[LWN Logo]
[LWN.net]

Bringing you the latest news from the Linux World.
Dedicated to keeping Linux users up-to-date, with concise news for all interests


Sections:
 Main page
 Security
 Kernel
 Distributions
 Development
 Commerce
 Linux in the news
 Announcements
 Back page
All in one big page

Other LWN stuff:
 Daily Updates
 Calendar
 Linux Stocks Page
 Book reviews
 Penguin Gallery

 Archives/search
 Use LWN headlines
 Advertise here
 Contact us

Recent features:
- RMS Interview
- 2001 Timeline
- O'Reilly Open Source Conference
- OLS 2001
- GaŽl Duval
- Kernel Summit
- Singapore Linux Conference
- djbdns

Here is the permanent site for this page.

See also: last week's LWN.

Leading items and editorials


BT attempts to enforce its patent on linking. It must really be true that British Telecom is a foe of software patents. What else could explain its behavior, which is clearly intended to demonstrate just how ridiculous and destructive those patents are?

The company, remember, claims that an old patent (from 1976) covers the act of linking between pages on the web. As a result, says BT, anybody who makes such links without licensing the patent is guilty of infringement. There are, of course, a few people and companies here and there which are guilty of this nefarious deed, meaning that BT should have no shortage of possible legal targets.

One wonders, then, how Prodigy was chosen for the honor of being the first victim of BT's legal team? (That team, incidentally, is Kenyan & Kenyan, the same company that sent out the ":CueCat" letters for Digital Convergence. They seem to be making a determined effort to kick Canter and Siegal out of the Internet doghouse). Some legal eagle has presumably calculated that Prodigy has the proper combination of deep pockets and shallow legal coverage. A good result from the first case would certainly help BT in its task of shaking down the rest of the American Internet industry. BT must be hoping for a quick settlement.

One can only hope that Prodigy stands its ground. BT will have a very hard time winning this case on its merits, for a couple of reasons.

The first reason is that, of course, there is a certain amount of prior art out there. The patent was filed early - 1976 - so many of the technologies we are familiar with now do not apply. But the patent is predated by Ted Nelson's Xanadu work, the pioneering efforts of Doug Englebart, and, of course, Vannevar Bush's amazing As We May Think, which was published in 1945. In the light of that work, BT's patent looks decidedly unoriginal.

There is also the little difficulty of proving that anybody is actually infringing on the patent, even if it is held to be valid. There are a couple of independent problems here:

  • The patent involves the use of a dumb "remote terminal." Modern computers used to access the web do not really fit the description of a "remote terminal" as used in the patent.

  • Patents written in a "client/server" mode, as it turns out, are hard to enforce. An ISP like Prodigy is really only supplying the server side of the mechanism described in the patent; the client ("remote terminal") side is provided by the user. It is difficult to say who, if anybody, is really infringing on the patent, since nobody is implementing the full system described in the claims.

(The above is lifted from Greg Aharonian's PATNEWS newsletter, which covers the topic in detail. See the Internet Patent News Service page for more information on PATNEWS).

So BT is not going to have a easy time prosecuting this patent. With luck, however, the company will, in failure, succeed in convincing a wider group of people that software patents are a bad idea.

(See also: Don Marti's open letter to the CEO of Prodigy, Rick Collette's petition to British Telecom asking the company to back off, and news articles in The Register and News.com).

The LWN.net 2000 annual timeline. Our long-time readers will know that we have made a bit of a tradition out of our annual, year-end timeline. We're happy to announce that the initial version of our LWN.net 2000 timeline is now available. Check it out for a summary of the major events in the Linux community over the last year. As always, it has been an interesting ride.

For the curious, here are the timelines for 1999 and 1998.

