Capistrano deployment for Jetty-Rails

As I have developed the (soon to be available) UploadForMe application with Jruby on Rails, I have had to look at the best way to deploy it in production. In development mode, I was just running the script/server with Jruby, but I needed something better for production.

I had first tested the Glassfish approach of running an Java application server. And although it isn't hard at all, as illustrated by Charles Nutter in his post correctly titled "Zero to Production in 15 Minutes", it wasn't practical for me: just running Glasfish consumed 300MB of RAM and impossible to run on my development machine. So I had to look somewhere else.

That's how I found Jetty-Rails.The way to run your JRuby on rails apps is so simple I thought their documentation was lacking! But it is really as simple as installing the jetty_rails gem!

The only missing piece were recipes and scripts to start, stop and restart the application when deploying with Capistrano. It was a simple matter of writing 2 scripts and the Capistrano recipe, with some experimentation in between I must admit :-)

Anyway, here are the scripts. The pid of the running jerry_rails process is stored in the file shared/tmp/pids/jetty_rails.pid.

script/spin
<br />#!/bin/sh<br />. script/jetty_config<br />TIMESTAMP=$(date +%Y%m%d-%h:%m:%s)<br />#check that previous process has been stopped correctly<br />if [[ -f $BASE_DIR/shared/tmp/pids/jetty_rails.pid ]] && ps ax | grep -f $BASE_DIR/shared/tmp/pids/jetty_rails.pid >/dev/null 2>&1 ; then<br /> echo "We could have the previous process still running! $(cat $PID_FILE )" >&1<br /> exit 1<br />fi<br /><br />cd $BASE_DIR/current<br />$JRUBY -S jetty_rails -e production > log/$TIMESTAMP_jetty_rails.log &<br />echo $! > $BASE_DIR/shared/tmp/pids/jetty_rails.pid<br />



script/stop_jetty
<br /><br />#!/bin/bash<br />. script/jetty_config<br /><br />#check that we talk about a jetty_rails process<br />if [[ -f $PID_FILE ]] && (ps ax | grep -f $PID_FILE | grep jetty_rails > /dev/null) ; then<br /> echo "Killing process $(cat $PID_FILE)"<br />else<br /> echo "This doesn't seem to be a jetty_rails process! See PID: $(cat $PID_FILE) " >&1<br /> exit 1<br />fi<br /><br />kill $(<$PID_FILE)<br />rm $PID_FILE<br />

There's a config file shared by these 2 scripts, put in script/jetty_config:
<br />#!/bin/bash<br /># --- START SETTINGS TO EDIT ---<br />#add the JRE bin directory to path<br />PATH=$PATH:/usr/local/jre/bin/<br />#Base directory for deployment<br />BASE_DIR=/usr/local/jrailsapps/uploadforme<br />#Jruby executable to use<br />JRUBY=/usr/local/jruby/bin/jruby <br /># --- END SETTINGS TO EDIT ---<br /><br />PID_FILE=$BASE_DIR/shared/tmp/pids/jetty_rails.pid<br />

Put this in config/deploy.rb:
<br />namespace :deploy do<br /> desc "Restarting jetty_rails"<br /> task :restart, :roles => :app, :except => { :no_release => true } do<br /> stop <br /> start<br /> end<br /> desc "Stopping jetty_rails"<br /> task :stop, :roles => :app, :except => { :no_release => true } do<br /> run "cd #{current_path} && script/stop_jetty", :pty => true<br /> end<br /> desc "Starting rails app with jetty_rails"<br /> task :start, :roles => :app do<br /> run "cd #{current_path} && #{try_runner} nohup script/spin", :pty => true<br /> end<br />end<br />

Notes

The start recipe is overwritten to add the option :pty => true. Without that option the nohup command isn't effective.
Using the nohup command is needed because there's currently no way to send the process to the background as a daemon, nor to store the PID in a specific file. These scripts do both.

Apprenez le flamand!

Ca vous permettra d'écouter Radio 1, qui vous propose des

• interviews et interventions en VO (exemple d'aujourd'hui: réaction de Barack Obama en VO, avec explication par après pour ceux ne comprenant l'anglais)
• sujets variés, avec invités spécialistes, couverts rapidement, sans chichi
• les journalistes qui posent assez souvent les questions ennuyantes
• des infos routes précises et complètes
• de la musique variée (parfois même francophone)
  

au lieu de (sur La Première, euh, pardon, La Première, accéder au site sans www devant le nom de domaine ne fonctionne pas....)

• des interviews de personalités non francophones doublées, ce qui a parfois pour résultat que vous ne comprenez ni l'original, ni le doublage.
  
• sujets couverts trop longuement (émission du matin)
  
• à 18h un journaliste soporifique qui parlent plus que ses invités (avec tout le respect pour ses compétences journalistiques)
• un taux de ratage de lancement de sujets pré-enregistrés phénoménal
• infos-routes qui durent 20 secondes, la moitié de ce temps étant pris pour dire "et des files à tous les points habituels"
• la méteo contradictoire prévoyant tout et son contraire (je n'ai pas pris note, mais cet exemple inventé est inspiré d'un billet réel: "il fera bon toute la journée aujourd'hui avec un ciel dégagé.......blablabal.... et le retour des éclaircies est prévu en fin de soirée")

Bon, tout n'est pas parfait non plus à la BRT VRT. Par exemple les journalistes ne posent pas toujours les questions ennuyantes, notament je pense en ce qui concerne les problèmes communautaires. Mais ne m'intéressant plus à ce sujet depuis belle lurette, je zappe quand on en parle à l'antenne, quelle que soit la radio.

Un conseille donc: si vous voulez obtenir une information variée, mondiale, et présentée de manière professionnelle, écoutez Radio 1 de la VRT.

Message se basant sur les plages horaires 7h-9h et 16h-19h.

PCLI/VAPZ: which one to choose?

The end of the year is nearing, and it's time for me, and maybe all belgian starters, to take a look at what your income will be and optimise your fiscal situation.

