[UKUUG Logo] Copyright © 1995-2004 UKUUG Ltd

UKUUG


Previous Next

UKUUG News

Editor's Column

(Susan Small)

Sue

Our apologies for the lateness of this newsletter. As usual, contributors left their submissions until the last minute and there was the usual rush to get things to the printers.

We have a number of trip reports in this issue. These should be of special interest to those of you who missed the Eric Raymond and Richard Stallman evening talks in London. Also included is a report of the last Netproject Linux Conference.

Those of you involved in financial computing within the European Union should find Finnbarr Murphy's article of interest as the single European currency reaches our desktops and printers.

As promised, this month's cover CD is the Red Hat Linux 6.0 Distribution (for SPARC architectures).

There are a couple of thought-provoking articles which I hope you will enjoy and respond to. The first by Con Zymaris concerns the future dominance of open source and the second, by Steve Talbott, discusses the diversification of Amazon.com.

Finally, we need to build up our band of book/product reviewers. If you are interested in contributing reviews to news@UK, then please let me, or Jane at Owles Hall, know what you would like to review. Have a good summer holiday season and my best wishes to you all.


Report from the Chair

(Lindsay Marshall)

Lindsay

It seems a long time since I last sat down to write my last Chairman's report and again UNIX (well, Linux) has been much talked about in all the media worldwide. Eric Raymond and Richard Stallman have been in the UK airing their views on the open source movement at a couple of Linux-oriented conferences and they both took time to speak at UKUUG organised events which were gratifyingly well attended - read the reports in this issue. (And read my take on the open source movement as well.)

Coming up on June 25th and 26th we have our own Linux99 event which has something for everyone in the range of speakers. If you are interested in Linux I hope that you will go along - it promises a great couple of days and is very reasonably priced as well! Check out the UKUUG website for more details.

Various people have been prompting me to remind you about the mailing lists we have set up for members. These are run using Majordomo so the usual subscription methods apply - to join list xxx mail xxx-request@list.ukuug.org with "subscribe" in the message body. (No we haven't started a hardcore porn list, that's just an example!)

The lists are:

members@list.ukuug.org
This is a moderated list used for announcements which may be of interest to members, such as future events. Normally only council members will post to it - we recommend all members join this list. (I'd better join then I suppose!)
chat@list.ukuug.org
This is unmoderated and is for general discussion between members. You will get the big heave-ho if you spam this or behave in an unruly fashion though.
jobs@list.ukuug.org
Another moderated list with (at the moment) very low traffic. This is used for passing on information about jobs that get sent to us - employers still find it hard to get hold of qualified UNIX people.

That's about it for this time - and it might even be the last time from me as my six years on council are up this year and there might be a new chairman before the next issue of the newsletter comes out. There again there might be one more.


Linux Links

(David Hallowell)

This is a starting point to various sources of Linux information.

General Information

News

Software

X11 (X Window System)

Search Engines

Distributions

Interesting Software

Fun Stuff


News from Owles Hall

(Jane Morrison)

Jane

For those who did not attend the evening meeting on 23rd March, at the Commonwealth Institute, to hear Dr. Richard Stallman give his presentation entitled "The GNU Project and the GNU/LINUX System", I can inform you that the room was completely full with attendees and we guess that approximately 400 people attended [ Review here: Ed. ].

We were very pleased with this number and have been able to secure email addresses from the attendees to keep them informed of future UKUUG events.

We should mention that the event, (which was free to those attending), was very kindly sponsored by Goldman Sachs.

By the time you come to read this hopefully you will have booked your place at the forthcoming UKUUG LINUX Conference in Aston, Birmingham, on 25th & 26th June.

If you would like to find out more about the event visit the UKUUG web site - www.ukuug.org - or call the Secretariat.

With the last issue Newsletter (March) all members received their free 'Powered By Linux' badge - and lots of members have already ordered extra supplies!

If you would like to order some more badges please note the costs:

Money must be sent with order (cheque payable to UKUUG Ltd. - or credit card details) to UKUUG Secretariat (details under Contacts).

