Uwe's Blog

Persisting an NFS4 mount point on macOS Monterey

2022-03-06T21:23:00+01:00

How to auto-mount an NFS4 share after startup on macOS Monterey?

This blog post explains how to do it with basically one line in /etc/fstab:

192.168.200.200:/nfs_share /System/Volumes/Data/my_nfs_share nfs rw,resvport,hard,bg,intr,rw,tcp,nfc,rsize=65536,wsize=65536

Note that the nolockd option is not included here unlike in the original blog post because this option resulted in an illegal argument error when I tried to access the mount on macOS Monterey. It might work on earlier macOS versions. And the option is probably unnecessary for an NFS4 mount.

Setting up an NFS Server with Linux, Mac, and Kodi Clients

2021-06-22T19:25:00+02:00

Introduction

We'd like to set up an NFS 3 server on a Raspberry Pi with Linux, Mac, and Kodi clients. Kodi does not support NFS 4 out of the box, so we'll stick with NFS 3 for now.

Set up the Server

As a starting point, follow this tutorial to install and set up an NFS (4) server on the Raspberry Pi. Then, make sure to change the server configuration as follows for running an NFS 3 server (assuming you want to export the /nfsexport/myshare directory:

`/etc/exports`

/nfsexport             *(rw,sync,no_subtree_check)
/nfsexport/myshare     *(rw,sync,no_subtree_check)

`/etc/hosts.deny`

rpcbind mountd nfsd statd lockd rquotad : ALL

`/etc/hosts.allow`

rpcbind mountd nfsd statd lockd rquotad : 127.0.0.1 192.168.100.2 192.168.100.3 <more-ip-addresses>

Finally, run sudo systemctl restart nfs-kernel-server.service for the changes to take effect and check if your NFS server is running as expected using sudo rpcinfo -p. Your NFS share will them be available at nfs://<ip-address>/nfsexport/myshare

Linux Client Configuration

Add the following line to your /etc/fstab:

<ip-address>:/nfsexport/myshare /mnt/nfs nfs auto 0 0

Mac (Catalina+) Client Configuration

To mount the NFS volume at boot time, we have to add it to the automounter master map (source).

`/etc/auto_master`

Add the following auto_nfs line to the end of the file. This will try to auto-mount your nfs drive.

#
#
...
/-                 auto_nfs     -nobrowse,nosuid

`/etc/auto_nfs`

Create the file /etc/auto_nfs and use the template below as an example for your settings.

/System/Volumes/Data/Users/Shared/nfs_volume -fstype=nfs,noowners,nolockd,resvport,hard,bg,intr,rw,tcp,nfc,rsize=8192,wsize=8192 nfs://<ip-address>:/nfsexport/myshare

/System/Volumes/Data/Users/Shared/nfs_volume, the location where you want to mount the volume. Change nfs_volume to match the location where you want the volume to be visible in your local file system.
-fstype=nfs, the filetype to mount
noowners,nolockd,resvport,hard,bg,intr,rw,tcp,nfc,rsize=8192,wsize=8192, the mounting parameters where the parameter resvport might be required to connect to the NFS volume.
nfs://<ip-address>:/nfsexport/myshare, the location of the NFS volume

Set the proper rights on the file /etc/auto_nfs:

sudo chmod 644 /etc/auto_nfs

Mount the NFS Volume

Use the following command to initiate the automounter:

$ sudo automount -cv
automount: /net updated
automount: /home updated
automount: /System/Volumes/Data/Users/tisgoud/nfs_volume updated
automount: no unmounts

Your NFS volume is now mounted.

Kodi Client Configuration

Use the following address for your media source definitions:

nfs://<ip-address>/nfsexport/myshare/<sub-path>

Fixing the Bootloader of a Ubuntu RAID System

2019-08-19T17:18:00+02:00

If the remote Ubuntu machine does not boot showing the following error

error: unknown filesystem
 Entering rescue mode...
 grub rescue>

the GRUB bootloader is broken and can be fixed in one of the following ways.

Using the GRUB rescue concole

Follow the steps described in this blog post.

Using a Live System

Boot into a rescue system (can be done via your server provider's web interface).
Connect via ssh using ssh root@ip.address.
Once connected, make sure the mdadm tools are installed and assembled:

apt-get install mdadm mdadm --assemble --scan
Mount the system partition (e.g. /dev/md2):

sudo mount /dev/md2 /mnt
Check the mdadm.conf entries by looking at

mdadm --examine --scan
Mount the boot partition (e.g /dev/md1):

mount /dev/md1 /mnt/boot
Set up the other required mounts:

mount -o bind /dev /mnt/dev mount -o bind /sys /mnt/sys mount -t proc /proc /mnt/proc cp /proc/mounts /mnt/etc/mtab
Switch into the chroot environment:

chroot /mnt /bin/bash
Install GRUB 2 into the MBR of all relevant disks (e.g. /dev/sda and /dev/sdb2 in case of a RAID system with one mirror disk):

grub-install /dev/sda grub-install /dev/sdb
If any error occur during the last operation, try:

grub-install --recheck /dev/sdX
Recreate the file /boot/grub/grub.cfg using:

update-grub
Leave the chroot environment using CTRL+D.
Reboot.

Mounting a Samba Share on the Raspberry Pi

2018-12-07T20:24:00+01:00

Install the following packages:

apt-get install  samba-common smbclient samba-common-bin smbclient  cifs-utils

Create a local directory:

mkdir /mnt/abc

Mount the Samba share using your username:

mount -t cifs //server/share /mnt/abc -o user=uwe

If the Samba share is provided by a FritzBox via its NAS feature and assuming the share is named "MyNAS" on your FritzBox, then the above command would be:

mount -t cifs //<FritzBox IP>/MyNAS /mnt/abc -o user=uwe

Based on this SO answer.

Update: if you want the share to be mounted on every reboot, follow this guide (in German).

Upgrade Your Apache to a Newer PHP Version

2018-06-08T16:30:00+02:00

Upgrading PHP from 7.0 to 7.1 or 7.2 on Ubuntu 16.04 LTS is almost trivial as this blog post (in German) explains:

First, add a PPA (personal package archive):

sudo apt-get install -y python-software-properties (falls nicht bereits installiert)
sudo add-apt-repository -y ppa:ondrej/php
sudo apt-get update -y

Second, upgrade all packages:

sudo apt-get -y update && apt-get -y upgrade && apt-get -y dist-upgrade

Third, if you use Nextcloud, install all required PHP packages (for PHP 7.1 and 7.2 at the same time in this example):

sudo apt-get install -y php7.1 libapache2-mod-php7.1 php7.1-cli php7.1-common php7.1-mbstring php7.1-gd php7.1-intl php7.1-xml php7.1-mysql php7.1-mcrypt php7.1-zip php7.1-dev php7.1-curl libapache2-mod-php7.2 php7.2-cli php7.2-common php7.2-mbstring php7.2-gd php7.2-intl php7.2-xml php7.2-mysql php7.2-zip php7.2-dev php7.2-curl

