So why is this a big deal to me?

Last weekend was the third OggCamp conference, OggCamp 11, at the Farnham Maltings in Surrey. Two years ago, at the first OggCamp (a one-day event at the Connaught Hotel in Wolverhampton), we invited Andy Stanford-Clark (@andysc) to be our opening keynote speaker. Andy co-invented the MQTT messaging protocol about 10 years earlier and, while there was a server implementation of MQTT (Really Small Message Broker; RSMB) that you could download for free from IBM’s website, it was proprietary and there was no open source implementation available.

Andy wrote a new presentation, especially for OggCamp, describing the geeky innards of his Twittering house (as seen earlier that year on the BBC). The presentation was a fantastic kickstart to the day and (somewhat predictably for a conference with its foundations firmly in the open source world) Andy was questioned about what bits of his home automation system were built on open source software and open standards. The one significant part of the system that was proprietary was RSMB (the core part that enabled all the parts of his house to communicate).

Then OggCamp started, we had a good time, and we went home exhausted but happy.

And then, just two weeks later, Roger announced that he’d registered a new project called Mosquitto (as in MosQuiTTo) on Launchpad. He’d been inspired by Andy’s talk at OggCamp to write an open source alternative to RSMB. Within what seemed like days he had a working bit of code which was taken up and tested by others in the open source community and hardware-hacking communities like Homecamp.

I cannot claim any credit at all for all the hard work that Roger and others put in developing and testing Mosquitto. I’ve always been proud, though, that Mosquitto was born at OggCamp – we played our small part in helping connect the previously mostly corporate/business MQTT with the open source communities.

That Facebook announced they were adopting MQTT for their new messaging system the day before OggCamp 11 meant we could vicariously revel in Roger’s glory while we tried to find out just whether Facebook had adopted his code or their own implementation. The answer seems to be somewhere between the two.

And while I’m proud for OggCamp (of course), I’m also excited for Roger in his own right that his name is now in the licence agreement of apps from the mighty Facebook – that kind of recognition for your hard work must be such an amazing feeling!

This week, I decided to install Rational Software Architect so that I could try out (again) the User Interface Generator which comes with IBM InfoSphere Master Data Management Server–and other products too, I believe. You can find out more about what user modelling is and how you can use the UIG in RSA (lovin’ the IBM TLAs yet? ) in this series of articles on developerWorks. The rest of this post is not about the UIG but about how to install and configure Rational Software Architect for WebSphere, which is relevant to anyone wanting to do this, regardless of whether they’re wanting to use the UIG.

So I wanted to install Rational Software Architect (RSA; an Eclipse-based piece of software) on to Ubuntu, which is what I run on my work machine, a Thinkpad T61p. Not only that, but I wanted to install RSA for WebSphere, which includes WebSphere Application Server Test Environment (WAS). This meant that I was installing not one but two pieces of software that are not officially supported for Ubuntu Karmic (or indeed for Ubuntu/Debian as far as I know). But as the Linux binaries are .bin files rather than .rpm, life should be easy.

And indeed installing it is. It’s when you want to create a WAS profile that the fun starts. And that’s where it had all fallen apart for me when I tried much the same exercise this time last year. That time, I gave up.

This time, I tweeted my predicament, knowing there were people who might know the answer. Unfortunately, the person I thought might know the answer had failed in much the same way as me only 6 months ago and was now relegated to using Windows. Still, others came back to me, including Gavin Willingham, who I’d not met before but who works possibly 10 minutes walk from my desk, and Jay Limburn, who I’ve known for a year through Twitter and internal instant messaging but, though he also works max 10 minutes walk from my desk, never met in person until yesterday (he’s shorter and blonder in real life).

