[UKUUG Logo] Copyright © 1995-2004 UKUUG Ltd



Newsletter Section 1


Editor's Column

(Susan Small)

[sue] Welcome to the “Linux Issue” of the newsletter. With this issue you should also receive a copy of Linux World, a new publication devoted to Linux in the UK. Our Linux SIG organiser, Martin Houston, has written articles for this new publication (instead of providing copy for Linux@UK :-) so we hope you enjoy reading your complimentary copy of this new publication. If you have any comments about Linux World, please send them to me and I will pass them on to Martin.

Thinking ahead to the 21st birthday celebrations for the UKUUG (see the Report from the Chair), it would be nice to have one issue of the newsletter offering contributions about the past achievements of the UKUUG. Suitable milestones should be remembered, so please let me have any stories or anecdotes that you would like to see included.

I hope that the lively contributions in this issue to the “Linux is coming” debate will resurrect the letters page which has been dormant for some time! Send your ripostes or supports to the arguments to me for publication in the next newsletter.

Don't forget that you get to keep the book or software if you're willing to review it for us. Send me the details of anything you would like to review and we will do our best to get a copy for you.

Report from the Chair

(Mick Farmer)

[mick] First, our apologies for the late arrival of your newsletter. This was caused mainly by teething troubles with our CD mastering and production systems. However, we've learnt a lot about this and expect to be back on time in future, with a cover CD on each and every issue.

As you probably know, this issue of the newsletter features Linux. We have gathered together a wide range of articles, reviews, and announcements concerning Linux that we hope you'll enjoy. The cover CD includes a complete RedHat version of Linux that will install easily and painlessly on your machine (at least, that's what Martin tells me).

Also on the cover CD is an up-to-date snapshot of the CPAN archives. We intend to provide this regularly on future cover CDS so that you won't have to waste bandwidth, time, and money downloading them from your local archive site. You'll find full details about the contents of the cover CD elsewhere in this issue.

Although our intention is to have the cover CD contain material pertinent to the theme of the accompanying newsletter and snapshots of popular archives, there will be occasions when there's still some room left over. If there's anything that you think we should include on a CD, and this doesn't just have to be software, then please let us know. We will attempt to provide it when space allows.

The theme for the next issue of your newsletter will be UNIX and Network Security. This will coincide with our Winter Conference (and I hope to see many of you in Manchester) and with our next planned workshop. Jim Reid, who is well-known to many of you, will be running a three-day UNIX and Network Security workshop in London in early February. Full details will be sent out as soon as the arrangements have been finalised. If there is any security-related software that you think we should include on our cover CD, please let us know.

While I'm still on the subject of your newsletter, what themes would you like to see in future editions? There are lots of interesting developments going on. What about the WWW? What about thick/thin clients? What about mobile computing and teleworking? Just let us know!

So far, the number of you that have taken up our offer of a free “mailbox for life” and a free web page has been disappointing. I can't believe that you've all got connections to inexpensive ISPs providing such services. Even if you have, why not take advantage of having another link to your site? As someone once said in a well-known advert, “you know it makes sense”.

Finally, we think that the UKUUG is 21 years old in 1997 (the dates of the first few meetings leave room for minor doubts). We intend to celebrate this occasion with a special meeting, probably in the summer. We hope to gather together ex-luminaries of the UKUUG and persuade some of them to talk about the past and the future of UNIX and Open Systems. Therefore if you were involved, or know of any people who were involved, in the early days of the UKUUG, please get in touch with us.

Why Linux will win

(Martin Houston )

The runaway success and explosive growth of the Internet has taken many people outside of the Linux community by surprise.

One would have thought that Microsoft with all its money to employ expert pundits and computer know-how to model the future would have seen it – but they didn't! Bill Gates assumed that he could distract the world's eyes away from the open horizon of the Internet into the nice cosy, making more money for Microsoft, world of "The Microsoft Network".

Nevertheless Microsoft was forced to make a total about turn and now embraces the Internet as something central to the business goals that they had all along. It remains to be seen how committed Microsoft is to the spirit of the Internet, rather than just the technology. With the current moves to foist ActiveX and the custom extensions in Internet Explorer on the world, the answer is probably not!

Taking on board wellproven, standard technology is easy – but as binary data of arbitrary complexity can be transmitted over the Internet, it is just as easy to reduce the Internet to a mere transport mechanism between proprietary components, that are inoperable with anything other than a

company's own products. There can be all sorts of praiseworthy sounding reasons for this lock-in, such as data encryption and seamless integration, but lock-in is what it is!

The spirit of the Internet is the total opposite of this. The Internet started life as a project of the US military – to build a command and control system that was distributed and would therefore survive an atomic war. Luckily for us the technology involved was not developed and kept as top secret. Internet development from the early ARPANET days onwards has been a highly collaborative act. There is no one “Internet Inc.” from which you have to purchase licences to use TCP/IP and all the other software needed to make the Internet work. Indeed, if all the billions of man hours that must have been accumulated on development and polishing of Internet software were fully costed, then nobody would be able to afford a licence anyway!

A statement like this has the potential to confuse a lot of people. They ask “Why do I have to pay for my Internet Access Software then?”. The trite answer would be “If you run Linux, then you don't” but it brings up the big issue of how software that is just the implementation of public standards should be costed. Just because developers are free to cooperate on things does not remove all costs. Programmers still have to eat! Doing things “right” by following standards involves more work in some directions than the tempted “quick lash up – compatible with nothing”.

