How To Subscribe and Unsubscribe is at the end of this note.
Mailing List Trouble? See http://www.langa.com/help.txt
Want an easier-to read formatted HTML version? See http://www.langa.com/whats_new.htm

Please email the LangaList to a friend! (Use this super-fast form!)

The LangaList

In This Issue:
BT2K is Smokin'!
Early Test Results
Browser Weirdness
Special Tip#3 On Cookie Coughing!
Open Source Split?
The More the Merrier
Neat Net Tricks
Bug Report #1
Bug Report #2
Just For Grins
More!

BT2K Is Smokin'!

Wow, what a week!

I expected a few hundred souls to try out the "proof of concept" demo of BrowserTune 2000, but instead, thousands of you have tried it out!

And, with all the great feedback I've been getting, I've been able to make a number of adjustments and improvements already: So far, I've posted three complete new point-revisions of BT2K, so the BT2K demo running today is quite different from the one posted just a week ago!

For example: the throughput timing tests are based on how long it takes for your browser to receive a special "pseudo-random data block:" This is a long string of data churned out by a random number generator, converted into ASCII text, and then scrubbed of any characters that might be interpreted as HTML or JavaScript. What results is a block of totally inert, random characters. They're random for a reason:

Some web servers and virtually all modern modems offer data compression, which looks for repetition in the data they're handling. When they find repeated characters, they substitute a "token" for the repeated string, which speeds up transmission.

By analogy, it's faster to say "25 A's" than to pronounce "AAAAAAAAAAAAAAAAAAAAAAAAA" one letter at a time. The phrase "25 A's" is a six-byte token in place of the real data, which is the 25-byte string of separate letter A's. In this case, the six-byte token represents a compression ratio of about four to one.

BT2K's random data blocks have few repetitions, and thus resist data compression. This is good, because it forces the server and modem to send almost every character separately. With a known number of characters, it's then a simple matter to time how long it takes to send them and calculate a throughput rate.

But even random numbers repeat: If you flip a coin, for example, you can get several heads in a row, or a string of tails. Similarly, BT2K's random data blocks do have a low residue of repeated characters that can be compressed. These compressible character strings slightly inflate BT2K's calculated throughput speed.

Initially, this caused some minor problems with BT2K's advice: If a 64K ISDN connection showed a throughput of, say, 68Kbps, BT2K would cry foul: With 100% incompressible text, the data rate can't be higher than the full speed of the modem.

But now, with enough test results in hand to gauge the real-world effects, I've been able to adjust the advice BT2K gives so that legitimate test results aren't treated as errors, and the advice is in line with the real-world effects.

There was no way to perform this adjustment without the help of many of you, who sent in your test results. Thank you!

I've made literally dozens of other improvements to BT2K too, refining and debugging it. I hope you'll check it out again soon, and let me know what you think.

Come check out the rapidly-evolving BT2K demo at http://www.browsertune.com/bt2kdemo/ and then drop by the WINDOWS Magazine "Dialog Box" online discussion area (accessible via the front page at http://content.techweb.com/winmag/) to share any unusual results!

return to top of page

Some Early Test Results

Of those early BT2K users who sent test results, most are using IE by a surprising two to one margin over Netscape! About one in three IE users is using IE5, with almost everyone else using IE4.

Most browsers seem to be unaltered, but a few people sent in interesting "agent" strings that suggest their browsers have been heavily modified by an online service or other agency. For example, "Mozilla/4.05 [en]C-AtHome0404 (Win95; U)" looks like a custom version of Netscape for @Home users. "Mozilla/4.5 [en]C-SYMPA (Win95; U)" is another Netscape version; but I have no clue what "Sympa" is.

Some are much clearer: There can be little doubt about the origins of "MSIE 5.0b2; MSN 2.5)" for example. Similarly, we all can guess where this little beauty comes from: "MSIE 4.01; AOL 4.0; AOL 4.0; AOL 4.0)." Subtle, isn't it?

As for speeds, the very worst throughput reported so far is an astonishingly bad one Kbytes/sec from a very unlucky cable modem user; the highest is a whopping 2408Kbps, reported by a different---and much luckier--- cable modem user. Most users seem to be connecting with 56K modems, with 28.8 coming in second; there's also an interestingly large (but minority) number of cable modem and LAN users.