Then, add the following code to your /etc/php/7.1/apache2/php.ini or /etc/php/7.1/apache2/php.ini, respectively:

opcache.enable=1
opcache.enable_cli=1
opcache.interned_strings_buffer=8
opcache.max_accelerated_files=10000
opcache.memory_consumption=128
opcache.save_comments=1
opcache.revalidate_freq=1

Finally, enable the new PHP modules in your Apache:

sudo a2dismod php7.0
sudo a2enmod php7.1
service apache2 restart

Setting up a new Ubuntu Webserver

2017-11-11T20:15:00+01:00

Here's a collection of links that provide helpful information when you have a brandnew Ubuntu machine with root access and you want to set it up as a webserver:

If you want to use PHP 7 with mcrypt, you might need to add an extra apt repository like this:

sudo add-apt-repository universe
sudo add-apt-repository -y ppa:ondrej/php
sudo apt-get install -y php7.0 libapache2-mod-php7.0 php7.0-mcrypt php7.0-mysql

The following links help you secure your server:

Setting up a systemd Service under Ubuntu

2017-11-11T19:15:00+01:00

Starting with 15.04, Ubuntu uses systemd instead of Upstart as the default system and service manager. This page offers a good comparison between the two systems.

If you want to add a new systemd service (e.g. for running a Java application everytime your machine boots), create the following file under /etc/systemd/system/myservice.service:

[Unit]
Description=My Service Description
After=syslog.target

[Service]
WorkingDirectory=/path/to/workdir
SyslogIdentifier=MyService
EnvironmentFile=/path/to/workdir/myservice.environment
ExecStart=/bin/bash -c "java -jar /path/to/jar/MyService.jar"
User=uwe
# see https://stegard.net/2016/08/gracefully-killing-a-java-process-managed-by-systemd/
SuccessExitStatus=143
Type=simple

[Install]
WantedBy=multi-user.target

After creating or modifying a unit file, you should reload the systemd process itself to pick up your changes:

sudo systemctl daemon-reload

To enable a service to start automatically at boot, type:

sudo systemctl enable unsereernte.service

You can start the service by typing:

sudo systemctl start unsereernte.service

Aside from start, you can stop, restart, or reload the service.

A systemd component called journald collects and manages journal entries from all parts of the system. This is basically log information from applications and the kernel.

To see all log entries, starting at the oldest entry, type:

journalctl

A good tutorial on systemd services can be found here.

Redirecting Requests from Apache to Jetty

2017-11-05T21:55:00+01:00

Assuming you are running an Apache server with a virtual host for abs.mydomain.com, put this into the virtual host's .conf file to redirect all traffic to a Jetty web server running locally behind your Apache:

<Proxy *>
    Order deny,allow
    Allow from all
</Proxy>

ProxyPass / http://localhost:1234/ retry=1
ProxyPassReverse / http://localhost:1234/
ProxyVia On

Here, 1234is the port on which the Jetty server is listening. This port should be blocked from direct access via Apache. Using Ubuntu's default firewall ufw, this can be done by running a simple sudo ufw deny 1234 on the command line.

Added benefit: the internal Jetty server does not need to use SSL if the Apache server provides an SSL connection to the client.

See this SO answer for an explanation of the ProxyPassReverse directive.

40 Falsehoods Programmers Believe about Names

2017-07-27T07:25:00+02:00

Patrick McKenzie has never seen a computer system that handles names properly. Sad!

Programming as if the Domain (and Performance) Mattered

2017-07-26T18:55:00+02:00

A long read by my man Carlo Pescio.

It's not CI, it's just CI Theatre

2017-05-26T18:32:00+02:00

Are you practising CI? Find out by reading this post by Suzie Prince.

The Initializer Pattern in Java

2017-03-18T20:55:00+01:00

A "type-safe builder pattern in Java".

The Lean Inception

2017-03-16T19:43:00+01:00

I should probably read this series of articles by Paulo Caroli before I try to convince my boss why implementing an MVP is a good idea.

PiBakery

2017-03-09T21:33:00+01:00

A great way to configure a Raspberry Pi from scratch is PiBakery. After struggling a while with setting up a VNC server and some other things on a fresh Pi connected only via Ethernet (no display, keyboard, or mouse), finding PiBakery was very helpful.

Here's my configuration:

![PiBakery configuration]({static}images/pibakery_conf.png)

The long string in the second Run Command step is:

DEBIAN_FRONTEND=noninteractive DEBIAN_PRIORITY=critical apt-get -q -y -o "Dpkg::Options::=--force-confdef" -o "Dpkg::Options::=--force-confold" upgrade

Using your SSL certificate for your Spark web application

2017-03-09T21:12:00+01:00

I've fallen in love with Spark recently. Being unexperienced with web development, I stumbled upon Spark when I was looking for a small framework for setting up a simple web application based on Java. My learning curve has been steep and my first website (consisting of nothing but a simple form) is running smoothly.

One issue I had to deal with was using my domain's SSL certificate (proudly sponsored by Let's Encrypt) for the integrated web server that comes with Spark. As you can read here, a keystore and a password need to be specified in order to achieve this.

Alright, nothing easier than coming up with a password, but how do I set up a Java keystore (.JKS) that uses my certificate? A quick Startpage search directed me to Maximilian Böhm's tutorial on JKS and Let's Encrypt certificates.
I could skip step 1 as I had created and installed my certificates already. For step 2, I needed to locate the directory in which my Plesk installation stored the fullchain.pem and privkey.pem files. I found them in:

/opt/psa/var/modules/letsencrypt/etc/live/mydomain.com/

I copied them to a temp directory to make things easier, and then ran the following command:

openssl pkcs12 -export -in fullchain.pem -inkey privkey.pem -out pkcs.p12 -name mydomain_letsencrypt

After typing in and remembering the password, the resulting pkcs.p12 file could be used to create the JKS file as described in step 3:

keytool -importkeystore -deststorepass DEST_STORE_PASS -destkeypass KEY_PASS -destkeystore keystore.jks -srckeystore pkcs.p12 -srcstoretype PKCS12 -srcstorepass SRC_STORE_PASS -alias mydomain_letsencrypt

Here, SRC_STORE_PASS is the password chosen in step 2. Make sure to remember and distinguish those three passwords! Or use the same phrase for all of them.

The result was saved as keystore.pks. I changed my Spark code accordingly:

String keyStoreLocation = "keystore.jks";
String keyStorePassword = "DEST_STORE_PASS";
secure(keyStoreLocation, keyStorePassword, null, null);

and put the file next my JAR. Done.

Update 2017-09-02: Here's a script that updates the keystore entry (e.g in case the existing SSL certificate has been invalided and a new one has been created).