First thing: if you have income higher than 11000€ for the year and you decided to pay the minimum social security contribution of 600€, you will have a regularisation, in the sense that you will have to pay you social contribution on your revenue above 11000€. The trick is that those social contributions are deductible from the revenue you have to put in your tax declaration. But for that to apply, you have to make the payment this year! (here's some more info: fr, nl).

If you started after 30 march, also note that this year is not full, you have only 3 quarters. It means that the definitive social contributions you will have to pay will be based on your revenue of next year. For the social contribution, you better make profit this year than next. But remember that more profit this year mean you'll pay more taxes too. So honestly, I'm not sure how this all combines. Any hint?

And finally, the subject of this post: as an indépendant/zelfstandige you can subscribe to a additional retirement fund: FAPZ (in dutch)/PCLI (in french). You can contribute up to 8.17% of your revenue, with a maximum of 2686.05€ yearly. And it's 100% deductible from your revenue!

You'll see there is a social vapz/pcli, which includes a guaranteed revenue insurance, but it's not that interesting I've heard: the part of your payment going to the retirement fund is lower, as you pay your guaranteed revenue insurance of course.

Now my question: any advice on where I should take my vapz/pcli? Are there big differences between the providers? I'm currently at Acerta, which works with KBC. Any opinion on their offering?

Furthermore, any further advice on optimisation of revenue for independant workers?

Ever needed to schedule a mail in the future?

It regularly happens that I want to schedule a mail to be sent at a certain time in the future. The ideal tool would also let me invite others to subscribe to this one time mailing, for example to remind people of the next Profoss, even if they don't want to subscribe to the newsletter.

If oyu don't find it, build it they say. Well, here it is: Taktiki.com.

And it's simple: no authentication is needed: created you mail, validate the sender address, and add recipients or invite people to subscribe to that one-time mail. For example, to get a reminder on monday for Profoss' ERP event on tuesday, you can subscribe here

In the hope this can help other people out there!

PS: for those who wonder: yes, that's the best I can do in web design. And that's with the help of YUI's CSS and grid help

ERP event nearing!

If you're interested in Open Source ERP systems, be sure to attend the next Profoss event on 18 november!

I'm excited by the interesting content that will be presented: Adempiere, Openbravo, OpenERP will have a presentation, a method for evaluating open source ERPs will also be presented.

Registration is still open, but don't wait too long to register!

google.com/ads in my dreams?

Did I dream that there was an URL for Google Adwords and Adsense at www.google.com/ads? 'cause now all I get is an error page:

Profoss on twitter

I've not been a Twitter user until now, so I'm curious what impact this will have.
Follow Profoss at http://twitter.com/profoss

Ruby flexibility to the rescue

In the development of UploadForMe, I had a chain of methods called, each of which possibly returning nil. I really didn't want to check at each step if the returned value was nil, as I would have ended up with something like

if (o!=nil and o.first_call!=nil and o.first_call.second_call!=nil)

I quickly created a class which would help me get around this problem, and I think (I haven't checked further, honnestly) this goes in the direction of monads, which are well known in Haskell.

The solution is to have a class of which an instance is returned in place of nil. This instance responds to all methods by returning an instance of the same class, in turn responding to all methods. However such an instance responds to the method #nil? with true.

class SimpleMonad
  def initialize(h={:is_nil => true})
    @is_nil = h[:is_nil]
  end

  def method_missing(m, *args)
    return SimpleMonad.new
  end
  def nil?
    @is_nil
  end
end

And I'm using this class as return value in a class derived from Hash. method_missing returns either the value in the hash for which the key is the method name just called, or an instance of the class above:

class ServicesKeys < Hash
  def method_missing(m, *args)
    if keys.include?(m)
      return self[m]
    else
      return SimpleMonad.new
    end
  end
end

That way, we can chain calls as we want:

 #define the object we will work on
 h = ServicesKeys[  :uploadforme => ServicesKeys[:login => 'rb', :username => 'rb']]

 #we can easily access the values in there:
 h.uploadforme.login
=> "rb"
 #if we try to access values for an undefined service, we don't get in trouble:
h.myowndb.login.nil?
=> true

You have to be cautious in naming your keys to avoid clashes, but in my case this is an acceptable trade-off.

There might also be better and more advanced solutions out there. Don't hesitate to mention them in the comments.

Why use Chrome when you have Arora?

When Google's Chrome came out, I was rather surprised by the move, and was eager to try it. But when I saw it was Windows only, I andandoned the idea even if I could have tried with Wine: I couldn't see any good reason to use a non-native browser. Reading Kris's blog post about his test of Chrome on wine only confirmed my opinion.

I was interested to test a browser with the Webkit engine though, and was surprised there wasn't an open source browser based on Webkit available on my platform of choice (GNU/Linux). The Epiphany developers had announced a switch to webkit some time ago, but it still isn't effective it seems.

A search on the internet returned Midori, build with GTK2. I'm not a fan of GTK, and I was surprised there was no Qt based browser, as Webkit is now integrated to Qt (from Qt4.4). Finally, this week I discovered Arora, which was originally a demo browser integrated to Qt4.4.

Installing Arora was easy once I had installed Qt4.4 from sources. I had originally tried with a Qt4.5 snapshot through git, but it wouldn't compile. Compiling Arora itself was so fast I thought I had forgotten to do something.

And Arora is fast too. Using it is also a good experience (eg I'm writing this blog post with Arora).

I don't think it will replace Firefox as my primary browser just yet, but it will take an important place on my desktop as I expect it to become the fast browser without any plugins or extensions I'll keep running in the corner to access the web-based tools I run continuously. And yes, MyOwnDB runs great in Arora!

LuckyAs.Me siteof the day: learn german history

Looking for Germany, I got a page on the Eurodocs website, from the Brigham Young University, Provo, Utah, USA, documenting European history. IT is mainly a collection of links to resources on the web.

AntWeb: LuckyAs.Me site of the day

As a result for a search for Belgium on LuckyAs.Me, I got redirected to a website dedicated to ants: AntWeb.

There I learned that we have (had) 86 species of ant in Belgium (of which 9 haven't been spotted the last 50 years).

