Someone else wrote asking (very nicely) for advice in selling photographic prints. I haven’t a clue as it happens, so go buy one… his site only gets a few hits a day…

I learned to play Democracy by Killing Joke on the guitar as I watched swifts (I think) reel above me in sky, eating flies I imagine.

Someone else wore socks that were designed to look like bowling shoes.

For most of the day I felt guilty.

]]> http://www.theotherblog.com/Articles/2010/07/08/funny-sort-of-day/feed/ http://www.theotherblog.com/Articles/2010/06/22/killing-joke-in-excelsis/ http://www.theotherblog.com/Articles/2010/06/22/killing-joke-in-excelsis/#comments Tue, 22 Jun 2010 12:50:45 +0000 tom http://www.theotherblog.com/?p=5085

Anyone that knows me knows that this year I have been completely obsessed with listening to Killing Joke. I’ve listened to an album a day for almost every day for six months on my walk to work. I’ve even been studying Killing Joke albums that I loathe, the ones between 1983 and 1993 when Jaz Coleman had a bit of a breakdown of sorts, had his hair permed, grew a moustache, thought the world was about to end and Killing Joke the band went all Simple Minds on me. Bad times, bad albums… still.

I don’t consider myself the OCD type at all but I have to admit that it has to be slightly obsessed, to listen to stuff you don’t like in the hope that somehow you learn to love it. It’s like torture. You can browse both the good and the bad Joke stuff on the Dipity timeline I made of Killing Joke’s career here… or in my You Tube Killing Joke playlist here (which also has relevant bands in too)… ahem… Not obsessed at all.

I don’t really know why I like Killing Joke that much. Of course, everyone has the band, which for them just seemed to have meant so much, you know, when you were 14 or 15… don’t they? For most people I guess they grow out of it, or the band turns crap, like Adam & the Ants and Death Cult did or they break up or music simply starts to matter less.

All of these happened with Killing Joke, but in 1994 whilst I was stripping wallpaper, they got back together again and started making that strange discordant noise that only they can do … listen to Whiteout (spotify link) or Exorcism (spotify link) from the Pandemonium album and tell me it’s not the most noisiest fantastic din ever created). The latter was recorded in a Great Pyramid and has coughs in (like a lot of Killing Joke songs)  dontchaknow.

I mean, try to these two live songs below on YouTube (only 87 views btw), Requiem (one of their first) and Total Invasion (one of the most recent) and they’re pure punk… with a hint of disco. Guitars you can dance to. If you wanted.

I relax to this (The Hum from Revelations). Killing Joke are my whales singing… my birds tweeting in a forest… by a sea with crashing waves…

And so, on the longest day (yesterday), Killing Joke have release an EP called  In Excelsis (go listen for free) which I’m not so sure about.  It has a dub track which is fun and Endgame sounds to me like Lemmy from Motorhead (no bad thing) but it’s not quite Killing Joke, or rather the Killing Joke I love which falls between 1980 - 1984 and 1994 and now ( please don’t listen to Love Like Blood or Sanity or America which are so bad I still get embarrassed for them). I’ve give it another few thousand listens before I pass judgement but at the moment… I’m slightly disappointed…

Anyway, I have been both worrying about my ever increasing Killing Joke obsession and at the same time writing a TV script (a bit like those wonderful Rock Family Trees show) to explain “Why Killing Joke Are Ace” to my mate Dave because I’ve got tickets to see them in October and he needs a little persuasion that they, and not Pink Floyd are the best bend that ever walked the earth - even if they did go a little potty in the 80s… I mean, who didn’t?

Spheres of Collaboration, CTP Report and Use Cases

All about the Collaborative Tools Project… as a PDF.

I’ve been thinking about how best to “tell the collaboration story” from a practical standpoint. And it occurs to me that often people look at the tools and approaches and ask “what’s in it for me?” … when the better question might be, “what’s in it for everyone else?” …  The irony here being that you are one of “everyone else”… like everyone else.