#!/bin/bash

CERT_PATH=/opt/psa/var/modules/letsencrypt/etc/live/mydomain/
CERT_ALIAS=mydomain_letsencrypt
PKCS_FILENAME=pkcs.p12
PKCS_PW='my_pkcs_pw'
KEYSTORE_FILENAME=keystore.jks
KEYSTORE_PW='my_keystore_pw'

# generate PKCS12 file
openssl pkcs12 -export -in $CERT_PATH/fullchain.pem -inkey $CERT_PATH/privkey.pem -out $PKCS_FILENAME -name $CERT_ALIAS -passout pass:$PKCS_PW

# delete existing entry in Java keystore
keytool -delete -keystore $KEYSTORE_FILENAME -alias $CERT_ALIAS -storepass $KEYSTORE_PW

# add new Java keystore entry from PKCS12 file
keytool -importkeystore -deststorepass $KEYSTORE_PW -destkeypass $KEYSTORE_PW -destkeystore $KEYSTORE_FILENAME -srckeystore $PKCS_FILENAME -srcstoretype PKCS12 -srcstorepass $PKCS_PW -alias $CERT_ALIAS

rm $PKCS_FILENAME

Referencing a Bean in JasperReports

2016-09-10T18:53:00+02:00

Thanks to this answer on StackOverflow, I know how easy it is to reference the bean itself instead of one of its properties in a text field expression in JasperReports. The trick is to declare a field whose description is set to the keyword _THIS:

<field name="myBean" class="de.uwesander.MyBean">
    <fieldDescription>_THIS</fieldDescription>
</field>

Then it can be referenced in the text field expression as follows:

<textFieldExpression>
    <![CDATA[$F{myBean} instanceof ...]]>
</textFieldExpression>

Installing a Tor Relay from Source on a Raspberry Pi

2016-09-03T21:33:00+02:00

The other day I learned on Twitter that I should update the Tor relay software on my Raspberry Pi as the Tor project had introduced a new Bridge Authority. So I lazily ran sudo apt-get update followed by sudo apt-get upgrade only to notice that no new version of the tor package was available via the package manager. Bummer.

Waiting for the package manager to provide a new version seemed futile so I decided to build Tor from scratch. I found a very helpful answer on StackOverflow that listed all the steps required for my setup.

Make sure to have all dependencies needed for compiling the sources:

sudo apt-get install git build-essential automake libevent-dev libssl-dev
Clone Tor from git.torproject.org:

git clone https://git.torproject.org/tor.git
Go into the tor directory:

cd tor
Switch to the latest release branch:

git checkout release-x.y.z
Run ./autogen.sh
Run ./configure or ./configure --disable-asciidoc if you don't want to build the manpages.
Run make
(Optiona): run make install

This last step installed the new tor binary in /usr/local/bin/ whereas the existing tor binary was still located in /usr/bin/ from where it was picked up by /etc/init.d/tor (I want to run my tor relay as a service whenever the Raspberry reboots). Luckily, this problem had been solved by someone else who shared his findings on a mailing list.

If you're using `service tor {start, stop, reload, etc.}:

$ /usr/bin/tor --version; /usr/local/bin/tor --version $ grep DAEMON=/etc/init.d/tor
The latter will likely point to /usr/bin/tor, which might be outdated.
If that's the case, change that line in /etc/init.d/tor to point to the new Tor executable /usr/local/bin/tor - that's where it should be; if you don't like that, change BINDIR = /usr/local/bin in Tor's Makefile and make install again.
The Debian Tor packages seem to like to assume torrc will be placed in /etc/tor/torrc . If that's where your torrc resides, make a symlink to it from /usr/local/etc/tor , which is where the new Tor executable will look for it:

ln -s /etc/tor/torrc /usr/local/etc/tor/torrc
Stop the running tor service:

sudo /etc/init.d/tor stop
Reload the tor daemon:

sudo systemctl daemon-reload
Restart the tor service:

sudo /etc/init.d/tor start

Regex for German Zip Codes

2016-07-05T19:56:00+02:00

I found the following regular expression (source) working fine for me:

^(?!01000|99999)(0[1-9]\\d{3}|[1-9]\\d{4})$

I tried two alternatives that did not work because they were too general:

[0-9]{5}

^([0]{1}[1-9]{1}|[1-9]{1}[0-9]{1})[0-9]{3}$

The first regex matches the zip code "00000" which is invalid in the German system. The second regex matches the zip code "99999" which is invalid, too.

Renewal of Let's Encrypt Certificates Using Plesk

2016-06-14T18:18:00+02:00

Note to self: If the Let's Encrypt extension for Plesk fails to renew a certificate (when triggered manually), disable the automatic forwarding to an SSL connection in the Apache settings. This forced secure connection seems to disturb the renewal script.

The error I got was something like this:

Let's Encrypt SSL certificate installation failed: Failed letsencrypt execution:
2016-01-19 13:56:07,885:WARNING:letsencrypt.cli:Root (sudo) is required to run most of letsencrypt functionality.
Failed authorization procedure.
removed.domain.com (http-01): urn:acme:error:unauthorized :: The client lacks sufficient authorization ::
Invalid response from http://removed.domain.com/.well-known/acme-challenge/
REMOVED-ID [REMOVED IP]: 401
IMPORTANT NOTES: - The following errors were reported by the server:
Domain: removed.domain.com Type: urn:acme:error:unauthorized
Detail: Invalid response from http://removed.domain.com /.well-known/acme-challenge/
REMOVED-ID [REMOVED IP]: 401

Create Custom Application Icons for MacOS X

2016-05-01T20:58:00+02:00

This blog post by Chris Stevens shows you how.

Testing Delegate Methods

2016-04-15T18:33:00+02:00

We had an interesting discussion the other day: should we write unit tests for delegate methods, that is, methods that only delegate the call to a delegate object? Here's an example of such a method:

public void do(Object o1, Object o2) {
    delegate.do(o1,o2);
}

I said I did not see much value in a unit test for this method as all the test could do is verify whether the do() method of the mocked delegate object was called. Instead of testing behaviour, we would test the wiring which is close to trivial in this case and should be tested implicitly as part of an integration test involving the delegate object and the main object.

Some of my colleagues disagreed with me. Their argument was that a unit test would detect changes in the delegate call and indicate the intended usage of the delegate object. They were concerned an incorrect wiring could be introduced without a unit test. And this unit test would be cheap to write.

I did not find any of these arguments very convincing - and I'm in good company: the JUnit team's FAQ webpage has a section on Best Practices which comments on testing delegate methods. You can read it here. Thank you, J. B. Rainsberger!

Headless Testing with TestFX

2016-04-11T18:35:00+02:00

Jérome Cambon wrote a nice introduction to TestFX and how to enable the headless mode using Monocle.

I tried to reproduce the steps he described for running my TestFX GUI tests in headless mode from Eclipse so I can keep using my mouse while they are running. The crux is setting the VM arguments in Eclipse. First, I tried to set

-Dtestfx.robot=glass -Dglass.platform=Monocle -Dmonocle.platform=Headless -Dprism.order=sw

in the VM arguments section of the JRE tab in my Maven run configuration:

![Maven run configuration in Eclipse]({static}/images/run_configuration.png)

This did not work. I could still see the TextFX robot move the mouse cursor on my screen.