Using search for fun

Why always be limited to the 10 first matches? www.luckyas.me is built on YAHOO boss and redirects you to a random site that is in the first 100 results returned by YAHOO for your search.

Maybe not useful, but can be fun when you have some time to change your mind. I discovered there's a Bauduin Preschool, a police inpector whose family name is Boulette, there's a restaurant called La Frite in the US, but also a mention of the Profoss OpenOffice event on BoycottNovell.

Have fun and let me know what you think of it.

When you're clueless... (2)

you should really shut up.

Sometimes I think of those talented artists who don't get their record published, while totally untalented people get selected in a reality show to have an album published, only to flop because they're so bad.

Now I think of those talented bloggers, who try to live from that activity, and who see such crap published on CNet, who paid its author.

As a reminder, here's the first post in what has just become a serie.

Disabling wifi on the Eee PC 701

I have installed the normal Ubuntu 8.04 distribution on my EeeP, applying the usual fixes.

I still had problems with the Wifi hotkey: the ath0 interface was still up and consuming power. I finally set it up this way:

In /etc/acpi/events/eee-wifi-off and /etc/acpi/events/eee-wifi-on, use the wireless.sh script provided with ubuntu:

action=/etc/acpi/wireless.sh

Then edit /etc/acpi/wireless.sh and add these lines before the call to toggleAllWirelessStates:

if [[ -d /sys/class/net/ath0 ]]; then
   /etc/acpi/eee-wifi-on-off.sh on
fi

After a reboot, I could totally disable the wifi interface: it isn't returned by ifconfig ath0 and the wifi led is out. I haven't had the time to check battery life without wifi, but it should help.

Now I still have to look at the configuration to get a notification when battery capacity is getting low....

Uploading is frustrating. Here's a solution.

So you're back from holyday, and you took a lot of pictures and you want to put them online or print them. But you have 2 gigs of them. I've been through that experience, and really hate it: my DSL connection is hijacked by the upload, the upload could break and then you have to look at which pics where uploaded, and which weren't. In short, quite a pain.

And it's from that painful experience that I've had the idea to start a new service: UploadForMe.com. Send us your data on DVD, and we'll bring it online either on our servers from where it can be imported into other services, or we'll import it directly into online services with which we have collaboration agreements (we don't want to know your login information!)

There are also offerings targetted to professionals in the works.

It was floating in my head since some time now, and when I started to talk about the idea, I was surprised of the interest. It's so unsexy and un-web2.0....

But hey, there are anterior examples, notably Netflix who built an empire on a snail mail based business.

I'm really looking forward to developing this, and am eager to read your reactions and suggestions, even negative (and I'm sure there will be). If you want to be part of the test phase starting in september, just let me know.

What's in a name?

A lot!

And this library must be really great stuff! :-)

Some things changes, some don't

This quote I remember from the movie Matrix Reloaded is really the best I could find to describe this post, even if it will mostly focus on something changing: Microsoft's stance towards Free and Open Source Software.

Of course, years spent thrashing FOSS and comparing the GPL to cancer can't be forgotten that rapidly. Old reflexes remain. But it seems clear things are changing.

And I was very surprised to experience it myself when I heard from Dirk Tombeur that Microsoft will be the Gold sponsor of the ERP event by Profoss. I had already met Dirk at Profoss events, but I saw his presence as a need for Microsoft to keep an eye on the competition. Becoming the Gold Sponsor of a Profoss event is clearly another approach.

It seems Microsoft is changing, and starts to understand that the IT world is more heterogeneous than it would have liked some years ago. Keeping its stance of some years ago can only lead it to irrelevance.

Of course Microsoft is a company, and as such is always looking to generate revenue. But now, it seems to be interested to also generate revenue with FOSS. They're working on IronRuby, worked on PHP performance on Windows, submitted 2 licenses to OSI, became Apache sponsor, ....

But let's hope this is only the beginning. There's still a lot Microsoft should do for Open Source.....

Before I finish writing this post, let me quicly talk about something that is not changing: Profoss' focus on informing objectively about the use of Free and Open Source Software in professional environments.

Really, some things change, some don't.....

Best backup solution?

I've recently set up a new server, and needed a quick backup solution. I made my own shell scripts that do exactly what I need (collect file, dump databases, encrypt file, transfer them), and put them on github in case it can help anyone.

However, there is certainly an easy to configure tool out there to set up backups. Any advice?

How do Amanda, Bacula and other similar tools do these days?

Taking the Erlang plunge

It's decided: I will now learn Erlang. I've kept an eye on it for some time, and things seems to move in interesting ways. Take Vertebra and Github's upcoming git daemon as interesting signs.

I found it great to join the Linux and Ruby communities before their popularity exploded. I feel like this could be the third time, but I'll have a better opinion after reading the book I just ordered.

Potverdekke, it's a shame to be a belgian

An economic crisis is looming, a financial crisis could be coming, more and more people have difficulties to pay for all their daily costs, the petrol price is sky rocketing.

And what happens in Belgium? The government falls on communautaire problems.

Shame on you, belgian politicians!

Updating rubygems on debian

Note to myself:

if you want to update the rubygems package installed by Debian, edit /usr/bin/gem and add those two lines after "require 'rubygems'":

require 'rubygems/gem_runner'
require 'rubygems/open-uri'

then you can update ruby gems with

gem update --system

Here's why Google is the search leader

Today I read an article about Yahoo's BOSS: Build you Own Search System, but it didn't specify a URL for the new Yahoo offering. So I typed in the Firefox search box the words "yahoo boss", and got the links as the first item after the News result displayed at the top. Quite effective!






But then I thought: I'm looking for a Yahoo service, I might also make a search on Yahoo's website. Google is the default search engine for a lot of people, including me. Maybe I should take a look around, for example at Yahoo search!
Honnestly I was surprise to not see a single link about it on the first page. So I looked on the following pages and didn't even see it in the first 5 pages. I abandoned looking further.





This awakened my interest: Would Microsoft do better? Not at all! Nothing in the first 5 pages.




Ask.com maybe? Good try, but no.