When Linux companies go bad. What happened to Corel? Earlier this week C|Net's News.com reported that Corel was considering a sale of its Linux business to a venture capital group known as Linux Global Partners (LGP). While C|Net reported possible transactions, the National Post was a bit more firm: "Software maker Corel Corp. has agreed to sell its Linux line of products to a New York-based venture capital firm in a transaction that will close in January, sources close to the deal said [on December 15th]." The news wasn't the first report of Corel's fall from the Linux hierarchy - see ZDNet's report from November - and it probably won't be the last. With all the speculation circulating, what does Corel have to say? "We have no further comment on this issue."

Assuming the sale is a done deal, what does this say about Corel and, more importantly, about Linux distributors? Is this a trend for commercial Linux distributors? Hardly. Red Hat is more robust now than ever, having handily beaten analysts' 3rd quarter earnings estimates. And TurboLinux had a stellar year, garnering over $80 million in investments.

So the problem is essentially limited to Corel. Why? What happened? By May of 1999, Corel's WordPerfect had generated over a million downloads to beta users. Overall, Corel estimates they have 22 million WordPerfect users worldwide (all platforms). Quite a start to their entrance into the Linux business world. Later, the Debian-based Corel Linux distribution hit the streets to much fanfare. In early 2000, Corel announced their intention to port their popular graphics application Corel DRAW! to Linux, eventually releasing it to beta along with a free version Corel's PHOTO-PAINT. The future looked bright for desktop applications on Linux.

But then things got ugly. Earlier this year, Corel tried to acquire tools vendor Borland Software Corp. to extend its Linux offerings, but the deal fell through after Borland uncovered Corel's poor financial condition and Corel's stock price fell. This past summer, Corel cut 21% of its staff (320 employees). CEO Michael Cowpland resigned in August, starting speculation that the Linux emphasis at Corel might not extend much past Cowpland's reign.

In the November ZDNet article, CEO Derek Burney claimed Corel was refocusing, not relinquishing, its Linux business. Linux now is one of Corel's four business operations - the others are graphics software, business applications and new ventures. However, Corel's Linux operating system, and a Linux version of WordPerfect, generate only about 10 percent of Corel's revenue, which was $36.4 million in the quarter ended August 31. This is hardly the sort of income necessary to maintain the development staff necessary for expanded Linux support and still meet the needs of existing Windows customers.

Corel made several mistakes on the road to profitability with Linux:

  • They wasted time on a Corel branded distribution. They could easily have picked up one of the myriad of other distributions available (see the Distributions page for a long list of possibilities).
  • They required software sales to be their primary form of revenue for Linux products, while all other Linux software companies are mixing software sales with support and other services.
  • They relied too heavily on WINE and didn't consider how that one library could be such a limiting factor to overall product viability. Native ports are less prone to such limitations across an extended product line.
  • They focused on a non-existent desktop market for Linux.
Essentially, they jumped the gun in the desktop race - and they ran in high heels. They looked nice, but it was a painful run.

Corel wasn't completely dead just yet, though. In October, Microsoft invested $135 million in the company to work on .Net services, which Corel initially viewed as not including Linux. According to Burney, who spoke to ZDNet at Comdex: "we didn't understand that an operating system could be constituted as a set of .Net services," Burney explained. This is a realization that Corel has arrived at as it has explored .Net technologies under nondisclosure agreements with Microsoft, he said. In theory, at least, this would mean that a client accessing .Net back-end services wouldn't have to run Windows, Burney said. The investment gave Microsoft a 25% stake in the company.

But that wouldn't be enough to sustain the Linux effort. After Cowpland's departure, it became obvious something had to be done. In an interview with InfoWorld in November, Burney stated "To be successful in the Linux market, you need a wider product offering. There's got to be some kind of acquisition," he said. "It could go either way... there are no sacred cows." The sale was just a matter of time.

The choice of LGP isn't necessarily a bad one. LGP has an investment portfolio of key Linux companies including Helix Code, CodeWeavers, GNU Cash and Metro-Link. The sale price of $5 million in cash is in line with the $6.2 million Corel's Linux business pulled during its current fiscal calendar. The sale might just be a good thing, if LGP manages to find the right management team to run their newly purchased entity.

