New, faster browser on the way?

[Stormig]

http://www.wired.com/news/infostructure/0,...7,57393,00.html

 

Teen's Web Browser Wows Top Geeks

 

By Karlin Lillington | Also by this reporter Page 1 of 1

 

02:00 AM Jan. 27, 2003 PT

 

MULLINGAR, Ireland -- A 16-year-old Irish student has startled the Internet world by creating a new Web browser that may -- or may not -- be able to goose browsing times by 600 percent.

 

But the speed claims for the browser -- which comes with 120 search engines, all the major media players, a DVD player, and an HTML and WAP editor built in -- have drawn criticism because no one has yet been able to clock it.

 

Adnan Osmani, a shy but cheerful student at St. Finian's school in this town about 50 miles from Dublin, spent two years coding what he calls his "Xwebs megabrowser" for an annual Irish school science challenge, the Esat BT Young Scientist of the Year competition.

 

"I just wanted it faster, for myself," said Osmani, the son of two doctors. "I wanted to make the perfect browser."

 

Osmani's programming precocity is undoubted, said judges for the competition. Larry Taylor, a judge from the competition's sponsor Esat BT, said Xwebs "is of the level of a final-year project by a university student and would be given the highest marks."

 

Osmani said he has developed an algorithm that speeds up the way information is handled inside the browser. The technique takes advantage of the features of a particular type of server used widely on the Internet.

 

The browser handles multiple requests for information, he said. So, instead of a single stream of information, several streams are processed simultaneously. In essence, the task of bringing over a Web page is divided into a set of smaller tasks, cutting the time it takes to reassemble a Web page on the computer screen.

 

But the judges evaluated Osmani's browser without even considering its speed, as they could not independently benchmark it and lacked complete access to the browser's source code, Taylor said.

 

Osmani won't reveal more of the code because he is exploring the possibility of patenting or copyrighting aspects of his browser. Some judges said they were unsure whether the process would work in the real world, but others believe it could.

 

One judge, Intel Ireland's head of process engineering, Leonard Hobbs, said: "Certainly the capability is there. The technique he invented seems unique."

 

But Hobbs de-emphasized the speed claims. "What impressed us most of all is he absolutely knew what he was doing. It was a complete work, a whole." Osmani's project demonstrated "the science of the Web," he said.

 

MIT principal researcher with Media Lab Europe, Gary McDarby, said he was astonished by the teen's "years ahead" programming skills. Even if the speed claims prove false, McDarby said, "What he's certainly doing conceptually is raising the bar for the commercial companies."

 

The browser is based on the version of Microsoft Internet Explorer that third-party developers use. But rather than using Visual Basic, Osmani used an older language called Borland C++, which meant he also had to use some Microsoft tools to translate from one language to another.

 

These processes generate thousands of extra lines of code -- a point raised by critics on sites like Slashdot and Fark, who couldn't see how any single developer could write the one-million-plus lines first claimed for the browser.

 

Osmani, an entirely self-taught programmer who got his first PC at 10, said his computing heroes are Apple founders Steve Jobs and Steve Wozniak ("because they were hackers and hands-on"), as well as hacker Kevin Mitnick.

 

The teenager, who says he'd like to attend Harvard University, was inspired to start work on his browser after a previous winner, Sarah Flannery, received worldwide attention for a new cryptographic algorithm. The algorithm eventually proved to be breakable, but only after top international scientists spent months trying to crack it.

[Azat]

They are completely right to criticize as it is not the browser speed that counts but the connection speed.(just like you mentioned) The one thing that he has done that will increase some of the speed is that he saves the DNS info on the local machine. Thus cutting that half a second or whatever it takes to get the IP for the URL that you have put. His app has not proved much since the date that it was announced over a month ago.

 

You have got to be kidding. I bet you can put a team of 3 or 4 thousand and they can't come up with something even close to Windows in a year period.

[nairi]

I bet Sipan can come up with it all by himself. In fact, I bet he already has

[Sip]

Nairi: A+

Azat: C-

 

 

 

Azat, I disagree with you on this one! Programming is just one of those engineering/design/creative things that simply having more programmers, engineers, designers does not mean you will have a better outcome. Sure sometimes it's just a get-it-done type of thing where any b0z0 could do it. But to create something new and amazing requires a LOT more than just knowing "syntax and semantics".

 

Windows was maybe somewhat of a poor example on my part ... the main difference is that Microsoft has such vast amounts of resources that it is inevitable that they will make something cool happen. Just their beta-testing program alone is something amazing ... by having thousands and thousands of people beta-testing things, it's not unreasonable to expact that they eventually will figure out some of the serious bugs.

 

By the way, have you heard of Squeak? If you haven't, just check it out briefly if you have time: ABOUT and it's FEATURES. It's a very interesting piece of work ... it somewhat an example of what I was trying to get at above.

 

[ February 23, 2003, 05:42 PM: Message edited by: Sip ]

[Garo]

[Azat]

Okay. Even with that I can tell you that no matter how many developers(1-10000) you put together you cannot develop a product as complex as Windows in 1 year.