We have recently given members the opportunity to subscribe to two new mailing lists:

If you have not seen details of these lists, then this could mean that we have not got your email address in our database. If you wish to find out more about these mailing lists, please email the Secretariat.

Future UKUUG Events


A trip report on the Eric Raymond Talk

Kieran Barry

I travelled up from Uxbridge with 3 other CS students form Brunel University. We got there at about 6.45 and walked bang into Raymond (ESR) and Alan Cox (a long haired bloke wearing a genuine red hat).

It was actually kinda funny. I turned round and there was this short guy with a moustache standing half concealed by the pillar behing me. He had a yank accent, faded black jeans and a bue denim shirt on. I had my suspicions so I scanned his name badge and it was ESR. He was talking to the dude in the red hat, I scanned his badge. Alan Cox. I put on my best cheesy grin and leaned forward to overhear, then lost my nerve and went back the other side of said pillar.

We hung around for a ten minutes feeling a little awkward, then went downstairs to the lecture theatre to wait.

ESR was introduced, insisted that he needed a body mike (he said he learned his speaking style from american stand-up comics), then polled the audience on what they knew, and what they wanted to hear.

The first two questions were about who had read "Cathedral and Bazaar" and "Homesteading the Noosphere". Almost the whole audience had: this was a well informed audience :-).

He also asked how many in the audience programmed for a living (90%) and how many would lose their jobs if the sale value of software disappeared. The answer to this was maybe 20 people in an audience of 300-500. I got the impression in this and later questions that many in the audience didn't understand the question (I think it is based on the labour theory of value, which features often in Marxist economic theory.)

He then polled us about our preferences for talk theme between

  1. Anthropology
  2. Business models
  3. Being an effective open source advocate

Number three won by a short head from business models. He made a number of mentions of "hacker-type" personalities which I thought came from the anthro-talk, which would have been much newer ground for me in hindsight, but the talk he gave was excellent.

ESR is between 5'3" and 5'6" tall, with a distinct limp. He has sandy hair, sort of borderline long (about to make contact with his collar, and a bushy moustache which over runs his mouth by about half an inch on both sides. He has good comic timing, and is clearly a very confident, experienced public speaker. Also noticeable was that when asked a question, at the end of his answer, he would ask the questioner whether they were satisfied (it became almost a catchphrase: "was that a responsive answer?") On most occasions they were, and the answers (unlike a politician's) all addressed the question.

He started out by giving some background to the hacker culture, saying that the current open-source movement can trace its lineage continuously back to the late 70s. He believes that the modern free software culture has not changed greatly in that time, which suggests that the Linux/apache/gimp/etc effort is part of a culture which is likely to be durable. He also said that in that time, the hacker culture had been singularly unsuccessful in converting industry to use "free software".

The standard approach was for hacker 1 to approach a co-worker and convert them to the faith. This was repeated until all his peers saw the light. Then they went to work on the line mangaer, who converted his peers, and worked up the line until it got to the board room. Unfortunately, this had never happened.

Until, that is, Netscape released Mozilla.

ESR said he had not been informed of the decision or any of the planning that went into it until a friend phoned him up and said something like:

"Hi Eric, did you know Jim Barksdale, netscape CEO, has just announced that he is gonna post the source to Navigator on the web, and he says it's because of your paper. I just saw it on CNN!" (NOT a quote, not even close, but similar, played for laughs.)

So he went to talk to Netscape, and decided that:

  1. This was huge. A fortune 500 company was releasing the souce to its flagship product.
  2. It was important because Netscape had a rep as an "Industry Trend-setter" on Wall Street.
  3. If it failed, it would be another 20 years before open-source was tried again.

So he started thinking long and hard, and at the end of it decided that he was about the best qualified man to act as a free software advocate. (Open source didn't exist at this time.)

His thought process went like this.

  1. Hacker culture needed a main spokesman.
  2. He needed to have credibility in the hacker culture going way back.
  3. He needed to have plenty of free time.
  4. He needed to speak the language of corporate America
  5. He needed to be comfortable speaking in public (an extrovert).

At which point ESR noticed that the pool of candidates was getting small.

