On passwords

I have been getting exasperated with a non-work website which has been asking me frequently (more often than once a month) to change my password. The most likely consequence of being asked to do this is to reduce security by writing down the password. (This does seem to be a bug, though, as the site owners claim it is only necessary to reset the password once every three months.)

More entertainingly, on taking over a new piece of kit I had to set a Bitlocker password. The validity rules were:

  • 9 letters long
  • Consonants and vowels as follows: CVCCVCCVC
  • No repeated letters

The person equipping me had some suggestions about combinations of words that would satisfy the requirement, but I thought up one of my own, feeling rather like a Countdown contestant juggling vowels and consonants. On getting home, I wrote a short program to extract all single-word valid Bitlocker passwords from my online dictionary; there are 14 of them. As well as some common, easy to remember ones such as ‘fisherman’, ‘gunpowder’, ‘lethargic’, ‘nostalgic’ and ‘wonderful’, there was one word I didn’t know: mockernut. Of course, none of these should be used as an actual Bitlocker password as they must be among the first things hackers will try.

Mental Health Awareness Week and workplace kindness

This is Mental Health Awareness Week, and in line with an area of concern in recent times, the theme this year is kindness. In a workplace context this includes: avoiding mean-spiritedness; knowing you have a right to be listened to if something is bothering you; taking care with those whose mental health might be more fragile, for example at times of maternity leave or bereavement. Wise colleagues (and an organisation is fortunate if they are line managers or admin staff with ‘human resources’ responsibility) have never needed to be reminded of these things.

I wrote in an earlier post about how some of these skills were in short supply. Reading about this week’s initiative and the trend it represents, I have one question: why did it take so long for the penny to drop about this?

(Disclaimer: nothing in either of these posts relates to any issue in my current employment.)

the pony in the field

A few years ago Private Eye made a quip which clearly assumed that only the well-to-do could afford to keep a horse or pony. In the following issue a correspondent took them to task for this, saying that second-hand tack and the goodwill of a local farmer with a spare corner of a field made this possible for people who would not count themselves as wealthy.

The present-day model of a digital humanities project can be a little like that: a much smaller affair put together in odd moments with additional labour from cheap sources such as graduate students and living on a corner of a server in space that can be spared (the equivalent of that pony in the field). I have remarked on this before.

It wasn’t always like this. Recently I read an old conference presentation from IASSIST, which alluded to projects I was involved with or whose content I used back in the years around the turn of the millennium. (I have fond memories of an IASSIST conference I attended in Edinburgh.) It was an era of ambitious, generously funded projects linking many institutions. I think for example of Intute’s attempt to catalogue resources in the whole range of subject disciplines.

Even before the credit crunch the tide was turning against projects on this scale. Intute was killed off because no institution was willing to take it over; perhaps it would have been better if it had remained as several separate services, some of which might at least have found homes. Rather like the robotic scanner which was too costly for any one institution to house. I recall also the insistence that another project I worked on was ‘unsustainable’ and the data would have to be ‘repurposed’ (though it seemed to be serving its intended purpose well enough).

So many potentially valuable resources lost because there was no ‘corner of a field’ to put them in and a managerialism that will not allow a small amount of time to be dedicated to maintaining them. (I’m not really thinking of Intute now, which would have increasingly suffered from incompleteness and link rot without significant effort.) My work in research data management has made me aware of the long-term value of preserving digital material, but all that is necessary for it to be lost is for it to fall into the hands of those who don’t understand that value. And like that pony, it need not be very expensive to keep.

One website I developed, Hidden Lives Revealed, happily made the transition to a new host (migrating it was simple and quick), and is running smoothly to this day because all the software I equipped it with still works. It seemed like a large site when I created it (we stored the files on a supercomputer at Manchester University) but as is the way with these things the amount of filespace it occupies would now seem insignificant. As its upkeep is funded by a charity I cannot imagine it has been expensive to maintain. It seems to be a matter of chance whether a resource survives, quietly cropping the grass, in this way.

communication with candidates

Almost every job I apply for requires ‘attention to detail’ and ‘communication skills’, often with an interest in the ‘user experience’. I feel these are often lacking in the recruitment process itself.

