the rules of internal order (anybody nothing to tell the code name of the project outside of the working room and office of the head not to utter, and similar infringements on freedoms).

Unknown, of course, is intriguing and arouses curiosity, but that you would have signed up for execution is clearly something very trivial and terrible secret, not having any idea what you have to do?

In the days of Steve jobs it was a common practice at Apple. Almost all who created this company wonders, went through it. Crazy. The normal in such circumstances, the work did not agree.

But we digress. To cope with the daunting task required a year and a half. It wasn’t all bad – vacancies in the project were many, just like a madman with appropriate qualifications find it very difficult. A veteran of the development of other browsers, usually the first thing I told the interviewers that they are willing to do anything, just not browsers – hopeless, tired, tired.

Ken never participated in the development of browsers, so don had to spend two days on an educational program. To explain what HTML, CSS, JavaScript, and other alphabet soup.

As a true (and they were required not just quality but the best in the world) the browser from scratch in a year or two is absolutely impossible, don and Ken went in search of a suitable browser is open source and sparing license, which would be the basis for their development.

And that was natural first candidate for such a framework became Mozilla. In fact, they made up a list of potential “donors” in the sixth position which was little known Konqueror. But we started with the first position.

 

The source code of Mozilla was huge: almost half a million rows. Versions for Mac OS X in Mozilla distribution was not only for Linux, most Mac OS X was only three months old, and between Darwin (the system without a graphical interface within Mac OS X, a BSD clone), and Linux had much in common, but the differences were very many.

To just build Mozilla for Mac OS X, gone a week. The browser icon appeared in the dock and then disappeared. A week for the first time managed to get to open a blank window. But to open the web page failed.

Weeks of hard work (seven days a week, from early morning until late at night) did not yield any encouraging results. Guide (Scott Forstall) have showed impatience, and doubt their ability to do something useful.

Six weeks from the beginning of development and failed to open even the most simple web page, it was a failure.

Meanwhile, there were any crazy.

Richard Williamson

Soon don Melton interviewed a new candidate (the previous one was rejected or refused to do), who spoke with a British accent, and was so confident that I don have any doubts.all technical questions the candidate answered correctly and, apparently, the required qualifications possessed.

However, don Melton was ready to deny a candidate, but then intervened in the case of Bertrand Serlet, senior Vice President of software business supporting a candidate. In the early ‘ 90s, Richard worked under his direction in the NeXT, and did very well. Accepted it.

Richard started his first day with questions about what was done to him, and in what actually a problem. His confidence was annoying, but the questions were appropriate.

Not commenting on the answers of colleagues, Richard continued to ask questions, and then took time out to “think”. Two days later, he invited don and Ken, promising to show them something interesting. Intrigued.

On a Mac Richard browser with unknown icon famously launched without problems opened the requested web page, and then another, successfully passed the link on this page, came back – everything worked. What took six weeks reached two days?

However, the text in the browser appear crooked, the fonts were terrible, support of styles and sizes of fonts are absent in principle – but it didn’t matter. Genius?

Richard is not just asking questions. He looked at what was happening with fresh eyes, and come to unexpected conclusions, which did not report, deciding first to try, and for two days I implemented their ideas.

Instead he chose Mozilla Konqueror. At the feature set, this browser is in no way inferior to Mozilla, but the size of source code was much smaller.

Have Konqueror’and, like Mozilla, there was a version for Mac OS X, but Richard was not with the bat to adapt though smaller in size but still huge, source code in Mac OS X. Instead he wrote “jumper” (shim) between Mac OS X and the whole KDE system. It is very difficult, but if you do not pay attention to irrelevant details is not impossible. That is, after all, a genius.

Also, instead of Core Graphics (graphics libraries Mac OS X) he used X Windows, UNIX marketing graphics package was part of Mac OS X, and more or less compatible with KDE (the graphical inaccuracies and ugliness Richard decided not to pay attention to).

Now was that the answer to Scott on his daily and increasingly snide question. The news reached Steve jobs.

The scope of work and deadlines

In good, working on your own browser is only supposed to begin. And in order to meet the deadlines, the code “base” subjected to scrutiny. Next to Mac AMI set some LInux machines with installed on them Konqueror’ohms, to be able to study the source mechanisms in the dynamics.

The source code compared to Mozilla, looked simple and compact – in the top-level directory was only two subdirectories, KHTML and KJS. Only three hundred-odd files.

Counted the number of lines in the code (in this count ignores empty lines and lines with comments), turned out 120 thousand lines. In 2003, Steve was called a different number of rows – 140 thousand, who have been more right I don’t know.

In any case, it is more than ten times less than in Mozilla.

The approach was new crazy (who agreed to take up the project of the fact that they knew nothing, but in the opinion of its leaders is able to benefit him), and the first phase (the present transfer Konqueror’and Mac OS X), it was decided to complete in 4 months.

And the whole project was supposed to take 18 months, including the already spent two.

To be continued

 

Advertisements