I had to specify the VM arguments using Maven's 'argLine' parameter:

clean install -DargLine="-Dtestfx.robot=glass -Dglass.platform=Monocle -Dmonocle.platform=Headless -Dprism.order=sw"

![Maven argLine parameter]({static}/images/arg_line.png)

If anyone knows why setting the VM arguments in the run configuration did not work, please let me know.

A Potential Pitfall with the TextFormatter Class in JavaFX

2016-04-07T18:52:00+02:00

A common use case for the TextFormatter class in JavaFX is a text field that should convert any lower-case character input into an upper-case character. This can be achieved easily by defining a filter for the text formatter as described here. A straightforward implementation could look like this:

private UnaryOperator<Change> getFilter() {
    return change -> {
        String text = change.getText();
        change.setText(text.toUppercase());
        return change;
    };
}

This implementation comes with one pitfall, though. When the (German) user types the character "ß" into the text field, it will be converted into "SS" because that's how Germans roll. So the resulting string is one character longer than the input string. Why is this a problem? Well, the cursor position will be between the two "S" characters after the conversion:

![cursor in text field]({static}/images/cursor_ss.png)

When the user types in the next character, it will be inserted between the two "S" characters, not behind them. Bummer.

Luckily, the masterminds behind JavaFX give us the tools to correct this unfortunate behaviour since the change object that the filter works on allows for setting the anchor and caret position in the text field. We need to correct them by the difference between the string lengths before and after the conversion:

private UnaryOperator<Change> getFilter() {
    return change -> {
        String originalText = change.getText();
        String upperCaseText = originalText.toUpperCase();
        change.setText(upperCaseText);

        correctAnchorAndCaretPositions(change, originalText);

        return change;
    };
}

private void correctAnchorAndCaretPositions(Change change, String originalText) {
    int diff = change.getText().length() - originalText.length();
    if (diff != 0) {
        change.setAnchor(change.getAnchor() + diff);
        change.setCaretPosition(change.getCaretPosition() + diff);
    }
}

This puts the cursor at the end of the string no matter how many "ß" characters the user types in or pastes into the text field.

The TextFormatter Class in JavaFX: How to Restrict User Input in a Text Field

2016-04-05T20:01:00+02:00

There are a lot of code examples for restricting or modifying user input into a JavaFX text field. Most examples I have seen suggest adding a change listener to the text field's text property. Here's how you would allow only lower-case characters in your text field using the change listener approach:

package de.uwesander.javafx.textformatter;

import javafx.application.Application;
import javafx.scene.Scene;
import javafx.scene.control.TextField;
import javafx.scene.layout.StackPane;
import javafx.stage.Stage;

public class TextFieldListenerExample extends Application {

    @Override
    public void start(Stage stage) {
        TextField textField = new TextField();
        textField.setMaxWidth(300);

        textField.textProperty().addListener((obs, oldText, newText) -> {
            if (newText.matches("[a-z]")) {
                textField.setText(newText);
            } else {
                textField.setText(oldText);
            }
        });

        StackPane pane = new StackPane(textField);
        Scene scene = new Scene(pane, 400, 200);

        stage.setTitle("TextField Listener Example");
        stage.setScene(scene);
        stage.show();
    }

    public static void main(String[] args) {
        launch(args);
    }
}

This approach comes with one drawback: you'll have two events being triggered by two changes of the text property. The first change is caused by the direct user input, the second change is caused by the manipulation of the user input by the change listener. The user won't notice these two events, but somewhere in your code you have another listener for the text property, that listener will receive two events, one with the "invalid" change and a second one with the "valid" change.

Another approach was suggested by Richard Bair a long time ago. His suggestion resulted in the TextFormatter class being added to JavaFX with version 8u40. It's a clean way to format, filter, or restrict user input. Here's how it works:

package de.uwesander.javafx.textformatter;

import java.util.function.UnaryOperator;

import javafx.application.Application;
import javafx.scene.Scene;
import javafx.scene.control.TextField;
import javafx.scene.control.TextFormatter;
import javafx.scene.control.TextFormatter.Change;
import javafx.scene.layout.StackPane;
import javafx.stage.Stage;

public class TextFormatterExample extends Application {

    @Override
    public void start(Stage stage) {
        TextField textField = new TextField();
        textField.setMaxWidth(300);

        TextFormatter<String> textFormatter = getTextFormatter();
        textField.setTextFormatter(textFormatter);

        StackPane pane = new StackPane(textField);
        Scene scene = new Scene(pane, 400, 200);

        stage.setTitle("TextFormatter Example");
        stage.setScene(scene);
        stage.show();
    }

    private TextFormatter<String> getTextFormatter() {
        UnaryOperator<Change> filter = getFilter();
        TextFormatter<String> textFormatter = new TextFormatter<>(filter);
        return textFormatter;
    }

    private UnaryOperator<Change> getFilter() {
        return change -> {
            String text = change.getText();

            if (!change.isContentChange()) {
                return change;
            }

            if (text.matches("[a-z]*") || text.isEmpty()) {
                return change;
            }

            return null;
         };
    }

    public static void main(String[] args) {
        launch(args);
    }
}

What's still missing is support for the backspace key to delete a character in the text field, but that's just my silly example implementation.

[Update 2016-12-28: The implementation was improved to support the deletion and selection of characters.]

This approach intercepts the user input before it's written into the text property and thus fires only one event. Before that happens, you can examine and modify the Change object in the UnaryOperator defined in getFilter() method.

In addition to filtering, a TextFormatter object can convert a value to a string representation and vice versa. From the Javadoc:

A value converter and value can be used to provide a special format that represents a value of type V. If the control is editable and the text is changed by the user, the value is then updated to correspond to the text.

Sounds pretty handy.

Owncloud with Apache 2.4, PHP FPM, and Plesk

2016-04-04T19:46:00+02:00

After about two weeks of fiddling around, I finally managed to get my own instance of owncloud up and running with Apache behind Plesk. Here's how I did it:

Set up a sub-domain in Plesk, e.g. owcloud.mydomain.com
Set up a database for owncloud using Plesk.
Make sure to use HTTP strict transport security by adding the following Apache HTTPS directive:

<IfModule mod_headers.c> Header always set Strict-Transport-Security "max-age=15768000; includeSubDomains; preload" </IfModule>
Create a directory for your owncloud data, e.g. /var/oc_data
Give sufficient permissions to your HTTP user on the data directory:

chown -R $user1:psacln /var/oc_data/ chmod 750 /var/oc_data

In my case, user1 is the the name of the Plesk user with which I had created the sub-domain. psacln is the group assigned to this user by Plesk.
Update PHP to version 5.6.x. There is this bug in version 5.5.x which makes owncloud unusable if you don't want to use mod_php (which is discouraged).
Configure PHP in Plesk and Apache as follows:
- Run PHP as an FPM application
- Set open_basedir to {WEBSPACEROOT}{/}{:}{TMP}{/}{:}/dev/urandom{:}/var/oc_data{/}
- Set memory_limit to 256M (or something like that)
- Set max_execution_time to 120 (or something like that)
- Set max_input_time to 120 (or something like that)
- Set post_max_size to 128M (or something like that)
- Set upload_max_filesize to 64M (or something like that)
Check this owncloud documentation page for potential additional PHP settings when running in FPM mode.
Download the setup-owncloud.php web installer as described here and run it from your browser.
Have fun.

