Outer Web Thought Log
September 29, 2003
#100
Kudos to Carl Masak of the University of Uppsala. He's attendee #100 of the Cocoon GetTogether.
September 24, 2003
European software patents
Word is coming out: the European Parliament voted in favor of the proposed legislation on software patents. While a good deal of amendments made it through the vote, it remains to be seen how much the European legislation will differ from the US one, once it gets implemented. There's a mixed mess of hurrays and boos to be heard on the relevant lists and fora, as it isn't clear what difference the amendments exactly mean. Patent activists indicate that the overall legal text is different in many ways from the one in force in the US, but still there will be some important lobbying going on by BSA peeps to finally get what they want from the European Council.

I'm again thinking abuse-prone patent legislation should be countered through open source licenses, i.e. by adding clauses which disallow companies that abuse patent law for their own good to use a given open source project.

We don't have much by which we can fight with, but I'll have a serious look at the xReporter license in the next few weeks.

September 22, 2003
Steaming ahead
The Cocoon GetTogether date is approaching, and we're at 68 attendees now. Which is only 17 less than the people who have effectively showed up last year, and since this time, there's a small participation fee, I'm confident that registered people will show up. With two more weeks to go, that means we might get to 90 people - let's keep fingers crossed. So much for my initial anxiety - and a testament again for the faithful Cocoon community. This year, we are already registering 13 different countries, instead of 11 last year. Cool. :-)

My days are now slowly becoming consumed with fun event administration side-activities, like renting a piano (?), nagging speakers to send in their handouts, and arranging lodging for a group of 8 budget-conscious travelers. Next on my list is finding a company renting wireless access points, since I'm getting afraid that our SoHo APs won't cut it when trying to connect with 20+ laptops. Anyone in Belgium reading this with a clue, please yell.

September 18, 2003
Show me your business plan!
I've been hitting the roads lately, being involved in the creation of an Open Source Center in the Dutch-speaking part of Belgium, and also because the Belgian market is (slowly) picking up Cocoon as an alternative for building robust XML-centric web applications. We're getting out-of-the-blue phone calls asking about Cocoon guidance and support, which is rather cool after making the decision to intensify our company's focus on Cocoon, and not trying to support other frameworks out there (such as the ubiquitous Struts and godforbid EJB stuff). All good stuff happening, and a reassurance we did a wise bet.

Today, I had a mostly enjoyable lunch with Peter, CEO of a company focussed on Embedded Linux, a smart and like-minded guy who works in an entirely different market as we do, but also cares a lot about open source and ethic business models. Meeting Peter is like facing a mirror when doing a speech rehearsal, his careful questions bring out the best in you, and his down-to-earth, no-nonsensical way of looking at things mixes very well with my ideas about doing business on the frontier of open source.

The theme for this lunch was his curiosity about what we do, and how we do make a living out of it. Since he's also been going through rough company periods during his time, and is prepared to speak open about it, I felt free to describe at length our past 2 years (nearly there!) of existence, and the running thread throughout our activities, AKA our business strategy. Mind you that we don't quite discern ourselves a strategic direction in much of our daily activities, but in the end, we seem to be heading somewhere, so after two years, I can now safely stand up and do the elevator pitch of Outerthought, with facts and figures to back my claims.

After lunch, there was a meeting (which Peter attended as well) with the peeps who would be organizing this new Open Source Center initiative, with the obligatory roundtable of presenting ourselves and our companies. I spoke frankly, and rather well prepared since I just had been doing the same with Peter over lunch, about the things we believe that differentiate ourselves from the rest of the pack, or should I call them our "concullegas". During the lengthy ride home, I had a phone call with someone asking whether we would be interested to do a project for his organization, and again I had to explain who we are and what we do, a polite way of saying 'no', since his enquiries didn't fit with our - ahum - business strategy.

So after several confrontations with the mirror, I'm now sitting here, and wondering why so many people apparently have issues with understanding Outerthought as a business entity, and I now feel it might be appropriate to try and jot it down, hopefully attracting some reactions.

In explaining our activities, there's two main focal points: we don't do turn-key projects for customers, and we have a very narrow technical focus.