The source Omnigraffle files are here if you’d like to extend or improve on them. My goal was to, around the outside show some real world practical examples of stages of collaboration, from twitterfalls (or instant backchannels) to examples of social networks like LinkedIn or Mendeley that clearly demonstrate “what’s going on”… but ran out of time.

Almost 10 years ago, this blog was developed in Zope. It wasn’t big or clever and it’s why (to this day)  my URLs still have “/Articles/” in them. That was a Zope folder and I thought it made sense to name it properly with none of this lowercase nonsense. And then (was it 2003ish?) Plone came along and although I tried and tried and tried, Plone beat me… I didn’t get it. To be fair Plone didn’t offer me what I needed, which was a simple prototyping environment, I really didn’t need workflow or roles or permissions…. I just needed python at the end of a URL.

Feeling more mortal than I’d like after failing to get to grips with Plone, I sloped off elsewhere… doing more UI or usability work and much less development in general, but using still python for unusual web crawling projects. When tools like TurboGears came along I tried them, occasionally I had another run at Plone, I mean, how hard can Plone really be?

Then Django made web development both easy AND pythonic. Django has its quirks but, and this is probably because Django grew out of the publishing industry, the documentation is lots of different shades of lovely. And the documentation works (often because it is tied to version numbers of the software).

Now I know that Plone is a different beast from Django completely but I now really want and need to learn it and so this week I’ve been sitting down, gritting my teeth ( not a requirement I know but it helps) and doing Plone tutorials. So far, it’s not been all good… Did you catch my classic English understatement then?

I started with Professional Plone Development book, it’s title seemed to have all I want… except the professional bit I guess. To begin with, I couldn’t get past the first three lines of code, which gets easy_install and installs a Zope Skeleton. I couldn’t do this because I was working on a Mac with python2.5 and 2.6 and Plone only works on 2.4. So I tried to get 2.4 and make it the default python but it’s harder than it sounds. Wrong libraries get loaded, paths are wrong and there seems to be a history (and related mess) about what is the best way to distribute python code… I gave up…

Next I thought that with a copy of VirtualBox and Ubuntu I would be able to run a virtual machine with Plone on it. This kind of worked but the install was lengthy and getting my cursor back from the VM was a black art and usability nightmare so I went with VMWare. Wouldn’t it be possible for there to be VMs for Plone with all the right versions of everything with everything in the right place? Anyway… back to the book…

I then get to what is maybe the fourth line of real code to use in the book…

paster create -t plone3_buildout myproject

… and I have no idea what paster is, or what it does. I thought it might be one of those websites that you paste code onto and share (and then run) … I have no idea what a plone3_buildout is… or what -t is for…basically, a whole heap of concepts are dumped on me that I have no idea about. Why is there a tool called paster anyway? Why isn’t this part of Plone itself (a bit like the ./manage.py do_something format that python uses?

Then we move into a discussion about getting Eggs from the Cheese Shop? …and PyPi and aargh! The language and metaphors are a mess!!!  OK, I’ll admit, until now I’ve hated eggs because they rarely worked and a folder of code seems a much more sensible way of dealing with things, I’m starting to see why Eggs are important but can we now start again with the whole terminology thing. What started out as funny is now, frankly, not.

Then, whenever I want to install any extra bits to Plone it seems that there is “more than one way to do it” with a Buildout or old fashioned folders in the Products folder. Can’t we, for the sake of a tutorial pick one and stick with it.

In the book we are then invited to write Unit tests before we’ve written any code? Now don’t go getting all purist on me, but the Unit Tests didn’t work, were complicated and frankly…

The then shows Code fragments that break across pages making them un-copy-and-pastable (they lose tabs too) which makes every line a chore.

The buildout.cfg file/ approach is introduced with no explanation and looks slightly familiar as a format, but do tabs matter? I’ve no idea… I added comments that broke the buildout process (yes).

Then using /bin/buildout -No doesn’t work…some product complains about needing to go online

I’m still with the book an the example is Creating a “Policy Product”… A what? At this point I’m not sure if a Policy is concept from Plone or what..

I then found a bug in the code in that I found a permissions module that’d been renamed from something much longer meaning it couldn’t addPolicy(). Now at this point I’m guessing that my version of Plone isn’t compatible with the tutorial.. maybe because eggs have automatically updated themselves… who knows.