I helped to create a tutorial for support staff at Bristol University, now known as ‘InfoSafe’, which dealt with data curation and storage. Amongst other things, we explained how you could use file and directory names constructively for easy retrieval. Some employers could do with studying this advice. I’ve just downloaded further particulars of a job I’m interested in, and the file name is of the form ‘123456’ or simply ‘Job Description’. Would it be so hard to include the job title – in abbreviated form if necessary? Maybe the organisation’s name too, given that the file is likely to be downloaded by people outside it?

Another regrettable trend, which I put down to tidy-minded HR departments, is an absolutely formulaic invitation to interview. So much so that in two recent cases, instructions about a presentation to be given at interview were contained only in an attachment, with no reference to the presentation or the attachment in the body of the message. So easy to overlook, especially when many emails now come with a string of otiose attached files. It’s basic email etiquette that if there is important information in an attachment, the body of the message should indicate this.

No wonder the skills I referred to in my opening sentence are in such demand!

novel-length source code

I have a bookmarked collection of ‘horrors’: Web pages which really should have been designed differently, many of which are mercifully no longer online. Backgrounds were one bête noire; I shudder when I recall the page promoting Scotland on a repeated backdrop of saltires, for example. A more recent example is the Ure Museum’s display of images of its entire collection (scroll down!) behind the search box. For the full effect, however, one needed to include animation. Spinning geometric patterns on a page about an exhibition of Islamic art (at the Bodleian, no less) or reduplicated Flash clips as can be seen here. The prize was taken by the guest-house in Wales which displayed six simultaneously rotating 360-degree views of each of its rooms, one above the other. It was impossible to look at this without feeling genuinely queasy.

There are of course many ways in which Web pages may be unsatisfactory. Investigating a local antique shop (or rather, the premises it used to occupy) I came across the page at https://www.antique-glawibbless.co.uk/. (To lessen potential embarrassment, I haven’t given it correctly; remove ‘wibble’). It appears to be a very simple page thanking past customers, and while one might think the mix of seriffed and sans-serif fonts a bit last millennium, uncovering the source code is the true revelation. It is 4,963 lines and 427,273 characters long (a mere 401,361 without spaces), that is, about as long as a typical novel. All to display a couple of images (one of which has had the top cropped off) and a message that is 31 words long!

How can this happen? I used the Wayback machine to look at earlier versions of the page, and while they had links to what was then a more extensive site, there seems no justification for the baroque confection of Javascript and stylesheet that lies behind it. The single largest component is a long list of instructions relating to a long string of fonts, including Cyrillic, Devanagari, Vietnamese and other scripts, linking to Monotype Imaging’s software. Wix.com seems to be the immediate source of the code. I think that pages like this survive because they are commissioned by people who don’t look ‘under the bonnet’ and ask awkward questions about why so much code is needed.

The search engine and Us

To my horror I realise that it is nearly a year since I wrote an article for this blog. I have been busy blogging elsewhere, and it is not for want of topics. I’ll do a quick post – a cautionary tale about search-engine optimisation and branding.

My church supports a charity which for many years was known as USPG – standing for the United Society for the Propagation of the Gospel. Its origins lie in missionary activity in the days of the British Empire, but now its work emphasises partnership with churches overseas (our particular area of support is a project in Malawi).

A few years ago the Society decided to rebrand itself as the ‘United Society’ or simply ‘Us’. Getting rid of the awkward colonialist connotations of the full name, and also sounding inclusive and pally. But you can see the problem – a search on ‘Us’ will find instances of the personal pronoun, or pages about the United States, if indeed ‘us’ is not a stopword and excluded altogether from searches.

(Actually even leaving out search engines, there’s a problem with ‘Us’ in conversation. Saying “St Filofax’ is supporting Us this month” is ambiguous. It reminds me of the workplace I had which named its servers after parts of the body (don’t ask….). It was very hard dictating IP addresses or URLs involving the one called ‘colon’.)