How to Make Unit Tests more Likeable

2016-04-04T19:13:00+02:00

What are these unit tests good for? Are we testing Mockito?

This is what a colleague asked another colleague a couple of days ago. He was referring to some rather complicated and long unit tests that used a lot of mocks and verify statements. I knew what he was talking about since I had seen quite a few tests like that myself.

Whenever a JUnit test method has more than ten lines or so, I get suspicious (and sometimes start crying). Add a few ArgumentCaptors and I'm done with that test. If I'm supposed to fix a test like that, I tend to give up quickly and rather delete the test method than try to understand what is being tested and why it fails.

A good test, however, makes me smile. It's short, concise and readable because it follows the arrange, act, assert principle. It uses Mockito's fluent API if it uses Mockito at all. That's anyString() instead of Mockito.anyString() and times(2) instead of VerificationModeFactory.times(2) (which is Mockito's internal API btw). And mocks, stubs, spies, and dummy objects are set up in the @Before method instead of every test method. This saves a lot of boilerplate code and makes me want to fix the test if it breaks.

Adding JavaFX Properties to a DTO

2016-03-21T20:23:00+01:00

Now that I have understood what a DTO is good for, I'm thinking about ways to improve it. As they are right now, our DTOs are POJOs extended by JSR 303 annotations for defining constraints on their fields/methods. The annotations are duplicates of the annotations defined on the corresponding entities.

Since I have worked with the UI code most of my time so far, I have felt the drawback of not having JavaFX properties in our DTOs: they cannot be observed easily by the GUI controls. There is a wrapping mechanism to add this observability, but it is rather complicated to use and requires a lot of extra code.

So why not make the DTOs full JavaFX beans? After all, we would not be the first project to do this. Two common objections were raised when I brought this up:

DTOs are created on the server which should not have dependencies to a GUI/client technology.
We would need to mix JavaFX beans with JSR 303 annotations.

Let's address the first point. Yes, having a GUI dependency in your server code is not a good idea. But are JavaFX properties a GUI technology? Or even a client technology? It's a bit unfortunate that they were introduced as part of JavaFX in a javafx.* package such that everyone identifies them as a GUI technology at first sight. It's an image problem. I would argue they are a self-contained technology which is independent of any GUI parts of JavaFX and could therefore be used for any data model, no matter if it's on the client, on the server, or shared between both. Even if one does not use JavaFX for the frontend, depending on JavaFX properties would do almost no harm in practice since JavaFX is part of the JDK and your server code certainly depends on the JDK. So the no client dependencies on the server argument is merely a formal one in my opinion.

What about the second point, mixing JavaFX properties with bean validation annotations? I have never tried it myself, but I have the feeling it cannot be that big of a problem. If it works in the FXForm2 demo, why should there be a fundamental incompatibility between these two technologies? But I admit that a good prototype is needed to make a final judgment.

We will have a meeting on this tomorrow. Let's see what the others think.

What's up with DTOs?

2016-03-19T21:30:00+01:00

Having worked a lot with EMF and Eclipse RCP applications based on EMF as the framework of choice for transferring domain models into code, I never got in touch with real-life examples of Data Transfer Objects (DTOs). I was familiar with the pattern in theory, but I never saw a DTO in practice since EMF works fine without them, even in a client-server architecture.

This changed recently. My new project is a client-server system which does not use EMF for implementing its domain objects. DTOs are used heavily to transfer data from the server to the client and vice versa. During my first journeys through the unknown code base, I stumbled upon a lot of DTOs and wondered what they were needed for at all since they seemed to introduce a lot of redundancy and required a mapping to and from the entities managed by the server. Wouldn't it be more efficient and practical to simply use the entities as the shared domain objects on the server and the client? What is right in the EMF world cannot be wrong in the JavaEE world, I thought.

So I did some research and found this blog post by Adam Bien who seemed to confirm my doubts. "JavaEE 7 retired the DTO" was all I needed to hear in order to start a discussion with one of my new colleagues about the whole purpose of DTOs in our project. He was nice enough to take a long break from his current task and give me an extended explanation of why and how DTOs are necessary in our project. He answered all of my questions patiently, even if some of them were perhaps a bit naive and showed my lack of experience with some aspects of JavaEE. Needless to say it was a fruitful discussion.

I did some more research the next day based on his explanation. When I was finished, I had changed my mind. I am now convinced DTOs add value to our project despite the code duplication and the extra efforts for the domain mapping because they embody more than a simple transfer mechanism . Here are the pros and cons of DTOs based on my current understanding:

Cons

Duplicate code (duplicate data models, duplicate constraints,...)
A mapping to and from the domain objects is required

Pros

DTO exposes the domain model in a form that is tailored for the client's needs
Entity should be managed by the server all the time, never deattached from it
Can save bandwidth by carrying only a selected subset of the entity's attributes/dependencies
No client dependency on JPA if entity uses JPA annotations (e.g. for constraints)
More pros here on Adam Bien's blog (2014)

**Resources I found helpful **

Any Complaints?

2016-03-16T19:52:00+01:00

https://twitter.com/phillip_webb/status/705909774001377280

I usually go for "too much boilerplate".

Can There Be Too Much of a Good Thing?

2016-03-16T19:49:00+01:00

I recently worked on JavaFX UI code that was new to me. In order to extend the existing UI by a simple label, I needed access to a bean of the underlying data model. Much to my surprise, I had to dig deep to find it hidden under multiple layers of wrappers. This was the data structure:

![A bean wrapped in a BeanWrapper wrapped in an ObjectProperty wrapped in a WrapperPropertyProxy. Quite impressive.]({static}/images/wrapper_galore.png)

Wow. Clearly, someone tried to satisfy their fetish for wrappers and indirection. I'm going show this image to the next computer science student I come across and ask them how many design (anti-)patterns they can see.

Zooming in the Fancy Chart Control

2014-04-24T16:03:00+02:00

A new version of the fancy chart control is available at https://github.com/tesis-dynaware/fancy-chart

This version supports zooming for the XY chart: by dragging the mouse, you can select an area in the chart to scale the selected area up to full size of the chart. This allows you to have a close look at the selected data points.

Headless UI Testing with TestFX and JavaFX 8

2014-04-23T14:41:00+02:00

JavaFX is a great UI toolkit. TestFX is a great library for testing the user interfaces written in JavaFX. Writing graphical tests with TestFX is simple and fast, but one challenge remains when you build your software using a headless build machine: how can you perform your UI tests in headless mode during your build?

Luckily, JavaFX 8 Update 20 will contain Monocle, which should allow us to run these headless tests.