Later, he coined the term "open-source" and trademarked it, specifically so that he could stop MS claiming to comply with the standard. The trademark was given to an organisation called "Software in the Public Interest", which was closely associated with the Debian Project. Then he left, with the trademark and the president, who gave it to him, then helped him to set up the "Open Source Initiative".

The "Open source" trademark is apparently based on the Debian project's licensing, which specifically permitted a "base plus patches" license, which has been so controversial in the TrollTec/Qt/QPL case.

Then ESR moved on to effective Linux advocacy (he seemed to be pushing Linux big-time.) He said that there were two rules.

  1. Talk to the boss (CEO) of the biggest companies around, because working from the bottom up didn't work.
  2. Speak in a language that they understand. This generally means business-school-speak.

His pitch stands on three planks.

  1. Stability. Why put mission-critical applications on an unstable base? In science and other branches of engineering, peer-review is built into the culture, so that in biology or physics, an experiment is not accepted until it has been duplicated elsewhere. An experimenter is expected to publish not only his results, but also details of his equipment and procedures.

    This was an alien practice in the computer industry, except for one small area: the internet. In this field, overcoming legacy, non-standard hardware, different operating systems (exactly the sort of conditions that the software industry claims to be the most difficult possible0, a group had built a system for no commercial gain.

    The system got everything talking despite distance and unreliable hardware in a robust manner becaus it was peer-reviewed.

  2. The total cost of ownership argument. "Open-source is cheaper to run, and easier to train people on" I don't remember the reasoning exactly, but ESR also said that this was far from key, and given limited time could easily be left out.
    <comment> One of the points was that the education system loved access to source code, and would reduce training costs. Someone pointed out that in his university, this was not an obvious trend. Perhaps this suggests that this side of the pond will witness a rather weaker version of the "open-source effect", to coin a phrase. </comment>
  3. The strategic risk argument. This runs like this: The CEO gets a lot of money to make sure that nothing that he can control is allowed to mess with his company. Therefore, the prudent boss makes sure that he doesn't expose his company to the mercy of a monopoly supplier.
  4. This thinking, which is standard business practice, has never before extended to software. But why rely on a single company which may or may not respond to your need? With open source, you have a choice, because system support can be done by anyone with access to the source code.
And then we were told we had to leave.

I may have messed up on a couple of points, but the general sense is similar to what he said.

Hope you enjoy this!


A trip report on the Richard Stallman Talk

Kieran Barry

/* The UKUUG (United Kingdom UNIX Users' Group) organised a talk by the [legendary | mythic | influential | kooky] hacker, Richard Stallman (RMS) on Tuesday 23 March 1999 at the Commonwealth Institute in Kensington. This is an attempt to give others a feel of the event. */

Tuesday afternoon I attended a meeting about the alleged consultation for the forthcoming ECommerce bill, so I travelled to Kensington from the LSE, with three others, and we discussed various matters on the train. I was kind of surprised that Stallman was held in greater respect than I expected. In the introduction above, I suggested that RMS is often viewed as a visionary kook whose time has passed. From the conversation in the train, this is at the very least arguable.

The event was due to start at 7.00pm, but it didn't, because the audience couldn't get in quickly enough. The hall the event was held in holds about 500 people, and I am confident that at least 300 of those seats were occupied. The problem turned out to be that the organisers were trying to get people's contact details at a table in the hallway near the entrance to the auditorium. They used the cunning ploy of selling "Powered by Linux" sticky badges for 50p at the same spot, resulting in a huge scrum, even at 7.15 when I was passing.

The result was that as late as 7.15 there was a queue backed up out the door of the centre with at least 20 people in it, and I am trying hard to be conservative!

To find a seat, I walked round in front of the stage, and RMS was standing at the lecturn, stock-still. He is in his 40s, with long brown curly hair to below his shoulders, the kind you see on Hell's Angels, along with a beard and moustache. He wore a black t-shirt, and looked a bit tired.

The talk, we were told by the chair, would have to end by 9.15 sharp (due to the policy of the venue. At both this and the Eric Raymond (ESR) talk 2 months previously, only this really prevented the meeting going on forever! Probably a good thing for the speaker's sanity :-) and that RMS' speech was timed for about 2 hours, so could questions be please left to the end.