What is Linux? That may seem like a strange question from a publication like this; it's thus, perhaps, even stranger that it comes from Jon 'maddog' Hall. One would expect that the Executive Director of Linux International and the Director of Linux Evangelism for VA Linux Systems would have an answer...

It starts with an amusing quote from Sun CEO Scott McNealy, quoted in this ZDNet article:

You people just don't get it, do you? All Linux applications run on Solaris, which is our implementation of Linux.

Some people were annoyed by this statement, though most were simply amused. Or even flattered...after all, it's not that long ago that Sun would have taken pretty strong offense at the notion that Solaris is an implementation of Linux.

This is where maddog Hall comes in. In a rare appearance on the linux-kernel list, he points out that nobody has really defined what "Linux" is.

At least, nobody has made that definition in a way that is widely accepted. The Free Software Foundation is happy to define Linux as being just the kernel - the systems we actually use are, instead, "GNU" systems. This claim is, needless to say, controversial. But if Linux is just the kernel, what about RTLinux (a real-time version), mkLinux (a microkernel variant), the S/390 kernel (running on a virtual machine), and so on? The Linux kernel is an amorphous thing.

Increasingly, "Linux" means the binary interface that runs a certain class of applications. Vendors like Sun have been making an effort to implement that API - it was a very quick switch from "Linux has no applications" to "we run Linux applications." Quoting from Mr. Hall:

If it is true that "all Linux applications work on top of Solaris", what standard prevents them from calling Solaris just another implementation of Linux? And should it?

He suggests that the Linux Standard Base may be the right body to set a standard for what is called "Linux." The LSB, perhaps, really needs to deal with the tasks it has now before taking on new ones. But the question is worth asking: what, exactly, is Linux? And correspondingly, what is not Linux? How much do we want to be flexible and where will flexibility lead us astray? Can a proprietary operating system ever be considered "an implementation of Linux", when Linux has always been "Free"? There is a lot to ponder.

Michael Hammel officially joins the LWN Team. Many of you may have already noticed a different name on a response to an enquiry, comment or suggestion sent to LWN in the last couple of months. Michael Hammel actually joined our team in September, much to the joy and relief of the rest of our team members. The volume of news, development and general information about the Linux community has only continued to grow, as we're sure all of you have noticed. Michael's addition to the staff was therefore both a necessity and an opportunity of which we plan to take advantage.

Many of you already know Michael, either through his website, Graphics Muse, his book "The Artists' Guide to the Gimp", his articles for Linux Magazine, Linux Answers and Linux Journal, his talks at a variety of conferences or through his work as Chairman and co-founder of the Colorado Linux Info Quest conference (coming up again March 30th, 2001!). You'll now also meet him within the LWN halls as well, as primary editor of the LWN Daily Page and much more. Please join us in officially welcoming Michael to the LWN team.

Inside this week's Linux Weekly News:

  • Security: /tmp discussed again, a plague of Zope security hotfixes, GnuPG, JPilot and more.
  • Kernel: The Linux Quality Database project; the public nature of linux-kernel.
  • Distributions: Distribution year in review, state of the Woody, Red Hat does IA-64, and Caldera sponsors Samba
  • Development: Debian Jr packages, LyX GUIs, XFree86 4.0.2, SpaceChart 3D
  • Commerce: Embedded Linux training, Red Hat's financial results, Linux is hard at work.
  • Back page: Linux links, this week in Linux history, and letters to the editor
...plus the usual array of reports, updates, and announcements.

This Week's LWN was brought to you by:


December 21, 2000

 

Next: Security

 
Eklektix, Inc. Linux powered! Copyright © 2000 Eklektix, Inc., all rights reserved
Linux ® is a registered trademark of Linus Torvalds