A packaged and documented collection of material, which is itself essentially free, does not automatically inherit this freeness. The act of packaging, documenting and supporting adds value that can be charged for. PC Internet access packages almost always have various “luxury” touches, such as a GUI in the ping program, to differentiate them from standard Internet code.

The real benefit of the Freedom is that the Internet has been free to develop and mutate in a multitude of ways that would never have been possible with a commercial software regime. There is no artificial throttle of having to observe and work with somebody else's proprietary, and sometimes not well, or even willingly, documented technology.

One of the net mutations is of course Linux. If it were not for the Internet, and the comp.os.minix Usenet group in particular, Linux would have remained Linus' student project and gone no further. Linux is of course an independent entity – you don't need an Internet connection to run a packaged Linux distribution, but the usefulness of Linux will be somewhat curtailed as a result.

Although it is possibly the single biggest feat of freely given human cooperation ever seen... [What about BSD? Ed] there is nothing unique about the Linux kernel – it is just one of thousands of other software projects that the Internet has nurtured from tiny embryos to being major parts of our software landscape.

The one thing that the vast majority of this great software has in common, apart from evolving on and from the Internet, is that it runs on Linux, and a lot of it is even part of some Linux distribution or archive.

Running a Linux machine really opens up the Internet in a way that dabblers with their Windows PCs cannot even dream of!

Microsoft thinks its future is on the Internet; sorry to disappoint you Bill, but the Internet already has its own OS that is growing in synergy with the net itself. Microsoft will always be an awkward and ill-at-ease visitor.

Linux, with its absolute openness, is the totally natural choice of OS for Internet connected computers. You don't pay somebody a licence to use the massive collaborative human achievement that is the Internet, so why pay somebody a licence just for the right to use your computer Internet operating system?

I am in wholehearted agreement with Microsoft that support for the net belongs in the OS (although I think their sentiment was aimed at the “do everything” Netscape). Where I depart from them, is that we would be far better off embracing the net as a whole new way of using computers, rather than just hide it behind the old familiar facade.

The things that people are able to do with the Web are sending shockwaves through the traditional software industry. Many things that used to be hard and expensive are now easy, standard and cheap to do. Look at how much has been accomplished on the Web in just a couple of years! And this is with Internet bandwidth being in short supply. What will we be able to achieve a few years down the line with our gigabit-connected world network of Linux boxes?

In his book Accidental Empires, Robert X Crigley observes that the cost of developing any fundamentally better software than we have at the moment, will require more resources than any one corporation can afford. The amount of time and talent being put into Linux by some of the world's best and most productive programmers means that Linux development is already more lavish than anyone could afford to pay for.

Microsoft has already had to make one giant U-turn over the Internet. I eagerly await the arrival of “Microsoft Office for Linux” – you don't think Microsoft is going to miss out on the limitless possibilities of a global Open Systems marketplace out of pique do you?

Linux - On the Internet, no- one can hear you scream!

(The author, who wishes to remain anonymous for obvious reasons, is happy to reply to any letters that you send to the editor in response to this article.)

If you want another article extolling the virtues of Linux and telling the world what an awfully nice chap Linus is, don't read on. Because here we replace Linux Fantasy with reality. Here you will see how the GPL and the Free Software movement is stifling investment and growth. Why Linux is dying on its feet and why there has been little or no uptake over the last two years. Got your attention? So let's go!

Following the release of RedHat 4.0, (the perfect marketing announcement, short on facts and long on hype) I got talking to an old friend about the demise of Slackware Professional. A story which the kids at RedHat would do well to remember. Slackware Professional was produced by a (now defunct) outfit called Morse Telecommunication. Yggdrasil pioneered Linux CDs, and Morse provided the commercial world with the first real Linux product. Presented in a quality box, the clever marketing hid the fact that it was little different to the freely available Internet Slackware. Just like RedHat in fact: boxed RedHat having little over what you can download. Back to Slackware Pro; it was a landmark product and proved to a sceptical

reseller base that you could indeed make money from Linux. And it sold very, very well. But as with all things Linux, the GPL and religious fervour were soon to make themselves felt. The soon-to-be-familiar support nightmare started for anyone selling the product. It was Linux, essentially untested (yes, untested - more on this later) freeware. If the product did not work, the customer would return it (often abusively) and demand their money back. Companies would be berated and abused on the Internet for not doing things in the way that the customer expected. For these were the heady days, when Linux would change the world and commercial enterprise, while not precluded from the GPL, was definitely outlawed by many on the Internet. Those reselling Slackware Pro also found that the margins were being eroded and there was little they could do about it. But in the grip of Linux zeal, nobody was looking at reality.

Slackware Professional went from 2.0 to 2.1 over about 6 months. 2.2 never made it and 2.3 never had the same success. A landmark Linux product faded away as others took over; Linux Developer Resource, Slackware from Walnut Creek, Linux-FT, RedHat. Suddenly there were more Linux releases, more widely available and, what appeared to be, an insatiable market. And the book publishers realised that there was an untapped market. With a stampede more fitting to the Klondyke, Gold (sorry Linux) Fever took hold and almost everyone had a book on Linux. Now, in the fairly-land world of Linux distribution, market share is everything. Success is measured by how many use your distribution, so there was a small (and pointless) push to get the Book Publishers to use various distributions. They mostly used Slackware; it was the most widely used and free. The bookstores were soon full of Linux books with CDs in the back; it all added downward price pressure, increased availability and decreased the cashflow for everyone selling Linux.

By mid 1995, the Linux rate-of-change was decreasing. Infomagic, formerly having eight weekly releases of the immensely successful LDR, were slowing down the release schedules. Linux quality was, from a commercial perspective, taking a nose-dive and margins were getting smaller. The Internet Effect was also beginning to make its mark. The powerhouse behind commercial distributions was slowing down.