A couple of words about RMS' reputation as a kook: he single handedly made the Linux systems we know and love today possible (details of which will follow), which makes a towering figure in the software world, but he has a way of doing things which appear outlandish and provocative. An example of this is his request that Linux systems be called "GNU/Linux" which, to someone not knowing the background seems ridiculous. He is also known for turning up on mailing lists and newsgroups with provocative posts. Eg

"In late September, 1994, Richard Stallman of the Free Software Foundation posted an article to comp.lang.tcl and several other newsgroups titled "Why you should not use Tcl". Predictably, a flamewar ensued, which lasted in one form or another for almost a month (until it was pre-empted by the GNU project's announcement of plans for its own extension language, later dubbed GUILE)." (Details available at http://www.vanderburg.org/~glv/Tcl/war/, the TCL War web site)

In addition, his view on computer related issues stem from his world view. This should become more apparent later on in the post, but without knowledge of his belief system as context, single ideas can appear outlandish. RMS might be called the man who lauched a thousand stories, (the unchariable might say a thousand flame wars) and one of great things about his talk was that he included many of them, but gave the context as well.

As a speaker, he falls into the less than flashy category. Right at the beginning he told us that he was extremely tired, having flown recently from Asia, which can't have helped him, but he held the audience's attention entirely, except for whispered comment on what he had to say.

The key to his views on software lies in his experience in the MIT AI lab. He was a member of a hacker culture whose currency was cooperation and code sharing. The culture extended far beyond the walls of the AI lab, for example allowing him to obtain a cross-assembler from the PDP-10 to the PDP-11 from Stanford via somone at Harvard.

To me it seems that the concept of not sharing the source was beyond his comprehension: he didn't have the framework for the idea, it just couldn't happen.

And then it did, and slowly began killing the community he was a member of.

It started with broken printers. Printers jam, they run out of paper, and they need to be fixed: it was ever thus! In the AI lab, the printer was in another room, so they rigged a way of passing a message to a couple of people with jobs in the printer queue to let them know about any problem as soon as it occurred, and this system, "wiring the human element into the network" made everything more reliable.

Then Xerox donated a laser printer to the lab. It was top of the range, capable of printing hundreds of pages a minute at very high quality. But it still jammed or ran out of paper. RMS made inquiries to Xerox for the source code so he could get it to send the problem messages, but they wouldn't give it to him. He tracked down someone who had the source he needed at Carnegie-Mellon, and the guy refused saying that he had signed a Non-disclosure agreement.

The way RMS put this was instructive: to paraphrase, he asked for the the source, and the man with the source said, "No, I promised not to give it to you." "And he promised not to give it to you (pointing), and you, and you, and pretty much everyone here except those not born at that time." RMS' view of software is that it should make the world a better place, IMHO, and under those assumptions, NDAs become Very Bad Things, whose sole purpose was to stop him making the world a better place.

RMS then went on to explain the founding of the GNU project, saying that he felt unable to work on a project that didn't make the world a better place. (I guess that he's been called a lot of names over the years: ESR caricatured the view of a senior executive who was being encouraged to use free software as saying, "Free software? Isn't that the stuff that commie from Cambridge [where MIT is based] goes on about!")

The feeling is mutual. RMS attacked economists more than once, saying at one point that they approximate the real world step by step till they reach the conclusion they wanted in the first place. An example he pointed to was the idea that tax cuts for the rich benefit the whole of society. He doesn't agree!

Stuck in a moral quandary, he pondered giving up on programming in the early 80s. The hacker culture he identified with so strongly was bound tightly to the DEC PDP-10, and in 20 years worth of PDP-10 assembly language which made up the Incompatible Time-sharing System (ITS). Its death-knell rang when DEC ceased production of the machine, since that code, all unportable, would die. The new generation of machines came with closed operating systems, and were unpleasant to work with. He has no truck with programmers who say, "If I don't write code for this project, I'll starve!" responding that they could always work as a waiter.

