Previous Issue | Search NetBITS | NetBITS Home Page | Next Issue
Thinking about burning up the Internet with a blazing-fast 56 Kbps modem? In this issue, NetBITS Managing Editor Jeff Carlson looks at how 56K modems work, and why the competing K56Flex and X2 standards may be burning users instead. Also, we test real-world spam blocking, talk about image compression methods, lament the absence of Internet Gophers, and catch a wave with the originator of the idea of Net surfing.
Contents:
Copyright 1997 TidBITS Electronic Publishing. All rights reserved. To subscribe to our weekly list, email <netbits-on@netbits.net>. Thanks to our sponsors for their financial support of NetBITS.
You're on Report -- As we've noted in previous articles about spam, one method of fighting it is to report the abuse to the network on which the spammer has an account. This can sometimes be tricky to find out, but the top two Received: header lines in email are usually a good way to learn where the message originated and from where the sender had connected - these can't be forged. (In some mail programs, the headers may be out of order or reversed; in that case, look for the Received: header line that shows the email being received by your provider or company's mail server.)
I did an experiment one Sunday afternoon and reported 100 spammers back to their networks, in most cases UUNET, Netcom, PSI, and MCI. By the next week, I had received messages that most of these users' accounts had been cancelled, or their downstream ISP was informed of the spam. Several smaller ISPs whose mail servers had been hijacked had already installed relay safeguards to prevent recurrence. Sure, spammers can just open another account somewhere else or find another mailer to hijack, but there's some satisfaction in having made their lives a little more difficult. [GF]
Image Ownership -- Jamie Alexander <jamie_alexander@ccmgate.ihsenv.com> asked me why I didn't mention compression algorithm patent and licensing issues in the graphic file formats article in NetBITS-007. I didn't because in this case intellectual property ownership doesn't result in costs to those who create and view images. The costs come when you buy (or develop) software that reads or writes formats which have licensing deals behind them.
To clarify, GIF (and the popular prepress format TIFF, which has no bearing on the Internet) uses the LZW, or Lempel-Ziv-Welch, compression algorithm; Unisys owns the underlying patents for implementations of LZW compression. This algorithm involves finding patterns (called phrases) throughout a file and replacing them with tokens that correspond to the patterns as entries in a table. For example, if you had a text file with the word "hellzapoppin" in it 50 times, the word might be replaced with a one-byte value of 137. In the phrase table, an entry matches the value 137 with "hellzapoppin."
All the major image editing developers, like Adobe, Corel, Macromedia, and others, pay licensing fees to Unisys to use LZW with TIFF, GIF, and other formats; Adobe also uses LZW in PostScript. This issue rose to consciousness because even though Unisys had patented LZW in the late 1980s, they didn't bother to discover its use in the GIF format until several years after it was in wide use. (They had already made deals to license it for other file formats, like TIFF.)
PNG uses the zlib library, which is based on very similar ideas - using variations on an older set of Lempel-Ziv algorithms - but is implemented in such a way as to avoid licensing fees or patent violation. Zlib is the basis of the Unix gzip compression program, and it's available as a choice (ZIP) in Adobe Acrobat Distiller for compressing black-and-white bitmaps.
JPEG uses Discrete Cosine Transformation (DCT) and Huffman Encoding as two of the five steps it goes through to compress image data, and as far as my reading and understanding goes, these two algorithms aren't patented.
The bottom line: any commercial software you buy has to license LZW anyway, in order to deal with multiple file formats besides GIF. The author of every shareware package you use that displays or creates GIFs pays a minimal licensing fee. Although some people have advocated abandoning GIF because Unisys charges fees for a previously "free" format, don't avoid GIF just because someone owns part of its soul. Information may want to be free, and you can help contribute to that cause; but as an author I know that creativity must be rewarded even if a corporation is the one that ultimately cashes that check. [GF]
<http://www.unisys.com/LeadStory/lzwfaq.html>
by Jeff Carlson <jeffc@netbits.net>
When jet test pilots first sought to break the sound barrier, they did so without knowing exactly what was on the other side. Some believed the goal was impossible, that mankind had reached its speed limit. The more superstitious pilots believed that dragons waited at 760 miles per hour (Mach 1) to gobble up the men and their little flying machines. And yet, the pilots all continued to push the speed limits until the barrier broke - and then they tried to fly even faster.
Today, modem manufacturers are emulating those test pilots of legend, pushing against the speed limits imposed by analog telephone line connections. Presented with a new class of modems capable of achieving speeds approaching 56,000 bits per second (56 Kbps), it's tempting to ditch our pokey old 28.8 Kbps and 33.6 Kbps paperweights and buy the faster models. In this case, however, there really are dragons waiting on the other side - for now, anyway.