First analysis using the latest beta of JDK 8u20 indicates that these tests are feasible, but require some changes in the TestFX library. We used a simple standalone JavaFX application for which we wrote a basic TestFX test class. We ran this test using the following command line options:

-Dglass.platform=Monocle
-Dmonocle.platform=Headless
-Dprism.order=sw

Unfortunately, this was not enough to make the test work. So we took a deeper look at the inner workings of TestFX and found that a working solution required the following changes:

refactor a part of the TestFX code such that different implementations of the org.loadui.testfx.framework.robot.ScreenRobot interface can be used. This allowed us to make test use an other robot implementation, in this case an adapter for MonocleRobot. The current TestFX implementation relies on java.awt.Robot, which uses the system event queue to move the system cursor. This, of course, does not integrate with Monocle in headless mode.
implement a ScreenRobot adapter for MonocleRobot. MonocleRobot is in internal package and thus not available via public API. In order to obtain an instance of this class, we had to call:

java Platform.runLater(() -> { robot = com.sun.glass.ui.Application.GetApplication().createRobot(); ... });

This setup gave us a green test eventually. :-)

Enhanced Fancy Chart

2014-03-21T12:28:00+01:00

A new version of the fancy chart control is available at https://github.com/tesis-dynaware/fancy-chart

Data can now be imported and exported to CSV, XLS, and HDF5 files. A lot of code was refactored to make it easier to support other file formats in the future.

Source Code for the Fancy Chart Control

2014-03-14T15:38:00+01:00

Due to popular demand, I published the source code for the combined LineChart and TableView control mentioned in this post to Github. You can find the sources here:

https://github.com/tesis-dynaware/fancy-chart

Visit Zombieland

2014-03-13T16:43:00+01:00

We finally released our FBX importer code for JavaFX 8 on Github:

https://github.com/tesis-dynaware/fbx-importer

You can use it to load (binary) FBX files and show the result as a JavaFX 3D mesh. Have fun playing around and please don't hesitate to give us feedback.

A Combined LineChart and TableView Widget in JavaFX

2014-02-19T12:10:00+01:00

[![Video of a combined LineChart and Table widget in JavaFX](http://img.youtube.com/vi/w0No8gBxDi8/0.jpg)](http://www.youtube.com/watch?v=w0No8gBxDi8) (click to watch the video)

This video shows a widget for displaying three characteristics (velocity vs. distance) in a LineChart node and a tabbed TableView node. The two nodes provide different views on the same data.

The widget comes with the following features:

hovering over a data point in the chart will show a popup with the x and y values.
selecting a data point in the chart will select the corresponding data item in the table view and vice versa.
when a data item was edited in the table view, the chart will update immediately.
three colour pickers allow for changing the line colours in the chart.
import and export data from sources like CSV, Excel, HDF5 (to be done).

These features were implemented using JavaFX bindings, events, animations, and CSS styling.

A Widget Picker Control in JavaFX

2014-02-07T12:38:00+01:00

The following video shows a custom control in JavaFX that lets you select items from a list and position them in a grid. The items represent UI widgets, that is, JavaFX controls, with a defined size on the grid. The grid represents a page on which these widgets will be shown.

[![Video of a widget picker in JavaFX](http://img.youtube.com/vi/CvVRlZOfE6c/0.jpg)](http://www.youtube.com/watch?v=CvVRlZOfE6c) (click to watch the video)

A Graph Editor in JavaFX

2014-02-06T16:29:00+01:00

Another thing we've been working on is going to see the light of day soon: a graph editor for JavaFX. The current beta version provides the following main features:

create draggable, resizable nodes and connections between them
add multiple input or output connectors to a node
create joints inside a connection
add custom skins for nodes, connections, connectors, and joints
enhanced user experience through comprehensive styling during drag-and-drop actions

Internally, we use an EMF model to represent the UI objects and their state. The model is in sync with the UI at all times.

Here's a short video of our current development version:

[![Video of a graph editor in JavaFX](http://img.youtube.com/vi/LI\_aPjz1O00/0.jpg)](http://www.youtube.com/watch?v=LI\_aPjz1O00) (click to watch the video)

A Custom Breadcrumb Control in JavaFX

2014-02-06T14:49:00+01:00

After investing some serious brain power in the dynamic sizing of the breadcrumbs in a navigation bar (described in this post), I forgot to provide a closer look at the outcome.

Here's a short video of this custom control written in JavaFX:

[![Video of a graph editor in JavaFX](http://img.youtube.com/vi/59PPwihHiAw/0.jpg)](http://www.youtube.com/watch?v=59PPwihHiAw) (click to watch the video)

Using FXForm2 with EMF Models

2013-11-08T13:46:00+01:00

Based on these considerations from July, we gave FXForm2 a shot for generating forms from EMF models. It worked quite smoothly after we wrote the factories that defined how the custom data types used in our models should be displayed.

We found a few things that are worth noting:

Using Java Bean Validation
In order to validate our EMF models using the Java Bean Validation API (JSR-303) that comes with FXForm2, we wrote wrapper classes for those model elements that we wanted to add constraints to. We decided to not touch the generated code.
Using another validation implementationFXForm2 can use any implementation of JSR-303. Using the Hibernate implementation is suggested on the FXForm2 website, but we found it easy to supply our own implementation based on EMF's validation features such as the Diagnostician class.
Validation philosophies
When you enter an invalid value into a field, FXForm2 won't update the corresponding model attribute. The value you see in the form differs from the value in the model. The model thus always has a valid state.EMF validation typically works in a different way: you start the validation operation only after all model values were set. This allows for invalid states of the model.
Dynamic beans
An open issue seems to be (as least for us) the handling of a String-to-String Map attribute inside an EMF bean class that stores user-supplied key-value pairs which should be displayed as regular attributes of the bean. Antoine suggested a solution based on a custom control (with a factory and a skin), but it comes with a different visual appearance for the user-defined attributes. This is work in progress.

Implementing Breadcrumbs in JavaFX

2013-11-04T15:08:00+01:00

I've been struggling to implement a JavaFX breadcrumbs bar using a ToolBar with a few Button items in it. The breadcrumbs should have the shape of an arrow as in the example shown in Loop81's blog post.

The breadcrumbs bar in the Ensemble application uses the same approach. What I don't like about this implementation is the use of rather sophisticated CSS rules in combination with static images.

.breadcrumbs .item {
  -fx-background-color: null;
  -fx-background-radius: 0;
  -fx-background-insets: 0; 
  -fx-border-color: null;
  -fx-border-insets: 0;
  -fx-border-image-source: url("crumb.png");
  -fx-border-image-slice: 1 10 1 10 fill;
  -fx-border-image-width: 1 10 1 10;
  -fx-border-image-repeat: stretch;
  -fx-font-size: 18px;
}

.breadcrumbs .item:hover {
  -fx-border-image-source: url("crumb-hover.png");
}

.breadcrumbs .alone {
  -fx-border-image-width: 0 0 0 0;
  -fx-border-image-repeat: stretch;
  -fx-border-image-source: url("crumb-selected.png");
  -fx-border-color: transparent #6e737d transparent rgba(255,255,255,0.3) , transparent rgba(255,255,255,0.3) transparent transparent;
  -fx-border-insets: 0 0 1 0, 1;
}

...