The Internet Effect – Or, why the Linux newsgroups are actually worse than useless

The Internet has been a fundamental tool in Linux's development. Nobody could possibly deny its benefits. Commercial postings have always been criticised and the Commercial Linux Vendors have a wide variety of tricks to get around this “problem”. After all, we would not want to upset the delicate sensibilities of News Readers, or waste bandwidth that could be better spent debating Linux Logo's (or other essential matters). The most popular technique is to post some contrived piece of information (no matter how useless) and “helpfully” add your price list. Simple. It's not such a problem these days, the Linux Internet groups have matured slightly, although you will still find signs of it.

The Linux newsgroups have been, and are likely to remain, a constant source of amusement. The groups (and the various Linux mailing lists) provide a forum where the stupid and knowledgeable intermix. Where sad people can feel important, debating non-issues with expertise they don't possess. Where it is claimed that Linux is tested. And in reality where more damage has been done to the embryonic Linux Industry than anywhere else. It was one of the contributing factors to the demise of Slackware Professional. By having a set of forums that appear to be official, any drivel could be posted; incorrect technical information; malicious commercial postings; fraudulent commercial postings; and, very occasionally actually useful information. Let's have some examples (and the names have been changed to keep the lawyers out of work);

Case 1 – Commercial

There was this wonderful posting by a Slackware Professional distributor in the USA which basically said something like “Isn't this neat; a picture of Patrick Volkerding on the new Slackware Pro box”. It was, of course, accompanied by a price list and order form. Hey! This is marketing and any trick is a valid trick and why not play the Internet game. The anti-commercial posting rule (arbitrarily enforced by the masses) has always been a stupid one.

Case 2 – Technical

This is more serious. A well known computer manufacturer wanted to know if their hardware would run Linux. So they posted a question to a Linux news group. After all, the Internet is the famous testing ground for Linux so what better place to ask? They received a few replies, all of which were contradictory. They eventually concluded from the “authoritative” information they received that their hardware would be OK. Wrong! The replies they got were complete crap and it proved to be an expensive mistake.

So just how do you distil the accurate information from the drivel?

The constant torrent of drivel on the Linux news groups has been a major contributor to Linux's commercial failure:

.     People making libellous postings about commercial companies;
.     Almost useless technical support;
.     Overtly anti-commercial postings and threats at the start of every new University term;
.     The regular idiot threatening to make a commercial package available for anonymous FTP.

Until the Linux groups (mailing lists, archive sites etc) are regulated by neutral people, the problem will continue and get worse. We currently have “How To's” being maintained by commercial companies (specifically RedHat for one); are we seriously expected to believe that this is bias free? Do they think we are that stupid?

So, as the number of Linux's increased, the margins dropped and things started to get really bizarre. The Linux Journal would simultaneously print articles about the merits of Free Software, whilst at the same time

bemoaning that so few applications have been ported to Linux. “Free Software will revolutionise the world – Oh, and by the way, please advertise your Commercial product - we need the money.”

But in this Linux fantasy land, nobody wanted to even contemplate that there was no money in Linux and continuing to develop a Linux distribution was a financial no-hoper. So everyone put their faith in Applications. Yes - applications were the answer, at least you could make some money from them. Whilst there is a Commercial Greed factor in Linux, most companies are just happy to scratch a living and invest in Linux development. So, where are the Applications? Wasn't there an organisation doing something in this area? Yes - Linux International.

Let's talk about Linux International (LI). Started to promote a single Linux (which was actually a good idea, but doomed to failure as it required the demise of the existing Linux vendors). Linux International is probably the finest example of how not to do something. The joke for so long was that it was one guy working hard to give the impression that it was actually a larger organisation. And good for him; because it worked for some time. They were also successful in getting people to send them money for their good works - a testament to the appeal of Linux. The trouble with LI was that it had no power and no mandate. Things got worse when for $1000 you could become a Director. Now there was absolutely no chance that they would accomplish anything that was not in the commercial interests of their Directors. Still think Linux is free?

But, people will port applications to Linux. Right? Wrong! Whilst some applications have been ported, the Linux "Community" (sorry, the warring vendors and self-interest groups) have done little or nothing to help. For two years there has been a call for companies to port their products to Linux. So if Linux is so damned successful, where are the applications? There are a few packages ported over by small companies. But we have still to see a major database. The brutal reality is that the Applications Industry sees Linux as an interesting toy, but not something that is worth porting to. Why? Because Linux is so unattractive:

.     Which Linux - with so many to choose, all (for commercial purposes) deliberately different, what do we do? Pity that many elect to wait till things stabilise.

.     There is no money in this. And how true this is. There are too many Internet debates on the cost of Linux and associated products for anyone to hope to make money. And to port an application to Linux the company has to be convinced that they will get a return. More on this below.

.     Linux is unstable, untested and lacks direction and standards.

After two years, none of the above has changed. If anything, things are worse. Whilst there are some very successful applications for Linux, they are very specialised and fit into a niche market. There are no mass market products succeeding. And when a product is ported there is an immediate downward price pressure. A really great example of how not to market an

application is ApplixWare. It's such a worthy example that it has a section all to itself.

But, back to Linux and how much money can be made.

Let's cash in on the Linux stuff! – Or, how to go bankrupt in six months

We are going to sell Linux, we will limit ourselves to two products, the LDR from Infomagic and RedHat Linux (we won't discuss the extra problems of stockholding more distributions). We are purchasing them in bulk (500+ units) from Kaspers in the USA, who are a well-known International Distributor. We will find out the hard way that the actual manufacturers are unwilling to undercut Kaspers price. We also find that we have to pay 5% import duty and import VAT at 17.5%.