 

And it does not even have to be Windows. I have worked on huge projects(4-6 millions lines of instructions) that would do Enterprise Management and it may seem easy to develop that tool from an outsiders prospective but it is not. Huge projects take lots of planning and coordination. It seems like the time it takes to develop these products is exponentially elated to size.

[Garo]

I don't get what's unique with his browser. As the article says it is based on Internet Explorer's codebase and I think he may have just used/made faster algorythm for source code parsing. The parallel processing the article mentions does exists in Internet Explorer and other browsers (threading). The DNS caching Azat mentions existed in Windows all the time. You can simply edit the 'hosts' file in your Windows' directory and put any domain name/IP pairs as you want. It may increase the speed of name lookup but it has a built in problem - you simply can not put all the internet domains in this file, plus every time a domain changes it's DNS record you need to update your 'hosts' file otherwise you won't be able to access the domain.

[Azat]

Garo, I completely agree with you about the DNS issue and know about the hosts file.(by the way, it is pretty cool way to play jokes on people by sending them to some stupid site when the type msn.com ot yahoo.com or some such)

[Sip]

Sweet! Thanks for the info Garo. I had so much fun with Borland C/C++ 3.0 and Turbo Pascal when starting to learn programming. That was really fun back then.

 

Azat, I agree that 3 or 4 programmers can't possibly put together 60 million lines of code (that's what I heard XP is at right now) in a few months. I am just questioning the fact that whether 60 million lines of code is required to do what windows does!

 

I am not saying Microsoft has done a poor job ... not saying windows is an inferior product. I fully acknowledge the fact that it has been an evolutionary process and there are faaaar too many variables that play into developing something as complex and windows. But I still believe that a very small group of EXCEPTIONAL programmers, can write something superior in a very short time period.

 

Of course it's not going to be easy and there may only be a very select few on earth that could get it done, but I think it can be done. I admit, it just sounds like an empty opinion but never-the-less, I do believe it

[Azat]

Sip there are many famous cases of simple software errors that have caused millions of dollars of damage as well as human life.

 

Therac-25 - Faulty software in a Therac-25 radiation-treatment machine made by Atomic Energy of Canada Limited resulted in several cancer patients receiving lethal overdoses of radiation

 

Mariner 1 - A unmanned rocket(probe) that was suppose to go to Venus. After launch it went off course and NASA had to blow it up. They later discovered that it was 2 lines of Fortran code that caused the mis-function. Cost the government 80 million.

 

Patriot Middles - The U.S. Patriot missile's battery successfully headed off many Iraqi Scuds during the Gulf War. But the system also failed to track several incoming Scud missiles, including one that killed 28 U.S. soldiers in a barracks in Dhahran, Saudi Arabia. The problem stemmed from a software error that put the tracking system off by 0.34 of a second

 

Denver Airport - The Denver International Airport was intended to be a state-of-the-art airport, with a complex, computerized baggage-handling system and 5,300 miles of fiber-optic cabling. Unfortunately, bugs in the baggage system caused suitcases to be chewed up and drove automated baggage carts into walls. The airport eventually opened 16 months late, $3.2 billion over budget, and with a mainly manual baggage system

 

and there are many many more.

[Azat]

Did I ever tell you that I studied Software Quality Assurance as part of my MBA. BARF. I use to love it but I hate it now.

[Azat]

Garo, didn't Borland change their name to Enprise or Inprise or something like that?

[Sip]

oh and about this supposedly "fast" browser ... the article doesn't say ANYTHING as to what is the algorithm and why it's faster and how they tested is (I think it says there is no test) so I wouldn't put too much faith in what they are saying. Obviously it's more of a "heart-warming story" than a technical piece

[Sip]

Didn't know that. I'm impressed ... no wonder they pay you all the big bucks!!!! I also heard that those with MBAs get to go out with very good looking blonds all the time. Is that true?

[Garo]

Yes Borland became Inprise Corp for a few months but then changed their name back to Borland.

As I said before their marketing was their weakest place for years and this failed attempt to establish a new corporate identity is an example of that.

[Sip]

These "critics" just piss me off at times. It's so impossible for them to see the potentials of ONE amazing programmer.

 

You put 3 or 4 really trulley talented exceptional programmers together, they could rewrite the entire Windows in a matter of 3 or 4 months, not only making it many times faster, more stable, robust, and reliable, but also probably add tons of cool features and maybe even make it fit on a single floppy disk!!!!!

 

to this kid even though I have NO idea what he has done. Whatever it is, it's probably something good to have grabbed so much attention

 

But as far as "browser" speed, I really don't understand what these people are talking about. I can't remember ever sitting there are going "Gee, my browser seems so slow!" .. it's always been the connection that's been slow and I doubt this kid has done anything to fix that problem

 

On a side note, the Borland compiler he is using was one of the best I have ever used. Too bad Borland doesn't exist anymore ... they used to make such awesome compilers and development tools. But give credit to Microsoft too, their compilers have been pretty amazing too.

[Sip]

As a scary afterthought ... this is from a talk that Dave Tennenhouse (head of Intel research) gave at UCLA a few months ago: ... do you realize that something like car Antilock Break System (ABS) software code that we bet our lives on everyday was most probably written by a C-rate programmer who very likely was not the smartest (or even maybe in the top half) in his or her class?

 

Yah, that is SCARY.