So, if you’re trying to install RSA for WebSphere 7.5.4 with WAS Test Environment 7.0 (other versions probably work the same way), I’ll end the suspense and start here:

Running the installer

1. Download the many parts of RSA for WebSPhere 7.5.4 and WAS Test Environment 7.0 with licence/activation bits and pieces. (NB this isn’t free software; you have to buy it from IBM so I’m assuming you’ve got that far by now.)
2. Extract all the zip files. If you do a right-click > ‘Extract All’ on the zip file in Ubuntu, the extraction tool doesn’t like to overwrite directories of the same name, so you end up with directories called ‘RSA4WS’, ‘RSA4WS (2)’, ‘RSA4WS (3)’, and so on when actually, you want the contents of each of those directories to be in the same place. So move the directories around so that you have just 3 directories:
   • RSA4WS (contains 7 directories called ‘disk1′, ‘disk2′, ‘disk3′, etc)
   • RSA4WS_SETUP (don’t do anything with the contents of this one)
   • WAS70 (contains 4 directories called ‘disk1′, ‘disk2′, etc)
3. In the RSA4WS_SETUP directory, as sudo, run launchpad.sh to start the installation process.
4. Follow the installer through but when it asks you for a user name and password to create a WAS profile, select that you will create a profile later and that you don’t want to create one now. The installation should run cleanly (if you let it try to create a profile, it will fail part-way through the installation).

You should now have a nice installation of RSA with WAS on your Ubuntu box. Next, you need to create a WAS profile so that you can use the in-built WAS server for development and testing (or in my case, to run the user interface generator).

Creating a WAS profile in RSA

There are a few things that prevent this just happening (some are generic Linux things and some are specific to Ubuntu):

• On Ubuntu, /bin/sh uses dash, not bash, but WAS scripts seem to use bash specifically, so the profile creation scripts fail.
• Something extra that I have no understanding of is required in the eclipse.ini file (which is key in starting the Eclipse-based RSA environment).
• The default location in WAS’s profile creation wizards is in the /opt part of the main file system which you typically won’t have write-access to as a normal user.

So here’s what you need to do (at least, this is what I did and hopefully will work for you to to get a running WAS server in RSA):

1. Change Ubuntu’s /bin/sh to use bash instead of dash.
   In a terminal (sorry it’s the command line but you’re changing some system settings here that you would very very rarely have to do normally, or just if WAS didn’t specify bash specifically), run the following command and select the bash option as the default for /bin/sh:

   sudo dpkg-reconfigure dash

   (As pointed out by @oldmanuk (Dom Evans), this is the proper way to reconfigure where /bin/sh points to (though I used Gavin’s method). )

   After you’ve created your WAS profile and everything’s up and running nicely, run the command again to change back to using dash. The benefit of using dash is speedier boot time, which is lost if you leave the setting as bash (see https://wiki.ubuntu.com/DashAsBinSh – thanks Dom).

   Now you’ve sorted out the bash/dash problem. One down; two to go.

2. Check that you have a version of xulrunner installed (I have no idea what xulrunner is for but, looking in Synaptic Package Manager, my Ubuntu installation included xulrunner-1.9l.1-gnome-support, but not the xulrunner package itself, which seems to be fine for RSA purposes).
3. In your RSA installation, find the eclipse.ini file. I installed RSA to the default location so mine was in /opt/IBM/SDP. In a terminal change to that directory:
   

   cd /opt/IBM/SDP

4. Open the eclipse.ini file in a text editor, such as gedit:

   sudo gedit eclipse.ini

5. Add to the end of the file the following line:

   -Dorg.eclipse.swt.browser.XULRunnerPath=/usr/lib/xulrunner

   (From an Ubuntu forum post.)

   Now you’ve sorted out the ‘xulrunner’ problem and you can actually start RSA. Two down; one to go.

6. Start RSA from the Applications menu: Applications > IBM Software Delivery Platform > IBM Rational Software Architect for WebSphere 7.5.4.

7. RSA should suggest a directory in your home directory in which to put the RSA workspace. That’s fine; I just accepted the suggested location.

8. When RSA has opened (NB, my Welcome view doesn’t load – I don’t think that’s a problem), give it a moment to think, then it’ll pop up a wizard to create a WAS profile.

9. In the profile wizard, clear the option about security unless you know what you’re doing with WAS security and have a need to use it in a development environment.