And so.. after much sweat and tears… the first code from the first chapter wouldn’t install.

I moved onto another book from PACKT… and found similar problems.

Moving to the hopefully more up to date online tutorials, I then decided to follow the DVD Collection Archetype creating tutorial, only changing the names/namespaces to be something I actually need ( a “Minutes” object for recording meeting notes). It didn’t work and I couldn’t work out why.

Giving up on writing my own Products, I thought it might be best to install some other ones and see how they work and hope that would bring some insights..

I tried…

Faculty Staff Directory (wouldn’t install). I’ve left a traceback on their Issue Tracker. I really want to use this product.

Feedmixer (wouldn’t install). Thought this would be easy. Again, I’ve left a report.

Collective GEO … yay! … This worked almost first time and I now can add coordinates to any Plone object and have it displayed on an OpenStreetMap. THIS is what Plone development should be like… thought through and reliable.

And now I’m getting random …

ImportError: cannot import name IObjectMovedEvent

So… after a couple of weeks of getting stuck into, but mainly stuck by Plone I need to take a breather and start again.

Plone people, there really is a need for …

• A version of Plone that exactly matches clear and simple tutorials. Being simplistic isn’t the same.
• Stopping the assumption that I’m a unix master. I’m not, but I’m willing to learn the bits I need if you take it slow.
• A re-thinking of terminology used for the various bits and bobs (I mean… “paster”…? wtf?)
• Some consensus about “the best way”… I don’t want confusion I want guidance from people who know better
• The UI of the management screens in Plone could do with a big hand. Currently it’s like a tool shed with all the tools laying on the floor, on the bench on the window sills. Maybe building some helpful help text into these screens wouldn’t go amiss. Maybe group some relevant tools together.

Interestingly ANY Plone object can have coordinates (rather than there being a separate Location object type).

My interest in showing maps within the Collaborative Tools Project isn’t actually rooted in maps as such, although they are lovely things… maps aren’t they? What I’m also interested in is…

Flashmobbing and “Good Enough” User Generated Mapping Data

Parts of the University’s campus are very well defined in the OpenStreetMap map and other buildings such as the recently finished Berrick Saul that I look out on are missing completely. I suspect the trails on the map are the result of a StreetMap Party (but I don’t know) where mapping nuts gather with devices and leave their trail(s).

This is fascinating firstly because it happened in the first place, but also because it’d be great to help make it happen again. Or to put it another way, the University of York is awash with energetic young people with a GPS device in their hand… surely with a little gentle persuasion a little effort from a lot of people, say, adding a room’s geographic location could add up to something bigger, better and more useful… Like for example…

…Accessible Maps

There is a project looking to marry the maps that our Estates dept have, with data about the paths, roads and layout around the University to create something that is helpful to people with disabilities. It’s a lot of  (technical) work but I believe that in the meantime we could have a prototype (maybe an iPhone app) that gave enough information to help shape the design of a better system.

Or to put it another way… I suspect that, in order for map to be genuinely useful, it needs to be mobile (and probably not paper) and allow user contributions “Don’t go this way - there’s a massive puddle of duck poo!” etc.

Which leads on to…

… Instant Backchannels…

Or put it another way… Have you ever been at a presentation where the presenter has announced a hashtag that the audience can use to collect tweets from everyone? It’s quite an interesting experience, for the presenter it can be a challenge receiving instant feedback but it also gives everyone outside the room a flavour of what’s going on inside the room.

“Instant backchannels” can be a mess to invoke … almost as if once everyone is in on the secret it works magically, but if they aren’t it can’t. For example, people in the audience might need a twitter account, and to know about tools that pull hashtags searches together… I might want the backchannel displayed alongside (or over) my slides being presented… I also might want some “background” material available too in say a wiki. I’ve even remotely attended conference sessions where the presenter’s talk is video streamed (I think with uStream) and I even got to ask a question via Twitter.