As another approach, I tried to define an arrow-shaped SVG path (using the -fx-shape CSS rule) for the breadcrumb buttons, but I encountered scaling issues. The width of the shape must depend on the width of the text displayed in the button, which could not be achieved using CSS.

.breadcrumbs .item {
  ...
  -fx-shape: "M 0 0 L 100 0 L 130 15 L 100 30 L 0 30 z";
}

So I defined the shape in Java code and used it to set the button's clip. The width of the shape was computed dynamically based on the button's text plus two magic numbers for the correct scaling of the shape. This worked nicely, but came with the burden of those two magic numbers that worked well only for the particular font I used.

int textLength = button.getText.length();
// scaling and padding are magic numbers
int mainWidth = TEXT_SCALING * textLength + TEXT_PADDING;

// build the following shape
// --------
//
// / /
// --------
Path path = new Path();

// start in upper left corner
MoveTo e1 = new MoveTo(0, 0);

// draw upper horizontal line
HLineTo e2 = new HLineTo(ARROW_WIDTH + mainWidth);

// draw upper part of right arrow
LineTo e3 = new LineTo(ARROW_WIDTH + mainWidth + ARROW_WIDTH, ARROW_HEIGHT / 2.0);

// draw lower part of right arrow
LineTo e4 = new LineTo(ARROW_WIDTH + mainWidth, ARROW_HEIGHT);

// draw lower horizontal line
HLineTo e5 = new HLineTo(0);

// draw lower part of left arrow
LineTo e6 = new LineTo(ARROW_WIDTH, ARROW_HEIGHT / 2.0);

// close path
ClosePath e7 = new ClosePath();
path.getElements().addAll(e1, e2, e3, e4, e5, e6, e7);

// this is a dummy color to fill the shape, it won't be visible
path.setFill(Color.BLACK);

// set path as shape for button
button.setClip(path);

// make button width equal to width of shape
double buttonWidth = 2 * ARROW_WIDTH + mainWidth;
button.setMinWidth(buttonWidth);
button.setMaxWidth(buttonWidth);
button.setPrefWidth(buttonWidth);

In yet another attempt, I implemented a breadcrumbs bar in which every breadcrumb is represented by three buttons: one for the left-hand part of the arrow, one for the central part (showing the text), and one for the right-hand part. This allowed me to work with fixed-width shapes on the left and on the right plus a regular-shaped (rectangular) button in the middle. The obvious drawback was the management of all the extra buttons including the correct update of hover effects for the outer buttons when the mouse was over the inner button, and vice versa.

I don't think there a JavaFX library out there that provides the functionality I need, so is there anyone with experience in implementing a breadcrumbs bar? Any thoughts or recommendations? Are there solutions that do not come with any of the drawbacks mentioned above?

Any help will be much appreciated.

Update (2013-11-06):

Andy Till suggested an implementation that won my heart: bind the x property of the horizontal lines in the arrow shape to the button's width. Since the button's width gets updated to fit the text, the shape will always have the correct width. No need for magic numbers. Thanks, Andy!

// build the following shape
// --------
//
// / /
// --------
Path path = new Path();

// begin in the upper left corner
MoveTo e1 = new MoveTo(0, 0);

// draw a horizontal line that defines the width of the shape
HLineTo e2 = new HLineTo();
// bind the width of the shape to the width of the button
e2.xProperty().bind(button.widthProperty().subtract(ARROW_WIDTH));

// draw upper part of right arrow
LineTo e3 = new LineTo();
// the x endpoint of this line depends on the x property of line e2
e3.xProperty().bind(e2.xProperty().add(ARROW_WIDTH));
e3.setY(ARROW_HEIGHT / 2.0);

// draw lower part of right arrow
LineTo e4 = new LineTo();
// the x endpoint of this line depends on the x property of line e2
e4.xProperty().bind(e2.xProperty());
e4.setY(ARROW_HEIGHT);

// draw lower horizontal line
HLineTo e5 = new HLineTo(0);

// draw lower part of left arrow
LineTo e6 = new LineTo(ARROW_WIDTH, ARROW_HEIGHT / 2.0);

// close path
ClosePath e7 = new ClosePath();

path.getElements().addAll(e1, e2, e3, e4, e5, e6, e7);
// this is a dummy color to fill the shape, it won't be visible
path.setFill(Color.BLACK);

// set path as button shape
button.setClip(path);

And this is what my breadcrumbs bar looks like now:

![breadcrumbs]({static}/images/breadcrumbs.png)

Zombieland

2013-08-23T13:29:00+02:00

Based on Jasper Potts' 3DViewer app (with support for importing OBJ, MAX, and MAYA files), we are developing an importer for FBX files into JavaFX. Below is a picture of our developer trying to understand the FBX format a screenshot of the extended 3DViewer app after loading an FBX file containing a zombie.

![A JavaFX viewer application showing a Zombie object imported from an FBX file.]({static}/images/zombie.png)

For implementing this importer, we took

Jasper's 3DViewer
The FBX SDK (comes with a zombie)
JNI (as a bridge between the FBX SDK and our Java code)

The importer currently loads all mesh, material, and texture information from all nodes in the FBX graph. We support ASCII and binary file formats.

We do not yet load light, camera, and animation data (the former two should be quite simple, while animation will probably take longer).

Our work is still ongoing, and I hope we can have an improved version soon. Stay tuned!

GUI Generation with JavaFX

2013-07-12T07:48:00+02:00

Recently, we have had a look at JavaFX libraries for generating UIs in a dynamic way. The idea is that we have a data model (a JavaBean class, an EMF model class, a model class with JavaFX properties,...) as input for some GUI builder that takes the model and generates a matching user interface with bindings between the model and the JavaFX controls. In a perfect world, this library would allow us to use a variety of data models and to customise the generated user interface.

We have had three candidates (all open source). As we use EMF a lot, two of them are based on EMF models.

1. FXForm2

http://dooapp.github.io/FXForm2/

This pops up early when you are looking for UI generation libraries for JavaFX. As the name implies, it is meant for generating forms from model beans.

![A sample form generated by FXForm2](https://camo.githubusercontent.com/178463387c4056f6bad98224fb117de90ffd462e/687474703a2f2f646f6f6170702e6769746875622e636f6d2f4658466f726d322f636f6d706f6e656e74732e706e67) Source: [https://github.com/dooApp/FXForm2/wiki/Style-your-form-with-css](https://github.com/dooApp/FXForm2/wiki/Style-your-form-with-css\[/caption\])

The main features include:

Automatic form generation and binding to bean properties
CSS support
Bean Validation handling (JSR 303)
Fields reordering and filtering
Tooltips
Localization

It takes three lines to add a generated form to your JavaFX scene graph:

MyBean myBean = new MyBean(); // get an instance of the bean to be edited
Node fxForm = new FXForm(myBean);  // create the FXForm node for your bean
root.getChildren.add(fxForm);  // add it to your scene graph

2. EMF Edit for JavaFX

http://tomsondev.bestsolution.at/2012/12/13/emf-edit-support-is-coming-to-javafx-via-efxclipse

This library has a broader scope than FXForm2 as it can generate an entire RCP application with a sophisticated visual model editor rather than just a couple of form fields. Or you can define your own controls and easily bind your EMF model instances to these controls, including support for drag & drop and undo/redo commands. An example of a user interface based on a "Contacts" model can be seen in Tom's video from the link above.