  £   £  
Buy Price   7 .16   16 .63  
Import Costs   1 .40   3 .90  
Total   8 .56   20 .43  
RRP   16 .63   !28 .00  
Delivery*   3 .50   4 .00  
VAT   3 .52   5 .6  
Total UK Price   23 .65   37 .60  
Margin   8 .07   7 .57  

*    Delivery is by courier as we find that too many people claim that the product didn't arrive by post.
!    Forced discounted price from USA price cutting and direct sales.

This massive margin is further reduced by manufacturing defects. Whilst we could return them, the return costs are almost always more than the price of the product. Replacing a manufacturing defective product (in RedHat's case) costs £20.43 + £4.00 = £24.43 plus the hassle value. Add to this the unpleasantness of dealing with:

.     Those that deliberately give the wrong credit card number and then get abusive in the hope that they will get a free product.

.     Returns from those that don't “understand” Linux and, finding that the product doesn't work, return it for a refund.

.     Those who consider it fun or otherwise acceptable to cause as much unpleasantness as possible on the Internet because their reseller has “failed” in some way (charged too much, didn't deliver yesterday etc).

and we can reduce the margin by about 50%. So we are making a massive £5 on the sale.

And yet, incredibly as it may seem, there are still those that consider these margins excessive and that they have “paid” for technical support.

What these figures show is that there is no commercial point in selling Linux - and the effects of this reality are hitting home. The simple truth is that it is possible to get Linux almost anywhere. The base price in the USA is 99cents. That's 60p for a CD. The Free Software model taken to its logical conclusion. And very bad for the consumer. Because far from encouraging competition, it drives resellers away from selling Linux or any associated product. This in turn reduces the possible investment, forces up the reliance

on “volunteer effort” (which itself is declining) and makes Linux promotion even harder. Linux availability is declining as companies find that they are actually better off not selling it.

It's just as bad for the manufacturers. The Infomagic set has gone from four to six CDs plus a booklet, but the price is unchanged. Yet we still see people complaining that it is too expensive!

Yes - it's you, the customer that gets the blame for this mess. Every time there is a complaint that the delivery charges are too high, every time some cretin decides to do a mass import from the USA (to save money), every time there is a little price-comparison debate on the Internet, another Linux reseller decides to dump a non-product in a non-market. And exactly what does the customer get from the petty little vendor comparisons on the Internet? You get to save a quid or two. Let's save a quid and kill an industry. “I managed to beat them down and saved £3 off my student grant.” Great achievement. Because the next time you want a copy, the company you “beat down” won't be there. And maybe you won't be able to get Linux locally. Or maybe not at all. To all constant price-cutting cretins out there – you are getting a vast amount of software for very little money. Please don't insult those involved by demanding a free lunch. You are not owed, nor do you deserve one.

Linux Applications

Because Linux costs $25 (Infomagic), applications for Linux should be low cost as well. The Free Lunch brigade give a small concession here. At least they accept that they should pay something, but the tendency is that a Linux application should be much cheaper than its equivalent for other operating systems. Pricing is all rather arbitrary as the Application Company (lets call them ACME) are put in an awkward position. For whatever reason they have decided to port to Linux and now find that they have to charge less for the product. They are usually (and incorrectly) led to believe that the lower price will result in higher sales. But there is a tendency for the Linux price to “downsize” their other products and especially the same product on other platforms. So let's have a real-life example (name changed to protect the innocent). ACME have an application ported to Windows and UNIX. The application is split into components, each costing $150. During the “courtship” phase of convincing them to port to Linux they are (gently) told that the Linux price should be about 50% less than the UNIX price. They are not told why this is, and are certainly not told the reality of Linux margins. Whilst this is not exactly wonderful for ACME, they still proceed. Then the reseller margin news breaks. From the Kaspers example above, the reseller expects a 50% discount. So, we now have a product price of $75 giving a margin to ACME of $30 (Kasper get something). ACME have a stark decision; can we support the Linux port for $30 per copy. ACME check the likely support problems and investigate the Internet Effect. They decide not to port to Linux. It's just not viable. Of those companies that do port to Linux, a great many port as an experiment; a loss leader to promote their other products.

There are some really fabulous products ported to Linux. Applixware and StarOffice are examples. StarOffice is Free. ApplixWare is another example of proprietary software from RedHat. The Kaspers price is $365 with a RRP of $495. So the resellers get a margin of $150. Suddenly the resellers have a reason to sell Linux - they have been given a living.

It's good news then? Well, yes, it would have been, except that RedHat announced an October Special price of $199. A $300 price drop. This master stoke accomplishes the following:

.     It further destroys the reseller channels' confidence in Linux. After all, some of them just risked purchasing at $365 to find that the street price is now $165 lower. They send the product back and abandon Linux.

.     Customers who purchased at the old $495 price send the product back to the (now outraged) resellers with some excuse about it not working and demand their money back. They then purchase direct from RedHat for $199.

.     Having dropped the price by such an extent, it will be impossible to raise it again. Years of experience has proved this.

.     It shows that Applixware is not doing very well for Linux. If it was doing well, the price would not have been dropped. Before Applix get upset, Applixware is a fabulous product - it's just that the Linux pricing has been screwed up.

.     It proves to the reseller channel that the Linux Community (in this instance represented by RedHat) are a bunch of amateurs. Kids having fun. But they just don't know what they are doing.