[Note that I'm going to use the term "speed" here even though it's not accurate. Modem "speeds" are actually "throughputs" because speed is a measure of distance per time, whereas throughput is a measure of moving some unit (bits, in this case) per time.]
The Good News -- Most of the marketing material currently being pushed by modem manufacturers is accurate - at face value. Speeds of 56 Kbps are possible over existing telephone lines without the addition of extra hardware other than a new modem supporting 56K technology. In fact, some of the current 33.6 Kbps modems on the market today can even be upgraded to the 56K speed by a process known as a "flash upgrade": you transfer new firmware - the modem's internal operating instructions that control how it works - from your computer to your modem's permanent memory chip (sometimes referred to as EEPROM - Electrically Erasable Programmable Read-Only Memory).
Although your computer may claim a 28.8 or 33.6 Kbps modem connects at speeds of 57.6 Kbps or 115.2 Kbps, these speeds are just how fast the port between your computer and the modem is running, rather than how fast the connection between your modem and a remote modem is running. On the other hand, 56K modems are actually capable of sending information at the higher speed. This is due both to new modem technology and to the digital connections that link most Internet service providers' (ISPs) offices to the switching stations at local phone companies where calls are routed.
Flying Faster -- Here's how the data transfer works. An ISP's server relays a packet of information from the Internet over the ISP's internal network to the modem that you're connected to via a phone line. Up until that final stage, the information is digital; when it hits the modem on the ISP side, it's converted from digital signals (used by the ISP's systems) to analog frequencies (the noise you hear if you pick up the line while you're connected to the Internet).
The information, now in analog form, goes from the ISP's modem to the switch at the phone company's central office (or CO) to which the ISP's phone lines are connected. (There can be several or even dozens of CO's in a major city.) When the data hits the switch, it is converted from analog back to digital for its journey through the telephone company's network (officially known as the public switched telephone network, or PSTN), ending up at the switch connecting to your phone line. The data undergoes another digital-to-analog conversion to travel to your modem, which then takes the analog signals and converts them once again into digital signals that your computer receives. The numerous digital-to-analog conversions result in a significant amount of "noise," slowing the network speed to 35 Kbps (a figure reached by a mathematical formula known as Shannon's Law).
With 56K technology, the digital-to-analog conversions that aren't necessary are taken out of the picture. ISPs can connect themselves to the PSTN digitally, so information only needs to undergo a digital-to-analog conversion when it leaves the PSTN en route to your modem. This streamlining of the process reduces noise and increases capacity of the connection, since the signal stays digital from ISP to PSTN.
If you'd like to blow a few million neurons trying to understand the technology behind regular modems (and the Rockwell version of 56K), you can read a white paper at Rockwell's site.
<http://www.rss.rockwell.com/K56flex/whitepapers/56k_on_PSTN.html>
The Dragons -- So, you might ask, why don't we all have 56K modems on our desks right now? Unfortunately, there are several barriers blocking the wide-scale implementation of 56K technology.
One of the key problems is standardization. As yet, there is no generally accepted standard for 56K communications, and a ruling by the International Telecommunications Union (ITU) isn't expected until at least September 1998. In the meantime, several modem developers, including 3Com (recent purchaser of U.S. Robotics), Lucent Technologies, and Rockwell International, have emerged with functional, but incompatible, technologies in the hope of forcing standards by way of public and industry acceptance. Until one set of standards prevails, users buying either technology run the risk of ending up with the modem equivalent of the varying gauges of railroad tracks through history.
<http://www.nb.rockwell.com/platforms/network_access/>
<http://www.lucent.com/internet/reprint2/>
<http://x2.usr.com/>
<http://www.itu.int/PPI/press/releases/1997/np-09.html>
Another problem concerns ISPs. Already burdened with having to support multiple modem technologies, ISPs will need to invest heavily in upgrading their systems to support 56K devices. To make matters worse, they won't be able to just buy a rack of 56K modems and put them on the network. Instead, they must install "central site" modems that are digitally connected to the PSTN. (This also means that if two users buy 56K modems, they usually won't be able to create a direct connection between them at top speed.)
Even if a standard is reached, and ISPs invest in central site modems to serve their 56K customers, another problem remains: those users will only be able to connect at a top speed of 53 Kbps. Pushing a full 56 Kbps through existing telephone wires requires a higher voltage than the United States Federal Communications Commission (FCC) allows. Until the FCC increases the voltage limit, or grants waivers for 56K communications, adopters of the new technology will be limited to the slower 53 Kbps speed, which is itself not always achievable. Tests done by computer magazines show a 40 to 50 Kbps range is more typical.