Once again, please check out the rapidly-evolving BT2K demo at http://www.browsertune.com/bt2kdemo/ and then drop by the WINDOWS Magazine "Dialog Box" online discussion area (accessible via the front page at http://content.techweb.com/winmag/) to share any unusual results you get!

return to top of page

Browser Weirdness

I know some assume that if you say anything good about any Microsoft product, you must be in Bill Gates' pocket or some such, but after spending weeks deep in the guts of IE and Navigator, I have to say I think IE is a much better product.

Here's one example: Netscape invented JavaScript, a boon to the entire online world. And Netscape Browsers have always had a much better built-in JavaScript debugger (the "JavaScript Console") than Microsoft's. Netscape's Console will show you exactly where an error occurred and highlight the exact character that caused the problem: you see the actual code, with the error pointed out to you, in red.

Microsoft IE, in contrast, simply tells you a line number and character number where the error occurred. That might be OK, except that Microsoft's own HTML editors, FrontPage, doesn't use line numbers. To debug with IE, you have to use a third-party tool that does count line numbers, and that's a pain in the ASCII.

BUT---and it's a major, major, but---IE is more robust than Navigator in handling JavaScript. Navigator, for example, has had a bug since JavaScript first appeared where complex code cannot be placed within the <TD> and </TD> tags of a table. In IE, Microsoft has worked around this JavaScript bug, and IE has no problem with code of arbitrary complexity inside or outside of tables: It just works.

The Navigator bug is a major problem for BT2K because all of CMP's pages are table-based. This makes it, um, somewhat difficult to produce a complex JavaScript-based tool that will work on both IE and Netscape.

What's more, IE's JavaScript interpreter is much, much faster than Netscape's. I run the two browsers side by side when debugging, and IE is virtually always finishes loading and running the BT2K pages significantly faster than Navigator.

I believe in competition, and really want to support non-Microsoft alternatives, but Netscape makes it hard: Its browser is a pain in the butt to work with.

Of course, that's just my opinion: Your mileage may vary. 8-)

return to top of page

Cookie-Coughing, Take Three

Several weeks ago, I told you how typing

javascript:alert("Cookie is: " + document.cookie)

in your browser's address or location bar will make the browser cough up all the information any site you're visiting has collected in a Cookie about you.

Then, reader Lloyd Folden wrote to tell us of a Registry patch he created that lets you right-click on a page to see the cookies.

Many, many of you wrote in with yet another easy way to view Cookies. For example, Dante.diTommaso@se.statprobe.com wrote to say:

...you can accomplish the main objective simply by creating a bookmark (for any javascript capable browser?) which I'm calling 'Cookie Checker' where the Target URL is simply: javascript:alert("Cookie is: " + document.cookie)

Whatever Web page I'm viewing, I can choose this bookmark from my list, and see a report of all cookies set by the site.

Thanks!

So whether by typing, Registry hacking, or creating a bookmark, you now have three ultra-easy ways to see exactly what information any site has collected about you in a Cookie!

return to top of page

The More the Merrier

They're still pouring in---Hundreds of you continue to recommend this newsletter to your friends via the easy-to-use, 60-second recommendation form at http://www.langa.com/recommend.htm#2.

Please don't stop! 8-) It's great to see the subscriber rolls grow! And there's always room for more readers! If you could take just literally one minute and recommend the LangaList to just one friend, I'd really appreciate it.

Thanks!

return to top of page

Open Source Split?

The Open Source movement sparks deep-felt emotions. Almost a year ago, Netscape announced it was releasing ("opening") its source code, and at the same moment, some Open Source supporters were calling for Microsoft to release the source code for Windows. They pointed to the success of Linux as an example of how openly available source code could improve the quality of software.

At the time, I wrote a column on the topic of Open Sourcing (http://www.cmpnet.com/voices/archive/052798langa.html); although the specific topic then was addressing that call to open Windows' source, the column also asked the more general question, "Is it good business to give away your source code?" It's still a good question today, and it came up on a prominent Open Source forum last week:.