The turn-key thing has the side-effect that we require developers to be present at the customer's side, since we will learn these internal developers how to do the job themselves, quite possibly using a toolset and/or framework that we prepared for them, accompanied by formal training or on-the-job coaching and mentoring. When Marc and I started to plan the startup of our own company, we had this presentation, which we created more for ourselves than we have ever used it for sales purposes, which had a slide quoting a very old proverb: we prefer to teach people how to fish, rather than feeding them with canned seafood. And this still very much is true, as a matter of fact it is becoming something we explain very early-on in our sales talks. As of now, having developers at the customer's side has turned into a requirement rather than a nice-to-have. This has three main reasons: 1) we suck at the business-side of things, since we are really hardcore technical geeks in our domain, so we tend to make abstractions of the business domain anyhow, 2) it is in the interest of the customer, as a radical implementation of the OnsiteCustomer practice known from XP, and it is the only guarantee they will effectively become the owner and master of the project after we quit, and 3) in order to provide our customers with deep-technical advise, we need to expose ourselves as much as possible to the communities where such knowledge is grown and cultivated (community mailing lists, open source projects, etc...)

The knowledge transfer bit we are doing really is a side-effect of the fundamental difference between business developers and technology developers. We are very much technology-driven, where as a developer sitting at the customer's side has the customer's business to support as his highest priority, eventually reducing the amount of time he can spend on exploring technological ways of addressing the business requirements.

The other thing, and my personal pet peeve, is FOCUS. There's a lot of peeps out there who carry a CV listing a gazillion of frameworks they are (vaguely) acquainted with, and plenty of companies who will support or implement every framework the customer asks for. We realized that, with the three of us, we can only focus on one technological domain, which we now shortly describe as the crossing of Java and XML used in the context of building web applications, more specifically using Apache Cocoon. Which means the list of things we don't do is very long: EJB, Struts, JMS, O/R tools, WebWork, Tapestry and many, many others. Does that mean we are dumb or lazy, compared with others? I would beg to differ, since 1) we actually are quite smart ;-) , and 2) we believe being smart is all about focus. We can confidently state that we know Cocoon inside-out, and that we are able to do many, also non-obvious things with it. We know for sure we could do the same with Struts, and we know that we are able to learn and understand EJB upto the same level, but alas, since we're only three, we prefer to be focussed and very good in one thing, than being average in lots of things. We get our versatility because of the depth of our Cocoon knowledge: while some other platform could be equally well suited (or better) for a certain task, we actually know much faster how we could implement this inside or using Cocoon. And the fact that Cocoon really is a Swiss army knife, and comes - in pure Python style - with batteries included, sure helps a lot with that. So while diversity is required to win the rat race, we believe we should look for this diversity inside our knowledge about Cocoon, rather than shifting away to another tool or framework just because it sells better. We know we can do the job, while really understanding what we are doing, better, faster and more robust using our singular framework offer. Our focus is good, also for the customer: 'better' and 'faster' are really important to him.

OK. So much for mirror-gazing tonight, I'll get some sleep and maybe add onto this later on. BTW: we are lazy, of course, but laziness is a virtue. Hm. I must be a geek, spending another 10 minutes in researching that link. But let's face it: mr. Wall has some spiritualizing ideas. If only the syntax wasn't like modem noise.

September 17, 2003
Fighting DNS squatting with open source licenses
After browsing the Slashdot thread (on level '5', of course, to minimize the peanut gallery murmurs, and seeing many people posting technical hacks for various DNS server implementations, I started wondering about a stronger way to fight this kind of power abuse in the future.

I believe it's safe to assume that companies who are abusing patent law, who illegally redistribute non-source versions of patched GPL-licensed codebases, or who are playing dirty by technical tricks like Verisign is now doing, are all using several open source projects within the core of their operations.

What would happen if several popular open source licenses add exclusion clauses to their licenses, explicitly forbidding the use of these projects (like Bind, Debian, Python, sendmail/qmail, Apache) by such companies? Of course, this might be in conflict with clause 6 in the Open Source license definition, but if they don't play by the rules, why should we?

September 15, 2003
ApacheCon reservations are open
... and I'll be going myself. As in: I'll be doing the "Introducing Apache Cocoon" talk, basically "just introducing" Stefano who's presenting immediately after me. We'll be seeing each other during "CocoonCon" (the GetTogether), so we'll have plenty of time and opportunity to sync slides and presentation ideas, in order to make this into a two-hour Cocoon immersion bath. I'd love to see our presentations flow fluently one into another. Plenty of other Cocoon talks as well, and many other presentations that I'll be happy to attend. Just have a look for yourself over here: http://www.apachecon.com/.

Coming out: apart from my honeymoon trip somewhere in the Caribbean, I've never went transatlantic, let alone visit the US. So I'm dead-curious to find out about the other side of the pond.