10. You’ll also notice that there’s a warning about the currently selected location for creating the profile. This is the third problem I listed above. The default location shown is for creating the profile in the installation directory of RSA. Change the location to a directory in your home directory. For instance, I told it to use /home/laura/IBM/profiles.

11. When the wizard has created the profile (which will take a few moments – you can see the activity in the bottom-right of the RSA window), the default server for the profile is listed in the Servers view on the right-hand-side of the RSA window.
    The server is listed as ‘stopped’.

12. Right-click the server then click Profile. This opens a dialog box about the profile; I just accepted the defaults then it failed to start the server (the error said it had failed to start within 300 seconds). But when I repeated this step, the server started.

And that’s as far as I’ve got but it’s further than ever before. If I come across any more gotchas, or take some screenshots, I’ll update this post.

Yes, it is a pain to have to do all this but remember that WAS isn’t supported (as far as I know) on Ubuntu Karmic. If you want it easier, install it on something that is supported. If you want it on Ubuntu, I hope this post (and the others I cribbed information from) helps.

Eightbar (as in the IBM logo which is known as the ‘eight-bar logo‘) is a community of people in and around IBM Hursley who are into cool, techie or creative things, either in work, out of work, or both.

The thing about large corporations is that people forget that most of the most amazing things that happen in those corporations come down to individual people just getting on and doing them. It’s easy to think (from inside and outside) that employees are ‘just a cog’ and everything is decided from on-high and nothing can be done without getting it approved in triplicate.

In fact, while a corporation’s culture can play an important part in encouraging and supporting good ideas, it’s the individuals who try them that make the difference. Whether that’s coming up with a better way to do something in your ‘day-job’, or writing a cool app in your evenings which subsequently gets so many downloads it gets incorporated into a real product (several people I know spring to mind immediately), or you just do something like running Linux as your desktop when hardly anyone else is and then helping others do the same.

That kind of innovation and adventure just doesn’t happen because someone in a suit on high tells you to do it. It comes because you think it’s a good idea and decide to give it a go.

The motivation behind eightbar was the realisation that there are loads of cool things happening around IBM Hursley that no one ever finds out about. So 4 years ago the eightbar blog was started.

Until today, I’d never contributed to it because I was too intimidated – but as one of many people around Hursley who attends conferences and unconferences, maintains (mostly) a blog, twitters, and likes to talk to other people who are into cool and interesting stuff, I figured I should make the effort (and the lovely @andypiper hinted very unsubtley that I should too).

So I did.

Usability is more than a pretty face (or a good user interface)

(Yay! Someone else who gets annoyed when people assume usability == user interface.)

InfoSlicer is a small application that enables you to download articles from Wikipedia, drag-and-drop sections of them to create new articles, and then publish your collection of articles for others to install or view on their own laptops.

The ideal of InfoSlicer is to support teachers in schools where access to books is limited. They can use InfoSlicer to quickly obtain content from the internet (maybe at a cybercafe rather than at the school or at home) and to create customized versions of the information that are suitable for their pupils and can be viewed with needing access to the internet.

Since completing the initial prototype, however, it’s become apparent that InfoSlicer could actually be more useful to the pupils themselves than just as a means to receive information created by their teacher. The children themselves could use InfoSlicer to download articles and then learn how to re-organise information for a specific audience or purpose and how to attribute someone else’s content without plagiarising it; the outcome of creating the articles is then less educationally important than the process of doing it.

So if you have Sugar, download the first version of InfoSlicer and give it a go (or just find out more) from: http://sugarlabs.org/go/Activities/InfoSlicer

Update 13th April 2009:

On re-reading this article (which was intended to be just a short intro to publicise InfoSlicer), it sounds as if I wrote the software myself! I didn’t. It was the outcome of the brilliant efforts of the InfoSlicer Extreme Blue team during their internship at IBM Hursley last Summer. Here’s a photo of the team at their Expo stand in Germany:

Jessica Vernier, Matt Bailey, Chris Leonard, Jon Mace

So, I started learning BSL in September 2006 when IBM put on courses for employees at Hursley. We had two hours of teaching every Wednesday morning for 30 weeks, which culminated in being CACDP BSL Level 1 certified.