The world lost a potentially unforgetable waiter when RMS decided to fight back. He decided to write his own operating system, and release his own source code. His aim was that it would be adopted as a standard operating system, and the hacker culture could be reborn.

The ITS died because it was non-portable, so the operating system he modelled this on was the UNIX system, since at the time UNIX was the only portable OS around. He resigned from the AI lab (I've heard elsewhere that this was to avoid any risk of them laying claim to his work, but he didn't mention this at the talk) and was fortunate that the director of the AI lab allowed him to keep his key.

At the time he made contact with a number of companies with a view to securing funding for what he called the GNU system (short for "GNU's not UNIX"), but although several expressed interest, they wouldn't give him money.

When the first component of his putative GNU system, the rewritten editor GNU Emacs was finished, he made it available via anonymous ftp. Nevertheless, he began to receive requests for copies. He sold tapes at $150 a shot, even though it was already available elsewhere, and soon found himself selling ten tapes a month, steadily, which was enough for him to live on.

He made a point here that the assumptions of economics should rule out this situation. The price of something should be dictated by the market rate, and the prevailing market rate for GNU Emacs was $0, yet people kept on buying from him, despite charging considerably more.

The whole problem was that this was his experience!

Another major point that came out was RMS' model of the freeness of software.

Level 0 freedom means that you can use the software.
Level 1 freedom means that you can fix your copy of the software.
Level 2 freedom means that you are allowed to give copies of the software to others.
Level 3 freedom means that you are allowed to improve the source by providing your fixes to others. (I am not sure about the wording for these.)

/* I think that there is obvious confusion between the ordering of 1 & 2. Is there more freedom in being allowed to change a program than in being allowed to distribute it? But I digress! :) */

He also told us a story about X-Windows which was written as free software at MIT. During the early 80s, UNIX (which came in many versions) did not have a "windowing system", so the software houses would bundle X-Windows with the OS. But they didn't send the sources with OS!

This made a very strong impression on RMS. A piece of software which had been created at level 3 was being used "in the wild" as level 0. This inspired him to create the concept of "copyleft" (I don't remember him mentioning the GPL in his talk: although he may have, he certainly didn't in this section, where the emphasis was on freedom.)

Copyleft was inspired by copyright :-)

Copyright forbids unauthorized copying: copyleft permits it.
Copyright forbids changing things: copyleft permits it.
Copyright allows fair use: copyleft places conditions on the use, such as redistributing the source along with binaries.

All of Stallman's work was released under copyleft, which is more commonly known as the GNU Public License (GPL). And when people asked him to work for them, writing a training manual here, consulting there, he released the fruits under copyleft. And people started sending in improvements to the work he had already released. This prompted the founding of the Free Software Foundation. He created it and gave it the business of selling the tapes of Emacs and other GNU programs.

As an aside, he mentioned that when the FSF had enough money to start paying programmers, he decided that he couldn't bring himself to pay himself on the grounds that from the FSF's view, paid or not, Stallman would continue to code for them. I guess the man is just an ascetic :-)

He also discussed the long-running "story" of the naming of Linux. RMS some time ago "suggested/asked" that Linux be called GNU/Linux in recognition of the work of the GNU project: it sparked a long controversy. He felt that the enormous amount of work that the GNU project put into the code which is now being distributed as the Linux shell and utilities, (and we all know that Linux systems would be hamstrung without gcc, tar, gzip, make, etc.) should be recognised.

At the end of the night a question was asked about a project to create "BSD Linux" using the utilities which form part of the *BSD systems, and RMS responded that they were insane. It was a glimpse, I think of the way he debates. He doesn't care much for the diplomatic niceties.

BTW, he mentioned that the Hurd, GNU's original kernel is due to be released later this year.

He also commented on the concept of Open Source software, and specifically the licence under which Apple recently released some of its code. This licence included provisions which required any changes to be mailed to Apple, and gave them rights to use them, while Apple was not bound by this clause. In addition, if an allegation of patent breach came up, Apple reserved the right to withdraw the right to use their code, and any code based on their code, putting anyone who used it at the mercy of the whims of both Apple and the US Patent offfice, (about which he was truly scathing.) This nevertheless was accepted as meeting the definition of Open Source Software by the OSI (athough I believe that this may have changed, according to rumours on Slashdot.)