Whilst this work-of-genius improves RedHat's cash flow, it further destroys Linux's long term viability. Another short-term measure to stop the leaks in the sinking ship.

Linux (aka Titanic) is Sinking - Find another Iceberg

The commercial future for Linux is dire. If it is not enough that we have Linux vendors shooting themselves, and anyone in range, in the foot with insane price offers, that undermine their own resellers, the downward spiral is fuelled by idiotic downward price pressure, mostly by students, and others that consider Linux is a God Given Free Lunch. Is anyone listening? Whilst you are demanding that Linux CDs should be sold at almost cost, you are destroying the market and severely damaging Linux. Free Software and GPL zealots should take serious note that their advocacy is removing investment for Linux. Think this is an exaggeration?

Standards - a Truly Titanic Project

Without going into the details, Linux requires international standardisation. This would at least prove that Linux has been tested. Some

published estimates for getting Linux through UNIX Branding put the cost at about $400,000. A huge investment that is then given away under the GPL. Result? Stagnation as everyone waits for someone else to spend the money and do the work.

This depressing situation has recently been demonstrated in the RedHat 4.0 announcement. When other companies had been investing in pushing Linux through POSIX certification (and releasing the source under the GPL), RedHat did nothing to contribute. Now that the NIST have released the POSIX.1 test suites for free (and they were a mere £4000) RedHat applaud the move and have a copy for “consideration”. The NIST are doubtless very relieved that RedHat are “considering” their tests.

Why wait till it's free? Why not invest in Linux before? It's a good question to ask the vendors (ask them how they test Linux at the same time).

The Truth Is Out There – But they don't want you to discover it

Whilst the official reason for not investing in Standards for Linux has always been that the various bodies charge money for them, there are commercial considerations. Let's consider a non-distribution: BlueHat Linux. In keeping with other Linux distributions, BlueHat don't really do anything to Linux (why bother - they can't control the source and as it's all GPL'ed, their investment is commercially pointless), they invest in pretty installation and configuration systems. Whilst there is a limit to what can be done in this area, they do it anyway. If they were to purchase some standards certification kits, change the sources and give it back under the GPL, they would be condemned to maintain and update this source tree for evermore. They would be criticised (yes it's pathetic, but there are plenty of them on the Internet) for being proprietary and, it would cost money. The most successful Linux companies are those that don't do anything to Linux. Just grab someone else's distribution, plop it onto a CD and sell it. As soon as you start to modify Linux you are wasting money. Far better to let the Internet volunteers do the work and use that instead. So, with the recent POSIX.1 example, it is both easier and cheaper for BlueHat not to get the suites and certify. Until the NIST make the suites free for anyone, at which point BlueHat can wait for someone else to make the investment and use it under the GPL.

It may not be nice. It may not be in the Linux Spirit. But this is commercial survival where nobody, anywhere cares about Linux. Ideals are cheap and easy when someone else is paying for them.

No! Linux really is a Free Lunch!

Yes, but it's gone stale. The real-world has moved on at considerable speed and the GPL and its associated baggage has been left behind. The new Web technologies require real licencing and real money. Before someone gets all sweaty about commercial exploitation, the licence fees are (at the moment) actually funding technology development. In the real-world someone has to fund the infrastructure, technology and other things that have in the past enabled Free Software to exist at all (although this is another debate). No, the real-world is charging down the Web - Java, Active-X, COM/DCOM, Corba, Cobra, Jigsaw, SET.

New terms and exciting new technologies that will create new markets. And what is Linux contributing to this incredibly fast moving market?

Well, we have concluded the debate about Linux's name. It won't after all be called Lignux, and most people will continue to call it Linux. Except maybe Richard Stallman who started the whole, stupid, pointless, idiotic naming debate. Who cares? Does it matter? And we do have an awfully nice emblem. Betcha that Jack Dee is pleased that his Penguin (adverts) have been adopted by the world's fastest growing operating system. And the Linux Penguin does look awfully pretty. And Linux is still getting Good Press. And we have talk of light appearing at the end of the tunnel. Could this be the New Dawn for Linux? No, it's the headlight of the oncoming Open-NT express. But the penguin does look awfully cute.


It'll never catch on. Linux is faster, you get the source code. It's tested by hundreds of thousands of people on the Internet. All true. And all Open-NT has going for it is UNIX Branding, Windows compatibility, DCE, Active-X and all the new Web technologies. In short, all the things that real-world users want and expect. Things that Linux can't provide.

And let's face the facts, the vast majority of commercial users couldn't care less about the source code. Many actively don't want it - it's a liability (yet more people hacking it, Noooo thanks). What they want is a Company they can call who will fix bugs. They are not interested if it is called Linux, Lignux or has a pretty penguin. They want support and a feel-good factor that Linux just does not deliver.

And UnixWare 2.0 Delivers!

SCO ( http://www.sco.com ) have announced that UNIXWARE 2.0 (and others) are available for free (ok, $19 for the CD) for non-commercial use. The intention is to make UNIXWARE endemic in Universities in partial retaliation to the Linux threat (don't kid yourself, they don't take Linux that seriously) and mostly to fight off Open-NT. If SCO were to make the source code to UNIXWARE available under (say the X11 licence, not the GPL) licence, then the whole future of Linux would be cast in doubt.

But Linux is Tested!

Oh really? If (as RedHat's recent announcement proudly proclaims) Linux's biggest strength is testing XX get announcement textXX just where are the results? Surely the test suites and test results (complete with QA plans and tester details) would be available somewhere. After all, just collating the results of “hundreds of thousands” of testers would take time.