Microsoft might be interested in buying Yahoo's search business, but mixing 2 irrelevant search results stays irrelevant. They'd better work hard to deliver better search results. At that time I would possibly change Firefox's search box' default search engine. That time has not come yet though.

version 0.4 of YUI slideshow

I had developed this YUI-based slideshow to use it on the MyOwnDB.com page, and published the code in 2006, already 2 years ago. It's been used by some people, but none had feature requests, until yesterday :-)

So today I added a back button. "Wow, that's a feature!" I hear you say.
Well, the slideshow was meant to be non-interactive in the first place: you give it images, html, divs to rotate and it'll do it for you automatically. So it was not in the normal workflow of the slideshow to have a back button.

Anyway, I added it and that was enough for making a new release: version 0.4 is out!

Next Profoss: ERP software

The next Profoss event will talk be about open source ERP software. Here's a list of software I know and that I will possibly contact.

• http://www.xtuple.com/
• OpenERP
• Compiere
  
• OpenBravo
• http://www.erp5.org/
• http://www.opentaps.org/
• Apache Ofbiz
• https://www.jfire.org/,

Fabien from OpenERP already expressed his interest. If you have any suggestion or if you can help me get in tough with good speakers about these software, let me know!

Raph

Planet Profoss launched!

I have finally launched a Planet feed aggregation at http://planet.profoss.eu. I say finally not because it was long and hard to set it up, but because I wonder how did not think of it earlier.

Included in the Planet are blogs and websites that cover professional open source.

Currently in english, I am interested in publishing planets in other languages, especially dutch and french.

If you want to be included, let me know (by mail or leave a comment here). The goal is to only include posts related to professional open source, and avoid any posts unrelated to that subject. It should be achieved easily with the help of tags and categories though.

Profoss event round up

Yesterday's Profoss event about OpenOffice.org went pretty well: speakers where happy, but above all, attendants were happy of the event. Feedback forms were all positive.

This was the first half-day event organised by Profoss, and the conclusion is that it's a good format. You can expect more of these ;-)

And as always, absent people are wrong. No company sponsored the event, and Sun wasn't even present. So when the question was asked: " Which company can we contact to get help and support in an OOo migration", Machtelt could just name them, but how convincing is that when they're not event at this event?

Microsoft, to the contrary, was represented, and raised some interesting issues to get an interesting debate.

Search engines: some interesting stuff this week!

First, there was the wikia search results that became editable. I tested it when announced and it was terribly slow. I tried it again and speed is much better now. But I'm still not a fan....

But now, there's Yahoo's Search gallery, which let you add modules to enhance the results returned. There are quite some modules available enhancing results from Flickr, Firefox addons, LinkedIn profiles, Ruby RDoc, LastFM, etc
Speed seems absolutely OK, and you can even develop your own enhancements.

Let's hope they'll manage to stay independent, as Yahoo is bringing some interesting stuff. And don't forget about YUI, Zimbra, Pipes, Hadoop, ...

Wanna compete with Amazon's EC2?

I just discovered there's open source software available to build such a service and it's called Eucalyptus.

Now all you miss is the datacenter :-)

Server alias superfluous IV

A very good one spotted by Serge:

http://www.bebat.be gives you the good websites, but .....
http://bebat.be redirects you to http://www.necho.com/

I wonder what's worse. Rendering a wrong website, or returning an error.... I might go for the error....

Usual reactions to this are "Goe bezig" in dutch, and "Excellent" in french :-D

Are vendors really pushing OpenOffice.org?

The Profoss event on OpenOffice.org is taking place on 10th june, and the final list of speakers and schedule are now available.

Until now, the interest for the event from potential visitors is comparable to other Profoss events organised. But what is different, is the lack of interest from commercial OpenOffice.org proponents to actively support and sponsor this event.

The format of this event is different (one afternoon rather than 2 days), but I don't think this is the explanation.

With Sun being the main OOo contributor and a Profoss partner, I expected interest and active support from them, but only David helped me.

Novell is also a big OOo developer, but I never ever got an answer from them, though the person I contacted was advised to me by Michael Meeks, an OOo developer.

IBM didn't show interest either. Their marketing department seem to be very picky, and more interested in marketing show than really informative events.

Even smaller companies labeled as OOo specialists didn't come back to me.

Luckily, registration numbers are currently similar to other events.

So, what I wonder is: despite all the noise around ODF and OOo, is there real interest from those companies and other FOSS companies to push OpenOffice.org to the enterprise market? Or do they think the majority of companies are not ready for the switch? Or, worse, do they think OOo is not ready for wide adoption in the enterprise?

Lots of questions, let me know if you have answers.....

Vim: Shortest match of your search pattern

I just discovered that if a "-" appears immediately after the "{", then a shortest match first algorithm is used. In particular, "\{-}" is the same as "*" but uses the shortest match first algorithm.

An example from http://www.ph.unimelb.edu.au/~ssk/vim/pattern.html:

a[bc]\{-}[cd] matches "abc" in "abcd"
a[bc]*[cd] matches "abcd" in "abcd"

Note however that a match that starts earlier is preferred over a shorter match:

"a\{-}b" matches "aaab" in "xaaab".

Lemon POS solution in Profoss directory

I just added the Lemon Point of Sale solution in the Profoss Software Directory (in english, french and dutch). Still a young product, but quite interesting.

Bye bye EC2, Hello Gandi Hosting!

I have been using EC2 for some time now, for example hosting the Profoss news website, the Profoss newsletter, and other personal tools. But it was kind of expensive for what I used, so I was looking at alternatives, but as I had not time to tackle it, it was left as an idea.

But now my EC2 instance became unavailable. And for those who don't know, an EC2 instance has no persistent storage, meaning that a reboot leaves you with a blank instance. Which makes backup even more important.

As I needed to restore everything from my backups, I decided to go the Gandi Hosting service. They partition their servers in 64 shares, and (currently) you can decide to use between 1 and 32 of them. Ok, one share is not very powerful hardware: 256 MB of RAM (+512 MB of swap), 5 GB of RAID6 disk space reserved for your data (the system is on another disk). Network bandwidth stands at 5 Mbits guaranteed. But one share only costs.... 6€! And you can start small and add shares as you need them.