And, of course, there are lawsuits. In 1995, a Stanford University associate professor named Brent Townshend presented Rockwell Semiconductor with his ideas for a technology which would enable 56 Kbps transmissions. Allegedly, Rockwell assured him that they would obtain his permission if they were to use or disclose his ideas. Rockwell then developed their K56Flex chipset independently, so Townshend filed suit in October 1997 asking for compensation and an injunction against future use of K56Flex.
<http://www.wired.com/news/news/technology/story/7915.html>
While courts debate over who created the technology, they will also be looking at whether or not modem manufacturers have deceived consumers by inflating claims about the current 56K modems. In May 1997, a California lawyer filed a consumer-action lawsuit against several companies on the grounds that their advertising is promising speeds and features that the existing hardware cannot deliver. He is seeking injunctions on the alleged false advertising, and unspecified compensation for people who have bought 56K modems.
<http://www.wired.com/news/news/technology/story/3962.html>
Oh, and did I mention that the 56K speeds only apply to downloading information from ISPs, not uploading it? The ISP's central site modem can take advantage of network synchronization and a few more obscure features that allow it transmit at up to 53 Kbps; the user-to-ISP direction can't take advantage of those extra cues and is limited to the maximum analog speed of 33.6 Kbps.
Applying the Brakes -- Does this mean we're all relegated to 28.8 Kbps and 33.6 Kbps speeds? Does this mean that to cruise the Internet quickly, will you need to invest money and time in technologies like ISDN, frame-relay, ADSL, cable modems, or fractional T1 lines?
Nope. If you desperately need higher speeds now, by all means explore those other options. If you've been using the Internet with a 14.4 Kbps or 28.8 Kbps modem and are already looking to upgrade, spending a couple of hundred dollars on a 56K modem will probably pay off and allow you to work at 33.6 Kbps. Many manufacturers are guaranteeing that their products will allow a flash upgrade (generally at no cost) to whatever the final 56K standard turns out to be - even if it's a combination of the two current technologies. In the meantime, be sure to call the friendly folks at your ISP and ask them which direction they've taken. (Some ISPs have banks of both modems to avoid alienating their most eager customers.)
If it's not killing you to hold steady at 28.8 Kbps or 33.6 Kbps (the speed at which I currently connect), you'll probably want to sit back and wait for the dust to settle. Given the current demand for faster Internet access, it's inevitable that the 56K barrier will eventually evaporate, leaving us to look ahead to speedier connections.
[Jeff Carlson is the managing editor of NetBITS and TidBITS. This article is adapted and updated from "Speed Jockeys on the Internet," which originally appeared under his byline in the 15-Mar-97 issue of adobe.mag. Adapted with permission from Adobe Systems, Inc.]
Question: What's surfing got to do with it? Andree Hollander <andree_hollander@spidernet.nl> asks, "People continually talk about 'surfing the Internet.' Where does the word 'surf' come from? I didn't think that the word 'net' would be something with which the word 'surfing' would be connected."
We happen to know the person who's credited with the first use of the term "surfing the Internet" - Jean Armour Polly <mom@netmom.com>, author of The Internet Kids & Family Yellow Pages, 2nd Edition, and publisher of Net-mom News, a weekly 2,000 word email newsletter that focuses on issues surrounding kids on the Internet. For sample issues and free subscriptions, check it out at the URL below.
Rather than answer this question ourselves, we asked Jean to tell us the story.
Answer: Yes, Nexis credits me with the first published use of the term surfing in relation to the Internet, back in 1992. [Nexis-Lexis is a popular pay-per-use newspaper and magazine searchable archive, for which journalists have a fondness. -Glenn] You can read that original article and hear the tale, entitled Birth of a Metaphor on my Web site.
<http://www.netmom.com/about/surfing_main.htm>
This is has become a cause celebre in surfing circles, who equate the phrase with mindless "channel surfing." I certainly intended no harm; in fact I meant to compare navigating the Internet respectfully to a sport that takes great skill and ability. Back in the early 1990s, you needed great skill if you wanted to use the Internet for anything besides email.
Remember that in 1991 and 1992, the Internet was not as we know it today. It was much harder to use, there were no search engines or Web catalogs as we have now, and you had to memorize lots of arcane commands. It was an art, not a science. Today we navigate the oceans using the Global Positioning System (GPS); in those days navigating the Net was more like ancient Polynesian wayfinding: memorizing star pairs at each horizon, reading patterns of phosphorescence in the waves, and knowing the habits of pelagic birds.