Jeff, our tutor, is Deaf and taught us using a combination of signing, speech, writing on whiteboards, slides, and humour. Different tutors using different communication methods – for instance, BSL tutors don’t have to be deaf themselves, and some use speech and some don’t. Jeff doesn’t really lip-read so we got lots of practice at signing when talking to him during tea-breaks.

During the course, Jeff taught us a bit about Deaf culture as well as the language. This built on the deaf awareness workshop that we had attended early on in the course. In the workshop, another man (also deaf but deafened later in life; he speaks, uses a hearing aid, and lip-reads) taught us about what it’s like to be deaf, how (as hearing people) to communicate with deaf people, what the Deaf (signing) culture is, and attitudes of deaf people to their deafness.

I really enjoyed the course. It was difficult at first to deal with learning something without being able to write it down (BSL notation is a skill all to itself!). So learning to rely less on written notes was useful too. Learning BSL has been really useful, in particular in talking to my friend Ben at work who is profoundly deaf (without speech) and whose first language is BSL. It’s also handy in meetings or in the noisy canteen to be able to sign to colleagues.

I think it’s really cool that we could learn BSL at work. Aside from the actual language, learning about the Deaf culture and deafness in general has given me a different perspective on things and broadened my understanding of other people. In terms of my day-job, I have a better understanding of the issues around Accessibility.

For instance, here’s one of them….

Did you know that if BSL is your first language (and, therefore, English your second), written transcripts are not necessarily sufficient for a Deaf person to understand an audio recording***? The concepts and grammar of BSL are so different from English that moving between the two can be very difficult. That’s why you get BSL interpreters signing on TV (eg BBC News 24) instead of just providing subtitles.

A lot (a *lot*) of people don’t know that.

Update (16th July 2008):

***This is not to say that written transcripts are a waste of time, nor that Deaf people can’t generally understand written English! Also, if you can provide written transcripts, they provide a means for other people to translate those transcripts to other languages. So projects like this one are really cool: https://launchpad.net/~transcribers. For a start, a written transcription might one day be able to be converted automatically into BSL…(My SiSi blog post)…

I’d heard a bit about the Hursley-based SiSi project from a friend who was mentoring the team, so I moseyed on down to Hursley House and spent a good hour-and-a-half visiting the Expo stands and hearing about those and other projects from around IBM sites in Europe.

I’ve been learning British Sign Language (BSL) for about a year and, having learnt just the basics about how to communicate in BSL (that is, it’s not just hand signs or fingerspelling but also facial expressions, lip shapes, and the spatial location of the signs that matter), I couldn’t imagine how an avatar could convincingly sign – especially not translated in real time from speech, which is what the SiSi project aimed to do.

The SiSi team’s demo blew me away. They use a third-party piece of software to convert speech input to text. The text is then sent to the client machine (I think) where an avatar signs the text in BSL or American Sign Language (ASL), depending on the language you selected. I can’t remember any more of the technical details than that but the demo text they tried was translated to BSL at a reasonable speed, I thought (probably as fast as a human interpreter). The demo was on a local system but the students reckoned it did okay over remote systems.

The project was done with the University of East Anglia and the RNID (Royal National Institute for Deaf people) who supplied the database of signs (which I guess is probably a database of video clips and associated labels). I’m not sure who marked up the signs in Sign Language Markup Language (SLML), a form of XML, but I expect that’s the most intensive part of it.

The great thing for IBM and the Extreme Blue scheme is that, like last year’s LAMA project the SiSi project has attracted loads of press coverage, here and around the world.

SiSi aside, there were loads of other cool projects including (you can probably spot a theme in my interests here!) the Accessibility in Virtual Worlds project. For a change, the virtual world concerned was not Second Life but, instead, Active Worlds. Active Worlds enabled the project team to devise a way to mark up objects around the world using XML so that blind people can walk through the virtual worlds using sonar. The user wears headphones (or has speakers set up) and the nearer something is, I think, the louder the sound (or something like that).