So I decided to take 1 share, and see how things go. If I need more power, I'll add a share, and I'll pay 12€ a month. But honestly, with the tools running on my share, it could be some time before I take that step.

Profoss speakers: interesting as usual

Profoss tries hard to get interesting speakers with real world experience at their events, and this is also true for the next event about OpenOffice.org.

Are currently confirmed:

• Eric Descamps, the manager at the Belgian Post for the OpenOffice.org project, which went through a Proof of Concept phase, and is getting at the deployment phase. Eric will share his experience and lessons learned in the project he leads.
  
• Machtelt Garrels, co-founder of the OpenDoc Society, will give a talk on how to best avoid usual annoyances met by people switching to OpenOffice.org.

Further speakers will be confirmed soon!

This event will take place at the International Press Center from 14:00 to 18:00, and registration is open for all interested parties!

Ubuntu really available from Dell Belgium?

There's a page talking about it, but once you want to shop, you get an error page.... I have to buy a computer for a friend, and he's interested to try Ubuntu and I thought I'd see what Dell has to offer. I also didn't find a system where Ubuntu is proposed. If that's all they have, it's not very impressive and I'll go somewhere else....

Profoss in the newspaper!

Today L'écho published an article about free and open source as an opportunity for SMEs. Profoss is mentioned as it is as founder of Profoss that I got in contact with Fabian Lacasse, the journalist who wrote the article. But that's not the most important. What is really important is the quality of the article. It even reminds readers that "gratuité" is not what characterizes free and open source software! Of course (however,... that's not a given in today's journalism), it reflects what I explained, but the writer managed to keep the substance while making it understandable to non-IT people. This is really an article of a quality I wished we saw more often!


But this is only one more step in reaching Profoss' goal to inform professional users of the viability of free and open source software. There's still lots more to do, so don't hesitate to spread the word!

Profoss directories launch!

I've been so busy I didn't blog for too long. It's not that I haven't anything to report: BarCamp Gent was great, MyOwnDB is progressing well, Profoss OpenOffice.org in june was announced, I was without internet 2 work days (and that's really a PITA!), etc...

But what I wanted to report in this post is that I've just launched the Profoss Directories (currently mostly in french, see why below) with 2 categories: Professionals listing companies and freelance specialising in FOSS, and Software, listing FOSS commonly used in professional environments.

I'm still looking to complete these 2 categories and you can help in two ways:

1. Let me know what (your company?) I should add in which category
   
2. Help me maintain and translate the entries in these categories

I'm also interested in your feedback!

And the reason it's mostly in french: these directories will be (I was told so at least) referenced in an article in L'echo. This has accelerated things and I have had to concentrate on the french version, corresponding to the audience of the newspaper.

Google Appengine: a great opportunity for MyOwnDB?

Google launched their AppEngine, a platform to build web applications easily, with automatic scaling as a benefit from the limitations put on the available features. For example, the storage engine is a schemaless database based on BigTable. Kind of like Amazon's SimpleDB or CouchDB. But what if you need an information storage letting you define a structure and schema?

You can't install mysql locally, and outbound communications can only happen through the email api or the url fetch api...

Enter MyOwnDB. With the REST API we will provide shortly, you'll be able to use MyOwnDB as your data store from within AppEngine applications. Communication from AppEngine to the outside can happen through the URL fetch API provided, supporting GET, POST, HEAD, PUT and DELETE. Exactly what's needed to fully enjoy our upcoming REST API!

Arranging my holidays as a freelance

It might seem strange that after my first week as a freelance, I already talk about holidays, but I think it's more than normal. For example, I have developed a tool to manage the production lines of a company, and if that tool stops working, they're in big trouble. If this happens when I'm on holiday, in the middle of the mountains where a GSM is completely useless and internet connection is even more utopic, they're in HUGE trouble. Although I would enjoy my hollyday, ignoring what happens, my return would be that more catastrophic.

As it is not possible to find a solid solution in the last weeks before leaving, and although no holiday is planned yet, I am already putting all pieces in place. The first one is

Documentation
If I want someone else to be on call and the point of contact of my customers, that someone should have all what is needed to also solve the possible problems, know the history with the customer, etc. The best way to achieve that is to extensively document all what is done: source code, infrastructure, requests from the customer, etc
Although source code comments should be a natural reflex, I hadn't found the tool I wanted to use for the rest. But this is now ok: I'll use Dokuwiki. It's really the perfect tool for that task: simple formatting; storage done in text files, making the content available in any text editor, and making integration with source code control tools easy; tons of plugins; file attachments; private wikis possible.

A positive side effect is also that I'll build a knowledge base I'll be able to use for future projects.

But if documentation is a necessary condition, it's not sufficient. You also need

Someone you trust
If you share all your documentation, contacts, etc, you'd better be sure that person won't short-circuit you and steal your customer. Although I want to build a relation of trust with my customers, making such a move more difficult, I'd rather work with someone I trust.

And I have the chance to know the right person! It's been some time now that I work closely with Bart from Zeropoint.it for MyOwnDB and other things. This collaboration has let us build a relation of trust, to the point that I'll contract Zeropoint to care for my customers during my holidays. Zeropoint has a team spread over Belgium and Pakistan, providing customised IT solutions in all domains you can imagine. With the good documentation I'll provide them, I'll be able to go on holiday with peaceful mind!

So if you're ever going freelance, one of the first things you should think of is your holiday! Isn't that great?

First training day... as a trainer

Today was the first day of the training about "Mastering Unix shell scripting" that I'm giving for Zeropoint.it. As a lot of fellow FOSS people, I've learned a lot by myself and I have only followed a couple of trainings, so I can't say a lot of things like "it's strange to be on the other side".

What I find the strangest thing however is to be considered the specialist. I've never considered myself a specialist, as I prefer to learn a lot of different things. A trainer however is by definition the specialist in the room, and I hope I've made honour to that title :-)

It's hard to evaluate my performance though as it's really really hard to have interaction with the attendants. It's as if they don't dare to ask a question during the class. I've made sure however they understood what was taught today and I answered all questions that surfaced along the day. I hope I'll get to know the content of their anonymous feedback form at the end of the training!

