Friday, 28 August 2015

Numbers, reputation, and worth

In the past week I've been shedding twitter followers, while at the same time my Klout score (Klout is a website that claims to measure the 'worth' or 'impact' of your tweets) has increased by a few points.

It's long been my view that metrics, rankings and the rest don't mean much individually  in absolute terms but that in aggregate, higher scores indicate a degree of worth.

And this is sort of demonstrated by this week's little event. Normally if the number of followers had gone down you would expect that my Klout score would go down as what was being said was seen to be less valuable.

On the other hand if what was being said was felt to be more valuable my Klout score would go up and probably the number of followers would increase - and certainly this has seemed to be the case in past months.

But of course twitter is populated by a host of inactive accounts, perhaps related to dead and stalled projects, and of course it's the end of August, the time when in the northern hemisphere, academic projects are typically wrapped up and closed down.

So I'm guessing what I've been shedding is a slew of low worth accounts.

And this is a learning experience - what's being said is more valuable than the numbers listening, ie measuring influence/impact is more complicated than the things we can easily count ...

Monday, 24 August 2015

The disruptive chimera of the digital humanities ...


Over the past year I've become more and more convinced that Digital Humanities is a chimera, much as Eresearch is also a chimera.

Many disciplines in the physical sciences have always deal with large data seta and their manipulation. Many researchers in the social and health sciences have always carried out complex analyses of government statistical data to reveal both new trends and the impact of legislative changes.

Until recently the poster child for this was psychology - ore more accurately the cluster of closely related behavioural sciences from ethology through to neurobiology that are usually lumped together as 'psychology'.

Psychologists have used computers since they became widely available to contol experiments, present stimuli and illusions, and analyse data. Clever innovative work that has become increasingly more innovative as technology has become cheaper and more and more off the shelf components have bcome together.

But nothing was more than a logical extension of previous research. And that is what digital humanities are - a logical extension of preceding research. Yes, the easy access to large quantities of data and the availability of easy-to-use mapping systems, natural language toolkits, has allowed a step change in the nature of the research, but not a fundamental change.

For example, in a moment of rash enthusiasm, I thought you could do something with the tax return date in the Domesday book to graph the harrying of the north - after all the Domesday book is semi structured data and online as queryable resource - until someone pointed out that someone had looked at exactly that question some thirty years before, with some rather more traditional techniques.

In other words there was nothing special about applying digital techniques, they merely amplified what was already possible, and by extension there is nothing special about the digital humanities.

And because there is nothing special they need no special consideration in the provision of computing resources, they merely require consideration.

Where their disruptive effect comes from, and the thing that makes them look like something new and different is the scale of the step change - the large scale digitisation of resources though projects such as Google Books, and the comparitive cheapness of cloud based computing has meant that a guy with a laptop and a good idea can make a significant difference for a low cost, and unlike in the science the data collection cost is negligible.

However, even this difference will disappear as both various initiatives for the digitisation of legacy data in the sciences and the open science movement and its emphasis on data publication bear fruit. Just as in the humanities somone with a laptop and a good idea should be capable of disruptive change.

And in both these cases there is nothing special in the resources required. The real disruptor is that the person with the laptop and the good idea need no longer be at one of the big research centric institutions - meaning that research can spread outwards to smaller, and perhaps more nimble institions ...

Tuesday, 4 August 2015

No more playing with linux on old Imacs ...

Saturday was a sad day.

We bundled up all my old PPC iMacs and took them to recycling - I'd finally come to the realisation that I was never seriously play with PPC linux again. That's not to say I'm stopping playing with linux, because the one thing that the whole PPC linux on imacs thing taught me was that older hardward can have its working life usefully extended by being sidegraded to a lower demand linux environment.

Great for cash strapped schools and libraries, but it's not a panacea.

It does require that the linux distribution that you choose continues to be supported and maintained, simply because you need modern browsers, as well as all these security patches. It also means that the hardware you're working on has to have some upgrade potential - extra memory, more internal storage, simply because even the best linux implementation or application is not immune from bloat.

So, what now?