So I thought how easy would it be to create an Instant Backchannel for every room in the University? The idea being that it was there, all ready to go whether you use it or not. The point being that it probably wouldn’t get used much, but when it was, it would be invaluable.

What’s most interesting (to me) is that in order to work it doesn’t require accurate geographical data, it only needs a list of room names BUT the geographical location can be added later, as an when… as simply as tapping a button on your iPhone. It might look something like this…

So there… a very simple idea that might help people find their way around campus, participate remotely in presentations, get in touch with people who aren’t in their office, find information or people based on the discussions that other people had in a room the week before you were in it…

…that involves Crowd-sourcing the geographical location of important places, QR codes to bring the digital into the real world and back again, twitter, mapping, iPhone or mobile apps, accessibility issues and user generated “good enough” content.

Sounds like a plan.

• should be rolled into the development of Cyn.in. A good example of this is the UI for the Home Page when we have more than 40 spaces. It’s already apparent that you need a very different user interface for working with 4,000 people as opposed to 40. It would be good if Cyn.in “adapted” based on usage so that as the site gets bigger and more complex, the UI reflects this. One very easy (and cool) addition might be an “I’m going” button on events.
  
• already exist and are are about integration. Good examples of this are the FeedMixer product for displaying RSS feeds and the Faculty / Staff Directory product which lets people add more information about themselves in their profiles. the Anz.Jabber instant messaging product has caught my eye too.
  
• can be easily developed myself (or with a little help from my friends). I am currently looking to maybe use Archetypes to create a Location content type. There are others out there such as PloneWorldKit (uses Flash) and GeoLocation (uses GoogleMaps?) but I’d quite like one that worked with OpenStreetMaps and has less strict GEO-focussed functionality.

Other features such as “can I have a taxonomy of jargon-related abbreviations and acroynms that get appended or search-and-replaced when editing a document”… seem hard to define and even harder to work out if they are a low-level feature request to Cyn.in or something I could easily create myself. This isn’t essential, but it would be nice and could be made to work for peoples’ name, locations (such as room numbers) etc.

Another feature I find hard to explain is to do with how groups (and Spaces) are administered. I would quite like to create a Group/Space that automatically has a “private sub-space” … but that this isn’t set to be the default (otherwise people tend to work away in privacy and obscurity). I’d then like a “knock to join this group” button which the administrator manages. The alternative to this is a massive top-down administration overhead that I just don’t think will work.

One feature I would definitely like is the “First Use Ten Second Tutorial”… so that having accepted an invitation to a Cyn.in community you are presented with a one-page (maybe two) tutorial saying “this button does this… and don’t swear please (or similar)”…

Lastly, as I said, I don’t think the design of Cyn.in’s Home Page works very well for larger organisations and yet vanilla Plone does have a “Dashboard” screen. I like the notion of a personal dashboard where you can decide which portlets to show (or have them pre-configured for you). This, in one fell swoop solves the problem of “Fire Hosing”… it opens up the possibility of adding Google Gadgets (say for email or calendars)…

So in a nutshell… I need to…

• Attempt creating a new Archetype based object type (this may  be useful as a Minutes object for recording meetings).
• Add Location abilities (collective geo looks good but I can’t contact them)… this just needs to be very simple
• RSS Feedmixer and portlet
• Anz.Jabber for instant messaging (even if just for status … i.e “is online”)
• I tried at the WebServices product but get an error when I call it. This might be handy for hacking and integration (not essential).
• integrate Faculty / Staff Directory (does it do Twitter accounts and Blog URLs? If not needs adding)

And finally I need to work up and discuss with Cynapse some UI ideas, particularly with regards to the Home screen.

Why did Cyn.in pip the others at the post? There are lots of reasons…

It’s open source (and built on Plone). The importance of this can’t be stressed enough for me. It buys the project the ability to benefit from other extensions and development work. I’ve already seen an XMPP (instant messaging) plugin that we’ll be able to slot in. It also means, ironically, that we don’t have to make a decision about which tool to use just yet, we can continue trialling Cyn.in for as long as we like, even getting to grips with what it is like to bend and extend. And lastly, open source is a good fit ethically.