I really enjoyed it, so thanks Bart for giving me this opportunity. I could even start to think to prepare trainings I would love to give. For example about "Mastering vim". Any takers?

Google Translate API

I just saw Google published an API to their translation service. That's something that will be very useful, and the first Google API I'm really looking forward to use with a direct personal benefit!

Start of presentations day

Bart will talk about MyOwnDB at Plugg, and I'll be talking at LinuxWorld for Profoss. Thanks to Murphy those 2 talks were planned simulteanously....

Server alias superfluous III

• Huxley.be


• IBBT.be

I'm sure the IBBT one should be corrected rapidly.
See you tomorrow at Barcamp Gent hosted by..... IBBT ;-)

Which professional insurance?

Today was another milestone in my transition to freelance: I resigned from my position at Easynet. There's no turning back now!

And I need to look for a professional insurance. Imagine I'm at a customer premises, and I step on one of these power block with a power switch and I stop the central server, that doesn't want to boot up again, leading to losses for this customer.

Or there's data loss in the course of my project (this being my fault or not). Or a config change corrupts a database leading to 2 hours downtime. I'd like to be covered by an insurance. But what should I pay attention to? And how much does it cost? What do such insurances costs?

I'll have to look around, but any advice is welcome of course!

Quand "Ikben kwaad" devient "Je suis très fâché"

Un extrait d'interview d'un politicien par la VRT est passé sur la RTBF. Comme toujours, la RTBF a la bonne habitude de traduire par une vois supperposée, ce qui fait qu'on ne comprend parfois aucune des deux voix: ni l'originale, ni la traduction.

Lors de cet extrait, j'ai pu cependant comprendre le politicien commencer par "Ik ben kwaad", et le journaliste traduite par "Je suis très faché".

Un bon petit rappel qu'il est nécessaire de garder un exprit critique face au média. Si des nuances sont introduites dans une info directement vérifiable, que se passe-t-il avec les infos plus difficilement accessible?

En dit is gewoon een voorbeeld, het zelfde mag gezegd worden voor de alle media, ook de vlaamse media. Laten we dus critisch blijven tegenover informatie die wij door de media voorgeschoteld krijgen.

Going for it!

So now it's decided. I'll go as freelance.

As I was already freelance in addition to my full-time job, paperwork is very limited: I only have to change my profile at Acerta, where I am currently affiliated, but who I could leave next year as they're far from responsive....

I'll possibly take one of two profiles:

• specialist in data structuration and management with web applications, in which I have multiple years experience, as employee at Easynet, as freelance, and with MyOwnDB
• As I'm not sure I'd like to do only data management, and as I did a bit of everything in the past and still love learning new things, maybe I can propose to handle whatever problems a company encounters: database, LDAP, mail, web, networking, etc

As I am working on an EZ Publish project and enjoy it, I also might continue in that direction.

And I hope to also bring Profoss to the next level!

Looking at that, I guess you can see I have had to decide rapidly :-)
I haven't had the time to iron out all details, but getting in the action without time to ask too much question might be a good thing.

Lack of usability costs money

I needed a cartridge for my printer, and it all started well: the printer has an embedded webserver with a page displaying a button "Order supplies". I decided to give it a try.

Clicking on the button asks you if you want to send the printer information to HP so you don't have to enter printer details yourself. Great idea, there's no risk of error here!

As to spread sending costs over several cartridges, I decided to order 2 additional colors, which brings the total of 3 cartridges.

Until then, all went very well, and now HP will redirect me to an online shop, a partner of them. They actually let me choose between two shops, and I choose the cheapest. The problem is that when they transfer your order to the shop, they online transfer 1 product, and you're supposed to search and select yourself the additional products in the shop's products. Why? Didn't I already do that earlier? Why introduce a risk of error that was avoided earlier?

I gave up ordering, I'll do it another way.

HP will certainly not care a lot, I'll have to get the cartridge anyway (refills for these cartridges don't seem to be widely available at this time. The best I found was a kit with empty cartridge and refill ink, but costs 140$, or this one for the C7180. I'll keep these option for later). But the partnering shops should, as they're loosing sales. Mine at least....

Random image in Ez Publish

I'm working further with Ez, and it's an impressive tool. There's a lot to learn and discover, but it is a really powerful system. Here's how you can have a randomly chosen image displayed in your page:

{def $nodes=fetch( 'content', 'tree',hash( 'parent_node_id', 68 ,'class_filter_type','include', 'class_filter_array', array('image')))}
{def $number_of_images = count($nodes)}
{def $to_display = rand(0,$number_of_images|dec)}
{def $image = $nodes[$to_display]}
{attribute_view_gui attribute=$image.data_map.image}

Some lines and assignments might be superfluous, but were helpful when learning how to put that together. With this code in your template, one of the images stored as a child of the node with id 68 will be displayed.

One gotcha is that arithmetic operations don't work in the .tpl files, hence the use on |dec to decrement the variable $number_of_images.

Also, be sure to put this outside of all cache-block, or your image will be chosen once and for all the first time the template is compiled as it would be cached from then on!

And notice the call to attribute_view_gui to render the image.

You can do this with any other type of nodes, for example for the latest news, or customer quotes. I already see a use for the future Profoss website :-)

webdav in Ez publish on postgresql

So you're running Ez Publish 4.0 backed by a Postgresql database, and you want to use the webdav functionality. You followed the howto, but to no avail. Your webdav client does not have access (I had the message "folder or file does not exist" with Konqueror).
Here's a good advice: enable PHP logging, and take a close look to it. I found this error message for every webdav request sent:

PHP Fatal error:  Call to undefined function mysql_real_escape_string() in /var/www/ezpublish-4.0.0/lib/ezdb/classes/ezmysqldb.php on line 908

The solution is simple, just install the php-mysql package.

I hope this saves you some time :-)

Another FOSDEM success

The association "Postgresql Europe" was started during FOSDEM 2008. One more achievement for FOSDEM!