I'm personally convinced that linux on the desktop remains a viable alternative to proprietary operating systems. In fact things like printing and network configuration have got a lot easier over the years, and the general robustness of Libre Office makes it a viable alternative office suite, just as you can run your life on Evolution, just as you can on Outlook.

Using a linux based notebook I've found nothing I can't do on a machine with a commercial operating system with the exception of working with documents created with odd templates and change tracking - something that's not quite as portable as it should be. In fact things like straight forward text editing are a lot more straight forward on linux. I will say though that you do need a decent browser, purely because increasingly one ends up using browser based applications (Evernote for example)  rather than stand alone applications - however, Firefox admirably rises to the occasion.

So I'll keep on playing with linux - except it'll be i386 hardware from now on ...

Monday, 27 July 2015

Big phones and smartwatches ...


I got myself a new phone this weekend to replace my 3 year old Galaxy S2 – an S5.

My old S2 was basically running out of puff – it would occasionally crash or flatten its battery and even with an extra SD card to boost the space for data and applications it was always a little tight for updates, making it time to upgrade, especially as my mobile provider had a special deal on the S5.

Obviously it had been a popular offer – for when my phone arrived, instead of being Virgin branded it was Optus branded with the Optus extras. Virgin resells Optus bandwidth in Australia, and I'm guessing that Virgin had run out and Optus still had stock, and someone failed to reflash my phone.

Anyway, I have a new phone. Battery life is definitely better but it's also inconveniently large – too big to fit in a jacket or jeans pocket, meaning it'll have to live in my bag with my notebook and other gubbins. For the first time in a long time I've ordered a case for a phone, living in a bag it has more chance of getting scratched and banged. If a tablet or laptop needs protection in that sort of environment I'm going to guess a phone does as well.

The other thing is the realisation that the bluetooth based smartwatch concept makes some kind of sense. If your phone is stored somewhere difficult to get to – like your work bag, being able to do the Dick Tracy thing and answer your phone from your wrist brings back the convenience of a mobile phone, rather than desperately hunting for it when it rings. The same goes for email and text alerts, and of course you may simply not hear it because it's buried among a pile of other stuff.

However I'm not yet convinced that's $200 worth of convenience ...

Friday, 24 July 2015

The paperless office ...


I grew up in a world of paper.

There were no word processors, only typewriters. Memos were written by hand. Mail meant writing a letter putting it in an envelope, sticking a stamp on it and dropping it in a mailbox. Social media meant talking about a newspaper report over a drink with friends.

Later on there were things like troff and eventually LaTeX, but there wasn't anything like a proper word processor until the advent of WordStar. (For my sins I actually used to teach WordStar and can still remember the macro commands).

Even though eventually we all got access to wordprocessors and email storage was expensive – always the luddite I always used to just bump up students' filestore if they asked – so stuff tended to be printed out and filed just as it would have been in the nineteenth century.

Same with meeting paperwork, expenditure reports, and all the gubbins of system management and solution delivery.

And because I'm a creature of habit I ended up with a 3 drawer filing cabinet in my office full of paper that no one ever looked at.

Well we're moving to a new open plan office next week. All the documents in that filing cabinet exist on my computer, on the various project sites and wikis, or archived using evernote or onenote.

So I took the three drawers of paper, dumped the non-confidential stuff in the paper recycling bin and shredded the rest.

I reckon I can find most things if required. Yes my online indexing might not be the most systematic, but it's no worse than searching through drawers full of stuff.

It's just possible I've finally achieved the paperless office ...

Thursday, 23 July 2015

Eduroam and public wi-fi networks ...


I've previously sung the praises of Eduroam, and it remains by default networking solution when visiting other institutions, but yesterday I had an experience which made me question whether Eduroam is the only solution.

I was at a meeting at the University of Canberra and I'd taken my Xubuntu netbook as a writing device. When I got there, I discovered that I'd forgotten to configure Eduroam on it. Major fail on my part.

So I pulled out my old 7 inch note taking tablet, only to discover that while it was fully charged and connected to eduroam, its certificate was out of date, meaning it wouldn't authenticate (it could of course just be that I'd stuffed up the eduroam configuration – but the middle of a conference on e-research is not the place for network debugging).