All this is in the article referenced above, but here's a brief excerpt: "In casting about for a title for the article , I weighed many possible metaphors. I wanted something that expressed the fun I had using the Internet, as well as hit on the skill, and yes, endurance necessary to use it well. I also needed something that would evoke a sense of randomness, chaos, and even danger. I wanted something fishy, net-like, nautical.
"At that time I was using a mousepad from the recently defunct Apple Library in Cupertino, CA, famous for inventing and appropriating pithy sayings and printing them on sportswear and mousepads (e.g. 'A month in the Lab can save you an hour in the Library.') The one I had pictured a surfer on a big wave. "Information Surfer" it said. 'Eureka,' I said, and had my metaphor."
I should mention that the closest I have come to surfing is as follows: I've wandered into a surf shop once, I have watched people surf on TV, and I paid homage to the Duke (considered the Father of modern surfing) statue on Waikiki the last time I was there.
<http://www.surfart.com/water1.htm>
<http://planet-hawaii.com/duke/>
Question: Where have all the Gophers gone? Rich Schiek <rich_ls@pacbell.net> asks: "Several years ago when the Web was just starting there were some other technologies for sharing text-based data over the Internet. The two I recall are Gopher and WAIS, which could be used to search for weather information, university library catalogs, etc. Have these tools gone the way of the TTY terminal? Everyone has a Web page but does anyone still have a Gopher hole?"
Answer: Ah, Rich, thanks for reminding us of the good old days before the Web, when, to find much of anything on the Internet, we had to telnet barefoot to it in the snow, pushing our packets uphill all the way. Yes, Gopher and WAIS (which stands for Wide Area Information Servers) were alternate methods of publishing and searching for information on the Internet before the Web. Gopher was developed by the University of Minnesota - their mascot is a gopher - and pioneered the concept of linking sites together, since an item in a Gopher menu could be either a local file or a link to a file on another Gopher server. In fact, Gopher could handle data types other than text, but only as items in a menu, in contrast to the way graphics or movies can mix with text on Web pages.
WAIS was somewhat different. Developed by Brewster Kahle while at the supercomputer company Thinking Machines, WAIS servers were basically massive databases of text. WAIS servers provided English-like search facilities and presented the results with ranking (how likely the result was to answer your question) and relevance feedback (which enables you to say "Find me more like this one"), both of which were cutting edge concepts at the time.
What's happened to these services? Although it's still available, Gopher isn't used much any more, especially for new services, although most Web browsers work with Gopher servers. WAIS was spun out of Thinking Machines into a company called WAIS, Inc., and was then purchased by America Online. Since then the only place you might find a WAIS server is as the back end of search engines accessed via Web sites. If you're interested in learning more about these essentially obsolete services, you might check out what I wrote about them back in 1995 in Chapter 9 of the third edition of my Internet Starter Kit for Macintosh, which is old but still online in its entirety. [ACE]
<http://www.tidbits.com/tidbits/iskm.html>
[Please send us any and all questions surrounding the Web, email, or the best way to make brownies at <faqtoids@netbits.net>, and include your full name and email address. Questions may be edited for content and length. We cannot guarantee publication or a reply.]
Move Me -- Martyn Wilkinson <mgw@picdar.co.uk> writes in regards to the graphic file formats article in NetBITS-007:
There's one point Glenn forgot to mention: GIF files support multiple "frames," which means they can be used for simple animation. I believe this feature is exclusive to GIF since, until Netscape supported it, no one had bothered with this rather obscure part of the GIF specification. PNG was designed before Netscape supported multi-frame GIFs, and the designers didn't see the point of supporting this completely unused feature. Pity.
Myths of the Net -- Ken Lager <klager@pipeline.com> points us to a good resource for debunking stories like the AOL 4.0 rumor we lambasted last issue:
I have found that the best way to deal with mail warning me of a deadly computer virus is to fire up my Web browser and head on over to the Computer Virus Myths page at Rob Rosenberger's non-profit Web site. It always has the latest information to debunk all of these urban legends. Usually, I write up a reply to the sender and everyone who the message was sent to, including the information from the site as well as the address. That way, I won't have to worry about getting that same message again from mutual friends who are looking out for me.
<http://www.kumite.com/myths/home.htm>
[Please send letters to the editor to us at <letters@netbits.net>. Please include your full name and email address. Letters may be edited for content, grammar, and length. All letters become the property of TidBITS Electronic Publishing. We cannot guarantee publication or a reply.]
Non-profit, non-commercial publications and Web sites may reprint or link to articles if full credit is given. Others please contact us. We do not guarantee accuracy of articles. Caveat lector. Publication, product, and company names may be registered trademarks of their companies. NetBITS ISSN 1096-4908.
Previous Issue | Search NetBITS | NetBITS Home Page | Next Issue