I came away from the Expo with a handful of really professional-looking Moo cards and leaflets from the stands I had time to visit. I think the most amazing thing that occurred to me about the Expo was the amount and quality of work that the students were able to produce in just 12 weeks.

I was Chairing the HCI Practice Day (the Thursday) of the conference so it was all a little bit hectic but still, as usual, a lot of fun.

This year, in line with the times I guess, there was a fair number of papers on Second Life and other virtual environments, including one about BDSM in Second Life. In fact, there was definitely a bit of a trend this year for erotic HCI… (that, as a statement, either makes HCI cool, or it just goes to show that academics can make *anything* boring ).

There also seemed to be a fair amount about emotions – that is, how we engage with technology; eg why we happily waste an entire evening on Facebook or watching random videos on YouTube. Web2.0 was also in there, of course.

There were also some papers on bluetooth, several (as usual) on eye-tracking, and stuff about Accessibility, usability of the Web, methods of evaluating interface usability, and so on.

IBMers featured quite heavily in the HCI Practice Day (as you might imagine):

• Mark Farmer (IBM Warwick) introduced the IBM Task Modeler tool that he develops (the link takes you to the Task Modeler page on Alphaworks where you can download a copy to try yourself).
• Colin Bird (Master Inventor and Information Architect at IBM Hursley) followed up Mark’s introduction with a presentation about how you can (and we do) use Task Modeler to support information architecture: to model user tasks and create the navigation for information centres.
• Ben Fletcher (Senior Inventor at IBM Hursley) did a great presentation on deafblind technologies, including the possibilities of virtual worlds in supporting deafblind (and deaf or blind) users.
• Me (Technical Author at IBM Hursley) – I was raconteur for Alan Dix‘s panel discussing the HCI issues in Web2.0 technologies.

The keynote speaker for HCI Practice Day was Jared Spool (the usability guru who isn’t Jakob Nielsen – and is much better and more credible, IMO) who moved heaven and high water (kind-of) to get here. He did a fantastic presentation that was very very funny and entertaining while being relevant and interesting too. He also attended as many of the other conference sessions as he could and participated by asking questions and making suggestions.

In fact, all the keynote speakers were great this year. Sometimes keynotes fly in, do their thang, then collect their expenses and go. All three (the others being Stephen Payne from Manchester Uni and Elizabeth Churchill from Yahoo!) all got involved in the conference, especially Elizabeth who was able to stay for the whole conference and seemed to be on every discussion panel going!

You can get the full proceedings of HCI 2007 (and, at some point, previous HCI conferences too) from the BCS eWIC site.

As a delegate, I also got the full proceedings as pdfs on a funky little USB drive, which I like.

It’s not long now until the call for papers will go out for HCI 2008 (to be held in Liverpool, City of Culture). If this blog is still active by then, I’ll post the call here. I encourage you to get involved in HCI – it’s more than user interface design or usability; it’s also about being innovative in how to design technologies for human beings.

Anyway, I realised that most of what I write there (all of two posts so far!) is fine to go on my external blog too. Hence, two blogs for the price of one! )

Which explains today’s more prolific blogging than usual.

If you happen to install WebSphere Message Broker, the product documentation typically instructs you to verify your installation using some of the samples (I know; I wrote some of it). But this method assumes that you have installed the Message Brokers Toolkit, the GUI development environment. While I think this is a fair assumption if you are coming new to the product to try it out or just have a play, if you’re more hardcore, you maybe just want to install the runtime (ie the bit that isn’t the Message Brokers Toolkit and that does the actual broker work).

JT’s article basically steps you through how to verify your installation without using the GUI.

I’m also happy to plug his article because he cites my IBM Redbooks publication WebSphere Message Broker Basics, as does (I just found) the Wikipedia article about the product.

Not that I’m easily flattered or anything…