![A sample application using EMF Edit for JavaFX](http://videos.videopress.com/R4NEN2Ic/screenflow1_dvd.original.jpg) Source: [http://tomsondev.bestsolution.at/2012/12/13/emf-edit-support-is-coming-to-javafx-via-efxclipse/](http://tomsondev.bestsolution.at/2012/12/13/emf-edit-support-is-coming-to-javafx-via-efxclipse/)

3. EMF Client Platform

http://eclipse.org/emfclient/index.html

This takes the idea of EMF Edit one step further. As the ECP website says:

The goal is to provide a one-click application based on a given EMF model. Besides the EMF model, no additional components have to be developed or generated.

![A sample application as generated by the EMF Client Platform](http://eclipse.org/emfclient/public/images/application.png) Source: [http://eclipse.org/emfclient/](http://eclipse.org/emfclient/)

So when you have your EMF model, you are only one click away from a complete client application with a model explorer, an editor, a persistence mechanism, and a couple of other nice things. The editor for the model attributes is similar to the data form generated by FXForm2, but it is embedded in an entire application whose pieces are customisable and extensible.

The bad news is that there is currently no JavaFX support for ECP.

[Update]
As Tom Schindl has pointed out in the comments section, there is a first version of JavaFX support for ECP in one of the development branches of the e(fx)clipse project. See here.

4. Summary

We have not found the perfect solution yet. FXForm2 has a promising approach for easily generating forms, of which we are going to have a lot in our final product. It requires some tweaks to make it work with EMF models but that is a small obstacle. What worries me a bit is the maturity of this library: the latest version is 0.2.2. So is it ready for production use? Well, there's at least one real-world application that uses it.

EMF Edit is a very powerful framework but does not generate controls on the granularity we require for most parts of our application. It will certainly be of use for fancy features like drag & drop and the command stack. Maybe we can combine its strengths with the power of FXForm2 for generating really rich forms.

The EMF Client platform has a very appealing one-click approach but it lacks support for JavaFX. Once this limitation is gone, we may give it another look. A major question will then be how far we can customise the basic layout of the generated application which by default is very different from what we have in mind for our user interface.

Navigation

2013-05-23T16:43:00+02:00

Dear tree huggers,

I'm one of you. I like trees - preferably on the left of my application. They give me the impression that I exactly know where I am and where I can go next. Just recently I learnt how much I appreciate a tree as a navigation element when I worked with Confluence 5 for the first time and tried to replace the navigation tree on the left by the new space sidebar. It felt strange to me because I always looked for things that were not there. I tried to familiarise myself with the dynamic contents of the sidebar, but I gave up after a couple of days. The mental overload was too high.

Why was it so difficult to get rid of the tree?

As a developer I'm used to thinking about tree-structured data so trees are in my head all the time..
I encounter trees in many of my tools: my IDE, my file manager, my email client.

So I live in a forest. As this is my natural habitat, I'm pretty good with trees. I know how they work and how to find what I want as they are an accurate representation of my mental model in most cases. In other words: my brain was not ready for that sidebar thing.

But there are other people in the world. People who tend to think in different ways than a software developer. Scary, I know.
Some of these people use the software I build and at the same time do not like trees, in particular as navigational elements. Their mental model is different from mine, maybe workflow-centric rather than data-centric. They might not think of the data they're working with in the same way I do. It's not unlikely that some of them do not think about their data at all but simply execute a given (no-brainer) task.

In these use cases, a tree is of limited value at best and can even be harmful when the user is faced with too many options and too many paths that do not get him where he wants to go. It is better to restrict user choices and thereby make the user focus on the current task. A context-sensitive UI helps guide the user through the given workflow whereas a tree might encourage the user to meander through the application and explore things are are not to be explored.

No wonder a full chapter of Designing Interfaces is about navigation. No wonder the first paragraph in Microsoft's developer documentation on tree views is titled "Is this the right control?" No wonder even a small thing like a breadcrumb bar deserves a full-blown description on patternry. Because even with a limited set of navigational widgets, you have many options to do it right and many more to do it wrong. There's more to navigation than just putting an extra menu bar somewhere. Or worse: a tree.

Having written this already helps me to let go of the tree. Now it's time to look for alternatives. In my JavaFX world this means goodbye TreeView and hello ...

I'll save the answer for another blog post. Stay tuned.

Designing Interfaces with JavaFX

2013-04-24T05:35:00+02:00

Good design makes you happy. I was happy when I first saw an iPod, and even happier when I first held it in my hand. Sometimes, you will not notice the positive effects good design has on you, but you will always notice the effects of bad design.

I'm a software developer, not a designer. I've known that ever since the late 90's when I repeatedly failed at impressing my friends with my one-minute Flash intro for my two-paged personal website over at Geocities. Why haven't I learnt my lesson and still deal with visual design today? Because I believe a good visual design is essential for any software that has a user interface. If you don't believe me, read what Ben Northrop wrote about the Halo effect.

Designing a great user interface is hard, in particular when you're more excited about your domain models and your test coverage of 101%. Unfortunately, even smart developers can write bad user interfaces, as Karsten Lentsch explains in this Jax session (in German). It takes a lot of time, training, and user research until you come up with a user interface that doesn't suck and that is tailored to your user's needs.

If you're lucky, you have a thorough and well-organised process for that. If you're even luckier, you work with a professional user interface designer who can do all those things for you. That's also a great way to find a scapegoat in the end, by the way.
But even working with a designer doesn't save you from analysing your information architecture and from organising your application content. There's still a lot to do.

And this is what I'm currently doing. My job is to develop a user interface for a client-server business application offering a compelling user experience. Here's what I have so far:

a visual designer
the great book Designing Interfaces by Jenifer Tidwell
a prototype of the client
four personas
and lots of ideas

Oh wait! I have some fantastic tools, too. They are centred around JavaFX as we saw the potential of this platform pretty early in our project and it has not disappointed us since. JavaFX supports the design process by its separation into layout (FXML), logic (controller classes), and appearance (CSS) tools. All of these can be tied together nicely with our underlying Eclipse rich client platform thanks to Tom's e(fx)clipse project.
There's nothing like a good tool chain. Well, a good design maybe.

So I know JavaFX allows me to develop great user interfaces, but how do I actually do it? How do I put the pieces together? This is the main question I'm going to ask my designer and myself again and again over the course of the next weeks and months. We have a general layout for our application window, which is fairly common in other applications, I would say, but one of the first things to decide now will be the navigation. How can the user go forward and backward in our application? How does she know where she is? And how do we tell her what else there is?
Exploring these questions is going to be part of my next blog posts. Stay tuned.