Ian Jackson, former head of the Debian Project and current president of Software in the Public Interest, stood up and said that he considered SPI to hold the rights to the "Open Source" trademark, and didn't feel that Apple had fulfilled the requirements.

Stallman also spoke about defending against the predations of commercial software, which from the above it should be clear concern him greatly. Some companies have refused to release specs for their hardware, and instead, mirroring Xerox all those years ago, releasing closed drivers. RMS pointed out that the Linux kernel permits the use of binary-only drivers as modules, which he felt was the easiest place to take a stand against commercial, closed software. A telling point was that his objection seemed to be that now, the free software movement would always have to be on its guard. I felt that he was playing Cassandra, and using his reputation, as a warning against a potential threat in the future rather than a "clear and present danger."

His big plug of the night was for people to get involved in the GNU project. They need both coders and documenters, details of which are available from http://www.gnu.org.

An area of overlap between this talk and the ESR talk was his discussion of free software business models. He said that people like Cygnus Solutions had made a good business out of selling support for GNU software, and that this was a viable way. Unfortunately, he also said that Cygnus had abandoned the model recently.

Above, I spoke about him not being a flashy speaker, but he had a couple of stunts which came straight from the file marked "kitsch stunts".

As he received his ovation after his introduction, he turned around to applaud the curtain behind him, which probably cut the clapping short and allowed him to begin his speech. Later, he donned a gown and a fake plastic halo to give a "homily" from his alter ego, "St. Ignucius", which to my eyes at least, he didn't feel particularly eager to do it. Perhaps it was the jet lag, but I felt at the time that it demeaned him, and that it belongs to his past, when he cut a lone figure on a landscape of commercial software.

The "St Ignucius" section said that anyone who used only free software could join him in the Church of Emacs. (I have checked on the licence of Vim, and I am not sure that it fulfills the definition, but its pretty close. Vim users in the Church of Emacs? What's the world coming to!)

How to assess the man? The reception he got at the meeting was warm and receptive. It seemed to be a younger crowd than the ESR talk attracted, and the Eminences Gris from the UKUUG were not so prominent, so it may be that those who disagree with him stayed away or stayed quiet. On the other hand, the material in the talk is widely available (most of the stories above are available from other sources) so it may be that he only has one talk!

As a figure in the Free Software community, RMS is still a man of huge stature. He acts as a conscience for the movement, eternally vigilant. He is motivated by his own principles, rather than the views of others, which makes him a provocative influence. His principles sometimes make him look like a figure from the Life of Brian (splitters! :-)

His views seem to belong to the past, yet they cast a long shadow over the present, and his feeling that no community spirit can survive when the community is unhappy (his original objection to the Xerox software) is something that I have not seem mentioned before.

I hope that I haven't offended anyone with my rambling.


A trip report on the NetProject Linux Conference

(Alain Williams)

This was the second NetProject Linux conference, like the one in January it was held at the Commonwealth Institute in London. That was a good start, only three stops on the tube from Paddington.

I looked at the programme, what to go to ? The conference was dual tracked: with management and technical streams. Evelyn then explained that the reason my name badge had so many pretty ribbons on it was because I was chairing the first technical session - decision made.

First thing: upstairs for coffee. Meet plenty of the old friends that I had first met at the UNIX/UKUUG conferences almost two decades ago. One of the things that I like about Linux is that it has brought back the social spirit that existed in those days: technical ideas, problems and solutions were discussed and swapped; we were all in it together and keen to help each other. Later the marketing men took over and competitive walls were raised, hopefully Open Source will prevent that happening this time.

The first speaker was Ian Batten of Fujitsu Europe. He was talking about the problems of integrating Linux into a predominantly MicroSoft world. The main problems were that of file formats: there are several products that will make a reasonable job of reading and writing MicroSoft Word and Excel files, but none of them does a perfect job. Running the MicroSoft applications under something like WINE or WABI was an idea, but these didn't always work properly. The best solutions seemed to be Windows Terminal Server (which could be expensive) and the new VNC, these allow someone to work in their preferred environment and hop into Windows for the occasional application.