Unfortunately, every time someone in the Linux world makes this statement, the Applications industry has a good laugh. Apart from anything else, they are laughing at the sheer naivety of those putting forward an argument that is so patently untrue and unsupportable. It is very well known that, at best:

.     Linux is tested by the actual developers, but there is little or no coordination.

.     There are no published test methods (hardly surprising as there is no Linux specification and in any case, the thing changes too fast).

.     At very best, Linux is tried by people on the Internet. Some discuss their experiences with others in a similar position and occasionally an expert gets involved. Given that the news groups are mostly noise, the developers can't wade their way through it all – they have better things to do with their time.

No, Linux is “smoke tested”. Which explains why none of the Linux CD companies put out a 2.0 kernel until it had been around for some time. It also explains why anyone porting to Linux picks a comparatively old kernel whenever they can.

And in case you are thinking that this is just a commercial whinge article, let's look at the way that Linux is developed and presented.

Lies, Damned Lies, Statistics and Linux Announcements

What is that in the air? Is it spring? Coffee? Lunch? No, it's bullshit from another Linux Distribution announcement on COLA (that's comp.os.linux.announce for those fortunate enough not to have read the news group). Because there is no specification for Linux, there are almost no limits to the mis-information that a distribution announcement can include. A few examples:

.     You can mention that the BlueHat Linux distribution has a WWW browser that is faster than Netscape. Pah! Bet BlueHat don't say how this is measured.

.     All manner of “inferences” can be made about the level of testing and stability of the distribution. In perfect safety! Because the users can't disprove the claims!

.     You can include all manner of standards and acronyms that nobody (probably including the poster) understands, in the safe knowledge that anyone qualified to answer is almost certainly not reading the announcement.

In fact, you can make a perfectly false, inaccurate post and get away with it. You can be sure that none of the other Linux vendors will correct any “inaccuracies” (lies is such an ugly word) as they have their noses in the trough as well. Hell - they are all at it. What's the old (almost joke) advertising slogan? And it is likely to get much worse before it is sorted out.

Linux development is not much better. Yes, it's great fun, a wonderful ego boost and that penguin is really very pretty. But where is Linux going? What is the point of many of the developments? Let's take (in no particular order) the SPARC, MIPS and ALPHA ports.

SPARCs come with Solaris - a perfectly good operating system. OK, so it does not have the source code, is Penguin Free, is not as fast and (ahem) you have to pay for it. But at least Sun support it. Linux on a MIPS is a great idea - but exactly how many MIPS systems are there looking for an operating system? And then there is the ALPHA. At least this is a system that Linux can run on. Except that the Linux-on-ALPHA aficionados fail to mention that there are no applications and that the ALPHA is expensive compared with the equivalent Intel system.

So many Linux developments represent hideous over-kill, solutions to non-problems or rampant Creeping Featurism. For the unaware, Creeping Featurism is the physical manifestation of engineers' tendencies to change and otherwise “improve” things that have, in reality, been completed. Creeping Featurism is FUN! Much more fun than (for example) producing documentation or (dare we mention it) test routines. Almost textbook examples of (Galloping) Featurism are Linux packaging and configuration systems. All the vendors have their own system, all presented as being “good for Linux” (read “good for the vendor”) and all of them are different and mostly incompatible. Any attempt to standardise is pointless; they have to be different, it's one of the few areas where value can be added to Linux and where BlueHat Linux can be seen to be different from ScrapWare Linux, which is different from... Oh never mind - you get the idea. Creeping Featurism results in packaging systems with more accessories and pretty lights than a tree at Christmas. Are all the features required? Who knows and frankly it's too late for it to matter as the damage has been done. It is now expected that a distribution will contain far, far more than is required or useful.

Ever wondered why there are so many CDs in a distribution? Is it completeness? Are they useful? No, it's historical. Slackware Professional started the trend for including the Sunsite and TSX11 archives at a time when they were actually useful and could be fitted onto a couple of CDs. Nowadays we have a binary CD, a source CD and about as many CDs as the vendor cares to include. The current record ( please don't write in with challengers) is held by Pacific HiTech who have a ten (yes, 10) CD Linux set. Contains almost everything you could ever need. But, like vegetables, Linux goes stale very quickly; and when one CD becomes old, the others are affected too.

Linux - The Way Ahead

Linux will remain as a valued operating system with some unique abilities in specific markets. But to really survive it needs to totally shed the baggage and pointless, religious, blind adherence to the GPL and the Free Software “spirit”. It's the GPL that is stifling investment. And Linux will need investment just to get it to the point where it could be considered for the new Web requirements. The distribution channel needs to have a reason (ie a good margin) to sell Linux, otherwise the decline will continue. No quality distribution channel anywhere will sell Linux if there is no financial incentive. Any talk of Channel Partnership is a fantasy if they have no margin. The old idea of “converting” UNIX resellers to Linux is laughable, given the margins that the Linux Community have forced onto the resellers.

Yes - it's you, the users that are ultimately to blame for the demise of reseller incentive and the eventual demise of Linux in a commercial setting.

This has been a hard-hitting analysis which exposes a few sacred cows and exposes Linux as the Naked Emperor that it is. Unless the commercial model for Linux can be fixed, the following scenario could become nightmare reality.