And this of course highlights one of the problems with eduroam – the configuration is tricky, and non standard – it's not like most public wifi systems where you get a private ip address, and then sign into the network, provide some identity data and tick the box agreeing to abide by the conditions of use and not do anything involving naughty Nora and her oscillating hamsters.

Setting up to use eduroam involves installing a certificate on your machine and configuring some settings. Not difficult, but fiddly and outside most non-geeks' experience.

The other problem with eduroam is that it assumes that you have a university internet account and can authenticate appropriately. Not all visitors to campus do, such as visitors from government research organisations, commercial bodies, and overseas academic institutions, particularly those in SE Asia.

Like all universities in Australia, UC have an eduroam service. But they also have a new experimental service called UC-visitor, where, you guessed it, you sign in just as you would to a public network in an airport, on a train, or in a shopping centre. I'm assuming that they do some rate limiting to prevent abuse and track usage to avoid people using it as a substitute for their 3G connections.

In use, the service was perfectly adequate for email, tweeting and syncing a file to dropbox, which basically is all you want to do, ie write stuff, show people stuff, tell people about stuff.

Eduroam is a service that has its roots in the days when internet access and particularly high speed internet access was expensive and therefore rationed. We're not living in that world any more.

In Croatia and Slovenia, even Sri Lanka, internet is everywhere – in Croatia,  coffee shops and petrol stations offer it for free, without any need for authentication, and even in one case, a small coastal town (Drvenik to be precise) provided free connectivity on its beach strip. Interestingly, the University of York has recently brought the York city public wi-fi network onto campus, while also extending eduroam coverage to the city network.

In a world where free public internet is increasingly becoming the norm, does Eduroam require a reboot?

Tuesday, 14 July 2015

Cloudprint for linux ...


Last year, I upgraded my old EEEpc701SD to Crunchbang linux to make a distraction free writing machine – something that's worked out pretty well, especially since I started using Focuswriter as a basic writing application.

In the meantime my venerable HP PSC1200 inkjet upped and died – the scanner still works and it still sort of prints but quality is variable and sometimes print is intermittent – I suspect that the contacts between the cartridges and print mechanism are dirty or damaged. However, without going into the ins and outs our local big box office equipment supplier had an end of financial year special on Epson wifi workgroup inkjets so I bought one to replace the PSC1200.

Apart from being discounted, one of the inducements for going Epson was that it supported Google cloud print natively, allowing easy printing from Chromebooks and tablets, which is something that's becoming increasingly important to us.

At the same time as setting up cloudprint, I of course added the drivers to our windows machines as well as our increasingly venerable imac to allow them to print to it as if it were a normal local network printer.

 Adding the drivers to my Linux netbook to achieve to achieve the same turned out to be a bit more complex than I thought it would – Epson don't distribute a PPD file as such, you need to install their print management utility, something that meant spending some time with
dpkg -i and apt get -f install.

After that little detour, installing the Epson drivers on the Eee seemed to be asking a little to much against the minimalist spirit of what I was trying to achieve here.

So I went googling to see if anyone had written a linux Google Cloudprint client.

And they have

One of the advantages is that once installed,  all your Google Cloudprint printers, become available meaning that you can save stuff as a pdf to Google drive, which is quite neat as an alternative to emailing stuff to Dropbox which is what I've been doing up to now.

Installation onto Crunchbang wasn't quite as easy as it should be but the following script works (for me anyway, your mileage may vary):

wget https://niftyrepo.niftiestsoftware.com/cups-cloud-print/packages/cupscloudprint_20140814.2-1_all.deb
dpkg -i cupscloudprint_20140814.2-1_all.deb
apt-get -f install
/usr/share/cloudprint-cups/setupcloudprint.py

Obviously you need to run it with sudo rather than straight from the command line.

You also need to have a web browser installed on your machine. The install script will prompt you for the google account name you want to use and generate a magic url you need to paste into the browser url bar.

Google will prompt you to login and then generate a keycode you need then to copy and paste back into your terminal window to complete the authentication key.

And it works. For a writing machine it of course also means that you can work on something on the bus, and using the wifi, queue something for printing and proof reading at home ...