Many users don't care or want to know about such problems, they just want to get on with their job. The main issues are the applications, the operating system is a secondary consideration most most users. This is what lets MicroSoft get away with it's "extend and embrace" -- take existing protocols and warp them so that non-MicroSoft applications have difficulties.

Often these "enhancements" don't really help with the job in hand, after all Charles Dickens wrote all of his books in one font, are they bad because of that?

One thought that he left us with was: "Remember the problems last time MicroSoft updated Word. What are you going to read those .DOC files with in twenty years time". [I don't care, I still use troff].

Mike Banahan (GBdirect) talked about how his company had moved entirely to Linux 18 months ago. He talked about how they built a web based training directory ( http://www.trainingpages.co.uk) on top of Linux with Open Source tools: Apache, GNU C++, Perl, mySQL. They used C++ as they knew it well, thought that it would be faster and more scalable; that was a mistake and Perl throughout would have been better. mySQL has worked well and is fast.

The hardware costs were minimal, no software costs. They have never had a crash. Mike spent some time explaining how it worked and what they had learned: "Javascript is a nightmare", "Java doesn't always work", "Weblint is a good idea".

Time for a fix of coffee and a tour of the exhibitors (why does O'Reilly have to turn up? I spent £65, and so was grateful of the 25% discount that UKUUG membership gets me). Informix turned up to rival Oracle, there was a software publisher, a couple of consultancies and systems vendors (good to see the small guys doing well in the Linux business). I was tempted by the flat screen that Dell were showing Gnome off on -- until I learned what it cost, next year, maybe.

Both streams joined together for the star of the event: Miguel de Icaza, the man behind the Gnome project. Someone had compared him to the portrayal of Wolfgang Mozart in the film Amadeus "off the wall". He is a bright, open person who knows how to speak well. I can see why he is doing well at leading the Gnome project: not only does he understand the computing but he has a good personality that lets him lead well and keep people wanting to work with him, he held his audience well.

He explained why he started Gnome: Linux is mainly found working as a server, but servers only account for a small percentage of computers in use. There is plenty of "free" software for servers, not enough for the desktop. (Quick diversion to explain that "free" means liberty, not zero cost). To get good a desktop a consistent user interface is needed, and a means to to integrate tools into the system.

Where possible they have reused existing standards (they don't have the effort available to do lots of research). CORBA is used, this is an InterProcess Communication (IPC) mechanism that enables much code reuse -- like pipes but better. This allows applications to be written quickly and easily. Applications are small, they reuse other applications to do things that they can't themselves do, they can be reused by other applications. Drag and drop is all over the place.

He talked about the Bonobo document model (named after a species of highly sexed monkeys). In many ways this is rather like a filing system within a file and allows things like a document to contain a spreadsheet that contains pictures. It is based on MicroSoft's OLE2/ActiveX design (he made several compliments about the good design).

There are 288 developers in 21 countries that have write rights to the 500Mb of code in the CVS system, many others talk to these 288. Red Hat Advanced Development Labs have done good integration work and provide some seven full-time programmers. You can buy Gnome support at http://www.gnome-support.com.

There are many desktop and groupware tools in development: spreadsheet, mail calendar, address book integration, gui developer, music, presentation. There is no word processor, use AbiSource. Multibyte character support has recently been added. The recent 1.0 release was of about 20% of the available code -- what was ready.

See also http://www.gnome.org, and http://bugs.gnome.org.

He also mentioned a bit about the use of Linux in Mexico's schools. My notes say 1,000,000, but if that was 6 to 15 year olds or computers I can't remember. But at $100 per MicroSoft license it is a large amount of money saved. The schools are using Gnome/Linux.

I like buffet lunches, it means that you can wander round and get to speak to interesting people. I met someone who I think will help in my pet aim of trying to get Linux widely adopted in UK schools. I don't want to replace the existing MicroSoft (and few Macs), but to have Linux work alongside. I believe that children will greatly benefit from the use of different systems, it means that they will need to have to understand and think (i.e. work out what they are doing rather than working by rote); this is quite apart from the huge cost savings.