It’s not based on a per-person pricing model. Because I am planning to create a collaborative environment where it is easy to invite collaborators in, I’m expecting that people will set up projects that maybe only run a few months - and maybe some people in those projects might be invited in to “cast an eye” over sketchy, unfinished work.

Its user interface, which is based on jQuery is by far the most instantly usable and likeable. This can’t be stressed enough - people actually like it. And for me, seeing how well the wiki interface behaves is delightful being able to link from one page to another without any special codes or URLs is refreshing.

I am now getting to grips with Plone. I may be gone sometime. I’m trying to work out how hard it would be to mine some data from social media sites and research repositories and create a sort of discovery engine for people at the University of York (see below). I’ve done the mining and the visualisation bit, now I just need to figure out how to save objects in the ZODB and how to create Products in Plone.

If anyone knows of any good resources/books for learning Plone, please pass them on… So far, the two books I’ve tried have left me a bit cold with code that doesn’t work very, very early on in the “Getting Started” chapters (sigh!). Wish me luck…

As I was just saying to Richard Millwood, during trialling software with different teams we found that the way the permissions model is communicated (or who can see what) makes a huge difference to the feel of the community. Quite early on in the project I knew that as well as needing the things that people explicitly asked for that tended to “shut people up in corners” we needed spaces and places that pull people together, like public promenades.

And additionally, we needed public-ish places where people can share what they’ve used, what worked and what didn’t… sort of peer to peer training (or a pattern language even). By public here, I don’t mean totally public, I mean public to all the members of the community. More public spaces are sites like LinkedIn, Twitter etc…

This all breaks down, in my head at least as Tools, Toys and Training…

Later I’ll explain what’s in the Toy box… apart from things like the PPPeople browser and  heaps of fun.

Who else uses it - Are they “like us”

• SocialText: Citizen’s Advice Bureau. Boston College (ahem)
• Jive: Intel, Nike, VMWare, SAP are all people we’ve heard of.
• LifeRay: French MoD, Penn State University, York University in the States - York Case Study.pdf , Sesame Street isn’t too bad.
• Elgg: Leeds University, University of Sussex

Cost - If it is a per-user model then what would be the cost “if” half our students sneaked in

• SocialText: Approximately £30,000 a year (for 3,000 users)
• Jive: Roughly £40,000 a year
• LifeRay: Free (open source)
• Elgg: Free (open source)
• Confluence: £1,200

Extensibility options - Can we develop new modules - How far can we get with templates? How hard is creating portlets? How easy/dangerous is hacking PHP?

• SocialText: Gadgets. Restful API. Has template documents. SocialText seems to be PERL ( they’ve kept that quiet! )
• Jive: It looks like the development opportunities for Jive are better than I’d first thought. You can create and deploy plugins as JAR files. There is both a SOAP and RESTful API.
• LifeRay: The model here is that you develop portlets. There is a full API (even for JavaScript jQuery API access).
• Elgg: You create plugins for Elgg or adapt the source code.

Development options - Does the platform support our technologies / abilities?

We currently are more than happy with a TomCat/ Oracle or LAMP platforms. I am familiar with python.

• SocialText: Perl based. Actually open-source, although licensing is a moot issue.
• Jive: Java-based.
• LifeRay: Java-based.
• Elgg: PHP.
• Confluence: Java development.

Community and Documentation - Is it up to date, helpful, copius and easy to use.

• SocialText: There actually seems little in the way of developer relations, meaning this would ultimately be a client / vendor relationship. It then becomes important that the company have a transparent development process and are open to recommendations. I already have a growing list of usability issues/suggestions that I feel are important enough to raise before purchase.
• Jive: The Jive Developer Community use Jive and as such it seems well serviced. I received a reponse to a complaint from a Jive staff member in days.
• LifeRay: Being open source they LifeRay community seems to be vibrant, current and helpful.
• Elgg: Although the community seems willing and able, the discussion forums seem difficult to use, with each person seemingly forced to create a group for their particular question. This doesn’t bode well.

Deployment - Does this fit what we have?

Are there any issues?