September 14, 2003
Less is more?
I'm considering switching again: to PyBlosxom this time, moving away from MovableType. While MovableType is nice, it's server-centricity and its application language are sometimes an annoyance for me. I haven't made up my mind yet, but I'm thinking...
September 12, 2003
[OT] Recipe for a nice evening with your lover
September 11, 2003
Finishing some threads
Remember my issues with my old Compay Evo n600c laptop? Well, after many phonecalls with Compaq, of which one of them was particularly nasty (I want to talk to your supervisor, and I want a service engineer here tomorrow morning with spare parts for everything), I now am the proud owner of a non-factory-refurbished laptop. The shell is the same, the screen is still the same, but it has a brand-new motherboard (after three other tries with so-called repaired ones), a new hard drive and a new touchpad.

In the mean time, I've shipped my faulty hard drive to some data recovery company in the UK, which should now be working on recovering my 18Gig of data. Just FYI, this comes with a quote and an estimated cost of repair (upon recovery of 100% of the data) of £595. If they fail to recover part of the data (which is to be suspected since the media surface was somehow damaged as described in their preliminary analysis report), that fee should be proportionally reduced. The hairy bit is that I needed to sign off a sheet which stated that they could not be held responsible for any (additional) loss of data due to the recovery process, so if they mess up, I don't get to pay, but it won't be easy to ship the drive to another recovery shop. So it's an all or nothing operation. After spending some splendid F2F time with my new TiBook, I'm also slowly realizing how much data I lost, so I'm actually quite eager to hear from the recovery shop.

On the TiBook front, after the initial falling-in-love, I'm beginning to find some (mostly minor) annoyances as well. I really miss the Alt-Tab switching I've grown accustomed to, but IIRC that should be coming back in the forthcoming Panther edition. The fact that I need to install a (free) application to minimize all apps and show the desktop and that keyboard mappings for "selecting the adjacent word" aren't consistent at all between apps is troublesome at best. There must be some logical explanation for the difference between the Apple key and the Option key (or is the Command key?), but the logic evades me at the moment. Shipping an Azerty keyboard without labeling the keys needed for | (Alt-Shift-L), {} (Alt-5 and -°), [] (Alt-Shift-5 and -°) and ~ (Alt-N) seems very awkward for an OS which brags about having Unix underneath a nice GUI... ever tried bash without the pipe symbol, or Ant without curly braces?

But these are all minor gripes, and habits of the past I need to break free from: this is what switching is all about. My main problem today was finding some decent graphics app, similar to JASC's PaintShopPro on Windows. I tried installing the trial version of Fireworks MX 2004, but it bombed out (well, not literally) upon program execution - nearly trashing the OS along its way. So I'm not motivated at all to shell out 299$ for the real version, if I can't get the trial working as a test. I looked at Canvas, but it seems pretty boring and more destined towards technical drawing. So if anyone knows a decent, not too expensive graphics program for Mac OS X, please yell.

September 09, 2003
Hi Peter
I discovered Peter in my referrer logs a few minutes ago. Peter is one of the few people I actually miss having around in our offices - just like Rik. Besides being my personal materialization of a UML guru, he also paints. While working for the same company, we spent some wonderful (as in: spiritualizing) evenings together, talking about people, relations, and all the stuff that keeps us drifting away from pure Darwinistic behavior.

Welcome, Peter. Now go off and make sure we have an RSS feed to subscribe to. ;-)

The BeJug saga: part II
I often laugh at the predictability of human beings, including myself. But this time, it's the kind of hysterical laughter when someone becomes involved in a powerplay that surely ain't his own style. So here we go for another episode in the interesting sage of one of the World's top 25 JUGs, i.e. BeJug.

During the summer, I was contacted by someone who clearly wasn't aware of all this and more, and which invited me to do the same presentation about Cocoon which I gave at XML Europe. Since I was on holiday, playing with the kids when the phone rang, and I felt rather coy about the my so-called troubled past with BeJug, I basically reacted very enthusiastic and send him a talk description right away.

I'm not sure whether it came as a surprise, but this morning, I got the following email from the guy who originally contacted me (Dutch original):

Dag Steven,

Vorige week is er een steering vergadering geweest omtrent JavaPolis en
zijn de sprekers voorgesteld voor 3 en 4 december.

Blijkt dat U in het verleden een conflict heeft gehad met Stephan
Janssen (voorzitter Bejug) en dat wij daarom omtrent uw presentatie een
veto hebben gekregen.