The lunatic price cutting results in the only margin resting with the applications. Few people sell Linux, most sales are direct from the manufacturers (adding to their costs and reducing their already bad margins). Given that the Linux vendors would be suffering from cash problems, more idiot price cutting of the Applixware nature would happen. Application vendors start to pull out of Linux (as they have no reason to “compete” in a non-market). Most of the Linux vendors redirect their efforts to something else (the Web), leaving maybe three vendors. Whilst the survivors are congratulating themselves on having won commercially (not realising that the others have shown good commercial sense by jumping into a lifeboat), they will quickly realise that they don't have sufficient income to maintain their distributions. Linux implosion continues until the only CDs available are just copies of freely available material with no added value. This particular Linux market will have died. Or more cynically, the customers will have the Linux they paid for.

The New Road

But, as nature abhors a vacuum, commercial enterprise will fill gaps. And there is a Linux market that we have not even considered. One that does not have a totally broken business model, one that has a Reseller Channel program that really works and one that invests in the technology. The snag is that it costs money to do this and it is reflected in the price; above $99. That's about £66 + VAT. There are a couple of vendors and products in this market: Caldera, Craftworks, Linux-FT to name three. Whilst RedHat have the ALPHA and SPARC versions at $99, neither can be classed as real market products (they are products looking for a market - if there was a market there would be more distributions and applications). Their Distribution Channel is either not interested in Linux (selling them as Solution Packages) or is fully aware of how commercially broken Linux is. Either way, the days of low and lower cost Linux have gone and the pretty blunt message to those whingeing about the cost is very much one of “pay up or go away”.

Of the vendors in the new Linux Market, Caldera are by far and away the most organised and successful. Far from exploiting Linux, they are investing heavily in the technology and possibly represent Linux's last hope of real industry impact. It is amusing to note that they are being imitated by others in the industry. Most notably RedHat (their 4.0 release is a poor attempt to copy the Caldera CND package) who fail to mention that their RPM package management system was in fact specified and paid for by Caldera. But then marketing is the art of mis-information and concealment.

High end Linux products represent the future for both Linux and applications being ported to it. The companies working in this area have real financial backing and are highly organised. To a large extent it is irrelevant if the changes these companies bring are accepted or not. Linux has to change and adapt if it is to survive. It will be the commercial companies that bring and finance these changes.

Low-end Linux ($49 and below) will still continue, just don't expect any commitment from the few resellers risking sales.

So when you hear someone complaining about the price of Linux, lack of availability and support. Just remember the amount of money that the suppliers are making. And give thanks that you can purchase it at all.

In spite of the broken business model, suicidal pricing models and some potentially violent upheavals, Linux has a bright future. It's just that you will have to start paying a realistic price. The free lunch is over. Enjoy the coffee. And talking of Java, but that's another story......

Linux Users Group Meeting

(Ted Harding)

Some thought is being given to the next meeting of Man-LUG (Manchester Linux Users Group).

One possible theme which had occurred to me is database work on Linux. This had not been particularly well supported – or discussed – hitherto, but recently has begun to look much better than it used to. Some of us have been waiting with bated breath for years for the situation to improve.

However, I would personally like this to be pursued beyond what is required merely for an evening's meeting in Manchester, and this is why I am also writing to Linux-Users and Linux-UK-Discuss.

In the first instance, I would be grateful to receive brief but descriptive messages from people who do, or wish to do, serious database work on Linux, so that I know, and can attempt to coordinate, an interest group in the topic.

Your interest may include “heavy-weight” work with very large databases, databases of smaller size but requiring complex queries, long-life DBs with stable structure, “throw-away” DBs which can be put on the shelf once they have been used to report on a data set, etc.

The merits and shortcomings of the various DB packages available (SQL or non-SQL, free or commercial, DOS packages running on DOSemu, etc) would also be of interest.

I would be grateful if those of you within reach of the Manchester area, who would be likely to come along to a Man-LUG meeting, would indicate this. If you have any suggestions for “hot topics”, or for the organisation of the meeting, these would be welcome.

For convenience, I would like the longer-term discussion to occur on the mailing list Linux-Users@mcc.ac.uk .

To subscribe to Linux-Users, send a message to majordomo@mcc.ac.uk containing the line: subscribe linux-users your.email.address

Similarly, if you wish to be on the Man-LUG list (very low-traffic, purely for Man-LUG announcements etc), send a message to

majordomo@mcc.ac.uk containing the line: subscribe man-lug your.email.address

Java'd Little Pill

(Lindsay Marshall)

It's official. I'm hacked off. It all started when I picked up my mail a couple of days ago. Sitting in the mail rack was a big envelope, looking all thick and interesting. I bet quite a few of you got one just the same. So, I ripped it open and pulled out the contents. “Java Computing Changes Everything” was the first thing I saw. I immediately looked at my shoes and they were still red and the walls of the corridor were still University standard magnolia, so I was instantly suspicious. Then I saw - “Revolutionary Concept Now Taking Root in the Enterprise”. Instantly my bogosity detector went berserk. Yeah, that's right, it's the latest hype sheet from SUN about Java!

“Internet time” it said. And then a lot more, some of it expressed in sequences of words that really couldn't be called sentences, let alone English (or even American for that matter!). I was introduced to the concept of “Java years” which I gather is something like dog years but accelerating. (Funny that dog years are fast but dog days are slow isn't it?). Then I learnt about thin clients - some sort of super model computing, I suppose. Oh and lots more nebulous stuff that didn't quite name names, but suggested all kinds of interesting activity. The South Sea bubble springs to mind. Try this for size:

“A major European telecommunications company will use Java to deliver new on-line services to consumers connected via Java devices in the home”

Try substituting the word “Electricity” for Java in that sentence. Meaningful isn't it? Perhaps every house in Cloud-cuckoo land has Java devices in them, but hey, the cable won't even get to the top of my street for another two years, and I need a new fridge before I even think about Java devices. (Maybe my new fridge will let me run Java - I've seen stranger things).