• SocialText: Hosted or appliance.
• Jive: Hosted or software. “Not resiliant but reliable“
• LifeRay: Self-hosted TomCat portal.
• Elgg: Hosted or self hosted LAMP stack.

User Interface - Is it nice to use?

• SocialText: Missing a few Web2.0 niceties such as inline validation etc.
• Jive: Exemplary.
• LifeRay: Has a few styling rough edges but handles adding tags etc very well.
• Elgg: Competent and clean, but at time clunky.

Reports and Feedback - do people like it?

Elgg: Rachel from Leeds…

My personal perspective - I don’t like it much. I found it clunky and difficult to customise. I doesn’t display properly in different browsers (this may now have been resolved). I find it difficult to navigate. The best thing I can say is the file store was useful - also the ability to set different levels of access to files. But again, this process was clunky. As you may gather, I’m not really a fan.

LifeRay: InfoWorld has named Liferay Portal the “Best Open Source Portal” on the market.

Jive: Sheffield and their users really like it.

SocialText: Unknown.

Confluence: Lots of positive feedback, particularly from the development community (and Julie in Digital Library).

Integratability - From systems we have (calendar, email, LDAP, XMPP) to those we might have soon(Files/Sharepoint/Alfresco, VLE stuff, HR widgets? Portal etc)

This is one of the most amorphous and yet important factors, in that, we are looking not to duplicate work already underway in different departments. For example, the VLE is soon to be “socialised”, the portal project looks to unify some sort of dashboard mix of information and we have existing enterprise tools like email and Sun Calendar in place and potential tools lurking on the horizon, such as Sharepoint or Alfresco file sharing.

Any system that integrates well with LDAP, and that can be easily gadgetized will be at an advantage. Shibboleth also lurks.

• SocialText: LDAP.
• Jive: LDAP yes.
• LifeRay: LDAP yes. XMPP (I think so)
• Elgg: LDAP plugin currently alpha and untested.
• Confluence: LDAP. XMPP (unsure, plugin available for JIRA)

Change management - can we import from existing tools like MediaWiki / Drupal etc to provide a migration route for existing platforms?

• SocialText: Don’t believe so.
• Jive: Don’t believe so.
• LifeRay: It looks like this is possible in theory. May need some development.
• Elgg: Don’t believe so.
• Confluence: Yes.

Can group admins invite external people easily? This is kind of crucial…

• SocialText: Yes
• Jive: Yes.
• LifeRay: Unknown.
• Elgg: Yes.

Is the privacy and permissions model well communicated? This one “scares the horses” most… not when it’s bad, but when it’s not clear.

• SocialText: Could be improved. Need to ask further questions, for example, in the SocialText admin screen below what does “public” mean, is there a mode that means “logged in users”? Not completely clear… AND … The workspaces themselves do a poor job of communicating their Permission settings.
• Jive: Good.
• LifeRay: Adequate.
• Elgg: Slightly poor. Could be improved with minimal development.
• Confluence: Unknown.

Given the widespread adoption of Google tools, how well does it play with those?

• SocialText: There are gadgets for Gmail and Gcalendar.
• Jive: Doesn’t seem to support gadgets at all.
• LifeRay: Supports Google Gadgets.
• Elgg: Kind of.
• Confluence: Unknown.

Events Calendaring - almost everyone playing (seriously) with SocialText has created an “Events wiki page”… this should be noted.

• SocialText: Poor events handling. Seem to be missing completely instead adopting an “everything is a wiki page” stance.
• Jive: Doesn’t seem to support events, although a task oriented calendar is available within a Project.
• LifeRay: Social office has a pleasant calendar/events tool built in.
• Elgg: Groups have events.
• Confluence: Unknown.

Branding options for individual groups

In my travels, lots of people have expressed a keen interest in how things will look.

• SocialText: Little to none. You can add images to the “Home” page for a project.
• Jive: Little to none
• LifeRay: Unknown. It is possible to fully brand an install. Not sure about group level
• Elgg: Unknown. It is possible to fully brand an install. Not sure about group level
• Confluence: Looks totally skinnable, even for individual spaces