Going freelance or not? That's (one of) the question(s)

My sabbatical year ends in one months. During 10 months it was clear to me I would return as an employee, possibly in a 4/5 regime. But february has brought some opportunities to go full-time freelance. So I'm evaluating the possibility this right now.... It would have been easier if I had had these opportunities one year ago, but here I am now with all those questions....

It's not an easy decision. First impact is the salary of course. Will I be able to get enough contracts to survive? I'm not looking to get rich, but a minimum income is necessary of course. And some costs will be added compared to my employee's situation: car, internet connection, mobile phone, servers housing. All these items will cost more if I go freelance. There'll be some "marketing costs' too, in addition to insurance, social security, etc, etc...
On the other hand, if I get lots of contracts, will I still have the time to live a life?

Also, how will I profile myself to prospects? Web developer, database administrator, system administrator, project manager? I love learning new things, so would there be a market if I want to take projects in a domain I don't know yet, but where the customer doesn't find anyone?

And what are the type of contracts that a freelance can get? I'd prefer to not go work 6 months full-time for a customer in a team of 20 developers, but would rather work on smaller-scale and short projects. Isn't that a wrong approach for a freelancer? Won't that approach put me in a market of small companies that don't have the money to invest in such projects?

Another question: during my sabbatical period, I worked quite a lot, but from home, and organising my time as I wanted. Will it still be possible as a freelance? And will I still have the opportunity to put enough time in those projects?

As you see, it's a lot of questions. If you're working as freelance, and you want to share your experience, I'll be happy to read your comments or your mails (sent to rb at raphinou dot com).

Let's regionalise!

So that we can decide to work together because alone we have no impact.

Recent events show politicians haven't learned the lesson.

But maybe I am dumb, and they're only looking to increase the number of ministers and positions to fill....

FOSDEM: feels like home

Today was FOSDEM. I thought I would be able to see people do the FOSDEM Dance, but that didnt work out at Philip and the team called me on stage to join them in this tradition. Here's a picture:


They made me really laugh with this slide (zoom in to read the text):




I'm happy the organisation team keeps this humor in the event. A big thanks to them! It makes leaving much easier.

As other years, lots of people showed up as can be seen on this photos of the Janson, were the keynotes were held:


I volunteered to take photos this year, and I've taken 187 photos today. I'll upload them in the coming days.


I had a great day. I saw people I appreciate a lot but nearly don't see anymore, and I also saw people I meet only at FOSDEM such as Bruce Momjian, Peter Saint-Andre, Matthias Brossard, people from the Namur Lug, the Jabber booth staff, Mozilla people, etc, etc

It really felt like home.

FOSDEM: my first _visit_

Of course I'll be at FOSDEM. I wasn't sure how I would handle my retirement from the organisation team, but it has been really easy. I've closed the chapter, turned the page, started other projects and I've never felt the need to even ask how things were going. The fact that I am 100% trusting the organisation team has certainly helped.

What I wonder though is this: what's it like to be a FOSDEM visitor? Will I enjoy it as much as organising it? Verdict will be known really soon now!
So, see you there on saturday and sunday. I'll be happy to discuss about Profoss, MyOwnDB or any other subject and possibly lend a helping hand to the organisation team :-)

imagemagick for image combination

As we're looking to choose the logo for Dedomenon, I needed to print a bunch of propositions, and wanted to put several images on one page. The best solution I found was to combine multiple images in one with imagemagick:


montage -label %f -frame 5 -tile 1x2 -background '#ffffff' -geometry +4+4 *jpg montage.jpg


It's not an optimal composition as if you have a tiny image it will still take half a page, but it was good enough for me. I'm not sure how I could have combined the images and let the software optimise the placement......

MyOwnDB selected for Plugg's Startup Rally

That's great news! We're in the list of 20 startups selected to give a small (2 minutes) presentation about our product. Based on this presentation, 3 finalists will be selected to give a 10 minutes presentation.

Bart will give this talk as I'll be talking at Linuxworld for Profoss that same day, thanks Murphy.....
If you go to Plugg, keep me updated of how good Bart was :-D
If you're still not registered, use the Profoss coupon to get a 25% rebate!

YUI 2.5: impressive feature list

Wow. Just when I'm concerned of Microsoft's bid to acquire Yahoo and the influence it could have if it goes through, the team releases version 2.5 with some really impressive and time saving features. Some features I (should) like, but I still need to test:

• the uploader (even though it's flash, it can be handy)
  
• the charts (same remarks)
• the image cropper
• the cookie utility
  

Check the complete library to get a better idea of the impressive coverage. Combined with Extjs you get a really good toolkit for your Ajax applications.

Now, let's hope the future of YUI get cleared rapidly....

Server alias superfluous II

Joining the list:
http://Entreprendre2007.be, http://Entreprendre2008.be, http://Ondernemen2007.be, http://Ondernemen2008.be

Leaving the list:
http://Expansys.be

Plugg.eu coupon

Robin, organising the Plugg.eu conference, was so kind to give Profoss a coupon giving 25% rebate on the registration costs. Early bird discount expires at the end of the week, so hurry and register with the code PROFOSS.

Although I won't be able to be there as I'll be at Linuxworld for Profoss, we have registered MyOwnDB in the start-up rally. If we get through the selection process, Bart will get the opportunity to present MyOwnDB!

When you're clueless....

you shut up, or you risk saying stupid things.

Microsoft-Yahoo: what's the real motivation

Since it was announced, I found Microsoft's offer to acquire Yahoo rather strange. What a headache the integration of 2 different cultures would be.

Multiple explanations were proposed. One even said Ms wants to become more open source oriented thanks to the Yahoo acquisition, which I really don't believe.

The best I read until now is Robert X. Cringely's friday column. What if Microsoft didn't really want to acquire Yahoo? What if it was only maneuvering?

And this maneuver has some impact already: I wanted to standardize on YUI for the next release of Myowndb, but I'm not sure it's a good long-term solution. YUI is really a gem, but I don't think Ms would keep that level of quality. It would be more tempted to make it dependent on Silverlight..... And I don't think I'm the only one with these doubts.