Anyway, now bored with the top document I moved to the next one, written by one Bud Tribble. It claimed to tell me what Java Computing will mean for the General manager and the CIO. I didn't actually know what a CIO was, but it turns out to stand for Chief Information Officer. I still don't know what it means, but at least I can pretend now. But there was a lot more exciting, post-modern jargon to learn inside! How about “webtop” - a desktop environment that can support Java applications. And the wonderful “fat client desktops”. (Yes, I know that wasn't a sentence - I'm getting into the swing of things here!)

It gets better though. I discovered that Fat Clients run on unnecessarily complex operating systems. I ran through in my mind all the people who sold unnecessarily complex operating systems. Wow, they are trying to corner all sides of the market!

Anyway there were pages and pages of this stuff all telling me that I had to scrap everything and buy lots of new technology to save lots of money. So I got bored with this paper too, especially since I don't have any money in the first place, and I turned to the next item. And this was the real gem. A veritable Koh-I-Noor! Written by one Evan Quinn of International Data Corporation, it was a paean of praise, a hymn of hype. They could record it and call it Anthem to the SUN, if that hadn't already been done. (I wonder what the world would be like if Huxley had written a book called “The Windows of Perception”? My brother has something to do with a program called Doors by the way, which I am told is very good.)

Actually the allusion is apt, because this document reminded me most of all of some of the more bizarre outpourings of Timothy Leary. Just use Java and the world will be a better place! Java will bring world peace and, OK, they might be starving out there, but they'll be able to run Java applications on their thin clients hooked up to the village Intranet! (I won't start on Intranets, I promise).

In case you are wondering, the medium for this transformation will be Sun's Java Workshop which “renders all other application development tools technically obsolete”. Let's forget for a minute that it currently runs like a dog on the average machine, and that in fact it won't even run at all on my Sun because I don't have the right Motif libraries installed. I mean, it's all written in Java, what could be better - it's “NOT a Windows application, or a UNIX application, or a Mac application”.

Apparently, this means that it is not bound by the inherent limitations of these environments. Think about this for a minute, (or maybe a Java minute which is much quicker and less stressful and more secure and reliable). I have this program (sorry to use such an old fashioned word) that implements the Java virtual machine. This program runs (well possibly) on my UNIX machine. All the services it needs, it gets from the UNIX operating system, anything else it does by good old straightforward computation. So how does this break any bounds? Let's not even think about computability theory or Turing machines.

But, I hear you say, what about when you run it on your “Java Thin Client”! Well, it's got to get the services it needs from somewhere, and if I read the paperwork right then these come from other systems that seem to be running something called Solaris. Hmm.

There's pages of this guff. But let's just cut out the chase and let the Fat Lady sing (before she's replaced by a Java Thin Lady naturally):


Anyway, I'm going to get another cup of coffee and wash down that java'd little pill. Maybe I'll walk round naked in my living room. I can certainly recommend it.

P.S. Normal service will be resumed next issue. I am sinking under vast tomes on some

programming language called Java which need to be reviewed...

Lindsay Marshall is a lecturer in the Department of Computing Science at the University of Newcastle upon Tyne. He has some UNIX DECtapes and a paper tape machine in his office.

News from Owles Hall

(Jane Morrison)
Firstly an apology - at the time of writing this piece I'm estimating that this Newsletter may still reach you before the end of October ..... It's late because of the production of the CD on the front cover - sometimes these things take a little longer than originally expected. If it reaches you in early November then we do apologise for the lateness - some of these things are out of our hands.

Anyway we hope you like the CD, it should be the first of many given free to our members.

Also enclosed in this envelope you should find the details about the Winter '96 Conference in Manchester. The programme looks good and we hope that choosing an easily accessible venue outside London will enable a large number of members to attend.

Josette Garcia of O'Reilly books plans to bring her book stand with her to the Manchester event. She often gives good deals on the books at these kind of events and has T.Shirts etc. to give away.

The Winter Conference last year in York, although it did not attract vast numbers of delegates, was a definite success with discussions continuing in the bar until the very early hours! Let us hope the Manchester event will be a success also.

Future events, including a Seminar in February 1997, are also being planned and details will be sent to you very soon.

Any comments on the different paper used to print this issue? We felt it would look better, but please email office@ukuug.org and let me know if you agree.

Jane has worked at the Owles Hall Secretariat for 8 years. She looks after the administration for the UKUUG, SUN User Forum and EurOpen. When not working, her pastimes include gardening, swimming, painting and decorating, going on holiday and doing absolutely nothing and trying to keep her 18 year old son on the straight and narrow!

Stopgates T-Shirts for-sale

Now you can get your feelings off your chest! Stopgates T-Shirt now avalable in the UK! For one of these high quality white T shirts with the “Stopgates” logo on the front and “Before he stops you” on the back send a Cheque or Postal Order for the value of your order payable to Deluxe Technology Ltd to the address below. Check out http://www.mh01.demon.co.uk/stopgates.html for the latest information on availability and other offers. The T-Shirt is available in three sizes: Medium, Large and Extra Large. The cost per T-Shirt is £12, plus £3 Post & Packing (for any order). Send order to Deluxe Technology Ltd (see contact pages for address).

Please allow 28 days for delivery. All prices include V.A.T.

Order enquiries to:

Tel: 01763 273 475
Fax: 01763 273 255
Web: Webmaster
Queries: Ask Here
Join UKUUG Today!

UKUUG Secretariat