It was a panel discussion among some of the open Source world's luminaries, including Linux Creator Linus Torvalds, Perl scripting language creator Larry Wall, Python programming language creator Guido Van Rossum, and Free Software Foundation founder Richard Stallman. (See http://www.techweb.com/wire/story/TWB19990304S0018) The discussion revealed a deep and widening rift among two camps in the Open Software movement.

The more ardent Open Source advocates maintain that all software should be open and free. Code that's open and commercial doesn't cut it, and neither does software that's proprietary and free. And they regard the last variant---proprietary and commercial---as unmitigated evil, even though it's the model on which our entire software industry is built on.

Linux is the poster child of the Open Source movement: It's moved astonishingly far and fast in no small part due to the fact that it's open and free. But Linus Torvalds (the creator of Linux) surprised some of those attending the panel discussion when he said that he has no particular problem with proprietary code and believes in making money with software. "The work I do that brings in the money and feeds my kids is going to be commercial software that no one is going to play around with," he said. He also urged Linux users to embrace mainstream users and mainstream software companies.

I couldn't agree more, and that was one of the points raised a year ago. Labor-of-love software is all well and good, but only to a point.

In this week's InformationWeek column, I'll flesh out the examples more, and examine the cases in point. I'll tell you why I think the rabid Open Sourcers are naive, and why---if they get their wish---Linux will die.

What's your take? Is open and/or free software the Next Big Thing, or will commercial and/or proprietary software remain the primary way enterprise-class OSes and applications get done? Will open/free software ever become predominant? Will we see more large ISVs opening their codebase? Join in starting Wednesday (10-Mar-99) midday (EST; UT-5) at http://www.informationweek.com/langaletter.

return to top of page

The First Bugs of Spring (part 1)

return to top of page

The First Bugs of Spring (part 2)

return to top of page

Neat Net Tricks

I try to make each Langalist easy to read, with a minimum of jargon and abstruse technical discussion. But I still get mail from some readers who feel I go too fast.

One great way to get up to speed is NEAT NET TRICKS, which the author describes as a "twice-monthly collection of light-hearted not-too-technical computer and internet tips." I subscribe myself, and maybe you'd like it too: It's free, so there's nothing to lose by taking a look.

You can subscribe free with email to majordomo@neatnettricks.com and state SUBSCRIBE NEATNETTRICKS . Or, visit the web site at http://www.neatnettricks.com. Tell 'em Fred sent you. 8-)

return to top of page

Just for Grins: Mergers

Gagler's Joke-of-the-Day service (signup@gagler.com) sends these along:

Here are  the latest mergers we can expect to see:

Hale Business Systems, Mary Kay Cosmetics, Fuller Brush, and W.R. Grace Company merge to become Hale Mary Fuller Grace.

Polygram Records, Warner Brothers, and Keebler Crackers merge to become Polly-Warner-Cracker.

3M and Goodyear merge to become MMMGood.

John Deere and Abitibi-Price merge to become Deere Abi.

Zippo Manufacturing, Audi Motors, Dofasco, and Dakota Mining merge to become Zip Audi Do Da.

Honeywell, Imasco, and Home Oil merge to become Honey I'm Home.

Denison Mines, and Alliance and Metal Mining merge to become Mine, All Mine.

Knott's Berry Farm and the National Organization for Women merge to become Knott NOW.

return to top of page

See you next issue!

Best,
Fred
(fred@langa.com)

(P.S. Please email the LangaList to a friend! Use this super-fast form!)

An easier-to read formatted HTML version is available in the "what's new" section of http://www.langa.com. All past LangaList issues are also available via the same link.

return to top of page

SUBSCRIBE (it's free!): Send email to subscribe-langalist@lyris.dundee.net
UNSUBSCRIBE: Send email to unsubscribe-langalist@lyris.dundee.net
LIST TROUBLE? HAVE QUESTIONS? NEED HELP? See http://www.langa.com/help.txt

This newsletter is a free service of Langa Consulting LLC and is Copyright © 1999 Langa Consulting LLC. All rights reserved.

return to top of page