Also, what of Zimbra? I did never test it until now, and certainly won't because I imagine this would be one of the first casualties if Microsoft acquires Yahoo.

Also, Would a Microsoft-owned Yahoo still support Hadoop development? Would Yahoo still support multiple a-graded browsers?

Open sourcing videos?

I've read some post form people who can't stand open source as a verb.
I wonder what will be their reaction when they read that "R.E.M. open-sources its music videos". Does it really make sense to say a video is open source? Where's the source?

It's ironic that when open source was started to not emphasize the ambiguous word free (as in free beer/as in free speech?), in this case it would have been better to say "REM freed their videos" as it is also a free license.

Anyway, I think I just joined the club against making open source a verb, as I pretty much prefer to say that "R.E.M. published music videos under an open source license".

Picidae

excellent et astucieux. Testez-le!

Fosdem really reaches its goal

So apparently Opengov was started at FOSDEM 2007? Cool!
I'm really happy of what Fosdem has achieved, it has really gone beyond all expectations. It has enabled developers from different projects, big and small, and from different horizons, to meet and push things forward.

Keep the good work, FOSDEM team! I'll be happy to visit FOSDEM and follow conferences for the first time this year :-D

Unusable Flickr?

Is it me, or is Flickr's UI crap? More specifically: how do you navigate to view all pics in a set?
I have a Profoss set, and when looking at the first picture, how do I go to the next one? Is this mini-thumbnail lying silently and unnoticed on the right, the only way? What's more, this couple of thumbnail is certainly not intuitive: the left pic is the previous pic to the one displayed on the page, and the right one is the next. So when you go to the next picture, the one that was displayed goes the the left thumbnail, and the right thumbnail shows the new next picture. Let me tell you this is not intuitive, and certainly not for novice users.

And what if I want go to the 10th pic?

Compare this with Smugmug's interface, and you'll understand why people a ready to pay twice the price than what Flickr asks for their Pro account.
With Smumug, you get:

• rapid access to pictures, not only to the next one
• a layout adapting at the size of your browser window, showing the current picture in a bigger size when you maximise your window

Here's the gallery for my temporary account, which could become obsolete if I don't subscribe within 14 days.

Maybe Flickr's tools are more advanced (picture editor, stats), but what's the point if their interface is unusable for people not used to Flickr?

I'll keep using the free Flickr for now, but if I have to pay, I think I'll switch to Smugmug.

Any better suggestion?

VMWare at Profoss!

Yes, it's true. It's been a long journey, but it finally happened!
Jeremy van Doorn will represent VMWare in our panel discussion "Open or closed source, an open discussion". He joins Dirk Tombeur from Microsoft and Werner van Unen from Citrix. I expect Frank Kohler from Novell will participate too, but I'm waiting for a confirmation.

This will be a really interesting discussion, so don't hesitate and register! ;-)

Server alias: superfluous?

That's what I wonder when I try to visit

• rtbf.be
• la1.be
• la2.be
• easynet.be
  

I find the last one especially sad.....

N810 of EeePC?

De Nokia N810 is eindelijk beschikbaar in Belgie (online bij Nokia) voor €459.00. Maar de EeePC is ook te verkrijgen, het grootste model voor 380€ bij Expansys.

Ik zal nog wat reviews moeten lezen en goed nadenken om te beslissen wat het best past voor mijn gebruik. Asus heeft toch iets moois met de EeePC, en de N810 blijkt alles te hebben wat ik bij een PDA nodig heb.....

The importance of Microsoft's presence at Profoss

So today I got the confirmation that Dirk Tombeur from Microsoft will participate in the panel discussion at Profoss titled 'Open or closed source: an open discussion'. I think this is very good news for the event attendance: they'll get several point of views exposed, and will be able to choose the best solution for their environment.

Profoss has never been started to convince people that they should switch to free or open source solution. Profoss was started so that people can make their own informed decision. And when I say informed, I mean "by knowing there are possibly free and open source solutions that are worthy alternatives to closed source proposals that should be taken into account".

With its stated goal to be objective, Profoss will let people compare proprietary software to free/open source software. It's no fun when everybody agree in a room. There's no progress happening. Progress happens with a constructive discussion when people disagree and present alternatives. At least strong and weak points of each party appears in such a debate.

I'm really looking forward to this friendly panel discussion!

Batch operation adding a filename suffix

I wanted to resize all sponsors logos on the Profoss website. The file name is of the form logo_company.ext, and I wanted to rename it to something like logo_company.small.ext so I didn't loose the original files. The trick here is to split the filename in 2 parts: filename without the extension and the extension.

This is easy with bash, as it lets you manipulate strings quite easily. For example, if the complete filename is stored in the variable $f, the filename without extension is obtained with ${f%.*}, and the extension with ${f:(-3)} (my file extensions where always 3 characters long ;-)

With the convert tool from imagemagick, it was a piece of cake to get the expected result. Here's the command:
for f in logo_*; do file=${f%.*}; ext=${f:(-3)}; convert -scale 50% $f $file.small.$ext; done

Final schedule of Profoss

The final schedule for the Profoss event on virtualisation on 22 and 23 january in Brussels was published.

I'm really excited as I honnestly think the content that will be brought by the speakers will be very interesting. Each speaker is a specialist in the field, but coming from different horizons.

Here are the speakers, in no particular order:

• Frank Kohler, Virtualization project manager for SuSE
• John Abbott, chief analyst at The 451 Group
• Mike Kreiten, Product manager workstation and server at AMD
• Duncan Hardie, Solaris Product Manager at Sun
• Kris Buytaert, co-author of Virtualisation with Xen
• Tarry Singh, virtualisation advisor and Avastu CEO
• Kristof De Spiegeleer, Q-Layer CEO
• Herbert Poetzl, project leader of Linux-VServer
• Niko Nelissen, VP Business development Q-Layer
• Mattias Rechenburg, OpenQRM project manager
• Werner Fischer, Developer and OpenVZ specialist

For the price of 145€ excl VAT for one day, or 290€ excl VAT for two days, with the restaurant and coffee breaks included, who would want to miss this out? So if you're interestd in virtualisation, don't wait any longer and register!