After lunch Andrew Findlay (Brunel University) talked about one of the current NetProject efforts: working towards a Secure Single Sign On. The idea is to have one username and password/token for all services (or per role). The trouble with networked computer systems is that they all want passwords, you have to keep on quoting them and remembering the different ones for different uses -- then some time expire.

There is an important distinction between Authentication (you are who you claim to be) and Authorisation (you are allowed to do X with service Y).

Some strategies were looked at and compared. The difficulties are as much political as technical -- why should you be allowed to control part of the authentication on my machine, you are in a different company or (even worse) a different department in the same company.

The building blocks are Public Key Infrastructure (certificates, directories, agent/security libraries), and hooks into existing systems (UNIX PAM (Pluggable Authentication Modules) and NT GINA (replaceable authentication library)). NetProject is producing an Open Source demonstrator of these ideas.

John Logsden (Quantex Research) and John Brazier (Brazier consultants) talked about the recently released VMware product. [ See review elsewhere in this issue. Ed. ] This allows you to run Windows (98 or NT) on top of Linux, and soon Linux on top of Windows NT. They had a few problems with the first release, but basically the product works well.

This means that you can get the best of both worlds. Many of us took that to mean running MS Word on Linux, but they think that this will cause MicroSoft a lot of grief as it will provide a way in for dedicated MicroSoft shops to get to use a few Linux applications, and then once they get to know them to gradually use more. This will break the MS strangle hold on the desktop, people will genuinely be able to choose the best tools for the job.

There was some description of the virtual machine that was created. There seems to only be a small performance hit, but (unless you want it to thrash) you need lots of memory.

Up for some more coffee where someone came up with a brilliant use for this. You want to do network testing: configuration or secure IP tunneling or something. Under Linux run up several guest Linux systems and set up a private network on which you can do what you want -- without sending a byte out of your Ethernet card. That is what I call smart.

Miguel then treated us to a bit more of a technical view of Gnome. I must admit that I was impressed by the great emphasis on correctness, and a drive to reduce bugs. There is a great use of run-time consistency checks and asserts, every function checks its arguments before using them. The coding style has got to be consistent, code is rewritten if it isn't quite right. There is a focus on security: it is more important to be secure than simple. Stick to C, it is less complex than C++, avoid threads. This is a refreshing change from much of the commercial code that I meet: "if it passes a few quick tests: ship it".

They use lots of libraries rather than one big one as it is easier to maintain that way. Don't forget that the Gnome developers don't work on it full-time and so it has to be understandable by such a community, small components help in this. Some of these libraries were looked at.

Miguel told us how to pronounce his name (we Brits all got it wrong), and had many strong opinions: "TCL is a mistake which shouldn't have happened" [sorry Linday]. He also showed us a Gnome that Alan Cox had given him -- will this be the start of a major new market in Mexico for the UK garden Gnome industry ?

Alan Cox then talked about some work he had done a CymruNet in Wales a few years ago. They introduced cheap E-commerce to small businesses, how could they get credit card authorisation done over the net? Orders were to be made over the net and passed on to CymruNet customers.

The first problem was to find a bank that understood what they were trying to do (in 1996), security (40-bit Secure Socket Layer). What server: NT vs Linux or BSD (I won't bore you with the reasons why NT was discarded) Apache is the web server and a public domain shopping site system that Alan found in the Far East. The end result was very successful.

All the speakers then joined for a panel session: Gnome and VMware were the hot topics.

Then upstairs for a beer and to listen to the excellent jazz band.

I spent all of my time in the technical sessions but would have liked to spend some time in the management stream. I spend some time explaining to people that you should use the best tools for the job, I like to see what other people have done so that I can quote them as examples.

All in all an excellent day. Well done Eddie.


Previous Next
Last modified 06 Feb 2001 00:00
Tel: 01763 273 475
Fax: 01763 273 255
Web: Webmaster
Queries: Ask Here
Join UKUUG Today!

UKUUG Secretariat
PO BOX 37
Buntingford
Herts
SG9 9UQ