Spijtig genoeg zal uw presentatie op 4 december dus niet kunnen
doorgaan.

which literally translates into:

Hi Steven,

last week there was a steering board meeting about JavaPolis, and the
speakers for the 3rd and 4th of December were presented.

It appears that you had a conflict in the past with Stephan Janssen
(chairman BeJug), and that we received a veto for your presentation
because of that.

I'm sorry to say that your presentation on the 4th of December is
cancelled.

Note: I'm legally allowed to put up this email over here, since there is no disclaimer concerning privacy in the message body, and this website is my own.

Note: sorry about the funny English, but this is a very literal translation, and the Dutch version was equally awkward. But I won't shoot the messenger.

Now a little lesson about business ethics. When we are asked about peeps who know their J2EE stuff in Belgium, with a focus on EJB, Swing and JMS, we wholeheartedly have been recommending JCS in the past, even after the little quarrel with its CEO about "his JUG" (yep, the same Stephan). It's not because of this little quarrel that I should question the technical capabilities of JCS. People are multi-faceted beings, and I believe it's perfectly OK to disagree on one thing, while agreeing on another. What I perceive here is being declared persona non grata in the areas where Stephan can coerce his personal opinion upon others. Because of the size of the Java community in Belgium, I'm also pretty sure this won't be the last time.

A sad constatation. Hopefully, someone else, equally embedded in the Cocoon community, will be taking over my presentation. It would be even more sad if the subject is dropped off the agenda.

September 07, 2003
OpenOffice on Mac OSX: pain.
I know there's a fair amount of warning that the current Mac OSX version of OpenOffice is for 'Unix-savvy' users, but I'm afraid Unix-savvy means something different to me than to the OpenOffice team. Unix-savvy may mean 'with arcane settings you need to carefully adjust in some startup script', but in OO, it just means butt-ugly. OK, it's a really an X application rather than a normal OSX Carbon/Cocoa .app, so I should learn to live with messy fonts, strange keybindings, and starting up X when I just want to open a Word doc, but knowing that this will be it until Q1 2006 feels kinda left-alone. Yep, they currently plan to stop working on the Mac support in OO until some new foundation thingy is done, making a native Mac port easier (which, I assume, is the same kind of cross-platform-porting insulation framework sitting underneath Mozilla).

Some latenight-FUD-thinking of my own: OO is still very much tied into Sun. Being pissed off by the defeat of NetBeans against Eclipse (IBM), NetBeans being Sun's alternative for the much-touted MS development IDE (Visual something), they really want to make sure they effectively own the Open-Contender-against-MS-Office space. Due to Sun's persistent backward thinking with regards to their core business and competitors, with the success of Mac OSX and all these geeks buying TiBooks, they are worried about the future of the Sparc/Solaris combo, so they want to go after its contenders. XServe/MacOSX might be a contender for Sparc/Solaris in some arcane regions of the world (like graphic shops), so they decide not to support the MacOSX peeps anymore by giving them a nice port of OO. Hm... dubious reasoning for sure, which won't help much to alleviate my current I-need-to-work-with-office-documents pain. Should I purchase that $500 Office X for Mac license from MS? Or just live with the peculiarities of OO running under X11 for Mac? Who knows. I don't.

September 05, 2003
GetTogether 2003 registrations are running
On a Friday evening hacking spree, I added an attendee counter to the frontpage of the Cocoon GT2003 website. Last year, the counter ended at 111 people coming from 11 different countries. Let's see if we can double that this year. :-)

In case you want this counter to display on your website, I can provide you with a URL which emits the tiny HTML snippet of that table. Uh. Should I call this a WebService now?

September 04, 2003
Cocoon GetTogether 2003 website's up
Some late-night hacking sessions and my first endeavors to Go with the Flow resulted in the Cocoon GetTogether 2003 event site being up. Come and check us out!

The Cocoon GetTogether is an annual gathering of Cocoon enthusiasts, which was previously an outgrown birthday party of my company, but now has turned into an Orixo event.

Plenty of interesting people are coming over, so if you want to get into the Cocoon vibe, don't hesitate to sign up - it's a very low-cost thingy.

September 01, 2003
Java interviews and cropcircles
Some people are so much out of this world. Tom has a funny story about his blog on Java job interview questions, and I have been astonished by this sequence of comments on a very old blog of mine. In the end, these guys were about to freak me out, so I basically closed the comment ability for that blog entry.