So back to USPG it went. Nice idea, but really any rebranding needs to take account of the search engine test.

where did the casual work go?

One major change since I was on the job market at the end of the last millennium is the abolition of the way into employment in universities by doing casual work. This was in fact the way ILRT recruited a lot of its staff, and a lot of the best ones: someone was taken on for a few weeks, and if they proved to do the job well their employment was extended, in many cases being converted into a permanent job.

I contrast this with some posts recently advertised near here. First time round four candidates (disclaimer: I was one) were shortlisted for the two identical roles and went through a selection process, but no appointment resulted. The posts were re-advertised but again no one was appointed. They were advertised a third time, with vaguer indications of the salary, and the process seems to have been again unsuccessful.

In the past, this would have been much less likely to happen. Without a strict ‘point-scoring’ level that candidates had to reach, there would have been flexibility to offer the posts to any of the short-listed candidates, relying on the probation period in the job as a safety net if one proved inadequate in practice. The work was done as part of a closely-knit team, so there would be limited damage they could do.

Offering short-term contract work and seeing how people got on would have been another possibility, with the possibility of extending it; this is quite common in the commercial sector. Instead, an expensive recruitment process has been repeated over and over, and the backlog of work the appointees would have done has been piling up. It does provide employment, but in the Human Resources department!

what happened to subject gateways?

An ‘ILRT alumni and friends’ group has sprung up on Facebook, with several dozen members. We reminisce about events we attended and projects we worked on, and share photos of former staff jollies and awaydays, conference visits, giveaway items and so on. Many of us had something to do with the resource discovery site Intute (I still have an Intute plastic drinks coaster). It covered all subjects and subsumed ILRT’s own SOSIG (Social Services Information Gateway). As well as cross-searching, Intute also had add-ons such as newsfeeds and a personalised space for tagging and exporting content. It ran from 2006 until 2011, with content frozen till 2014, when the site was finally closed.

(Earlier when I worked at NISS – now Eduserv which has been in the news this week – we had a ‘Directory of Networked Resources’, which also covered all subjects. As well as cataloguing all resources on the NISS site, anyone could catalogue and submit a resource. It ceased to work in about 2002, I believe because there wasn’t the staff time available to check the resources that had been submitted, and by then it was only duplicating other gateways, including those that became Intute. BUBL met a similar fate.)

What went wrong with Intute, apart from the general shortage of funds after the credit crunch? We were proud of SOSIG, but it was apparent when I worked on merging it with other subject gateways, that they were at different levels of development and detail. The claimed advantage of Intute was that it enabled cross-searching, but it wasn’t so clear how much of an advantage that was. If you are a historian, do you need to be able to find resources relating to chemistry? And how much do you need a facility to export content, when most people are comfortable with cut and paste, and have their own way of storing information that they need to keep?

More generally, we don’t seem to hear much about subject gateways any more. A search on the term suggests that the main people who run ‘subject gateways’ now are professional organisations. They were labour-intensive – SOSIG had a distributed team of subject experts cataloguing resources, plus a team at ILRT keeping the whole thing going. And the whole process of resource discovery has changed. Better search engine algorithms have lessened the need for them, as relevant, popular resources rise to the top of the list of search results, so there is less need to go browsing through a catalogue.

But they haven’t gone away altogether. For a few years now I’ve been an editor of the Digital Classicist Wiki, which collects and catalogues ‘digital projects and tools of relevance to classicists’ (this last term is quite generously interpreted). The wiki works by voluntary labour from interested people, and is part of the Digital Classicist hub, which in turn is hosted by KCL’s Department of Digital Humanities. We meet for a sprint once a month, with a suggested list of resources to be added or edited, and contribute at other times. The interface is not flashy, and the tagging by category doesn’t catch everything (this is one of the areas I like to work on); the categories themselves are a folksonomy* of what people find useful, not a systematic hierarchical classification scheme. But it’s got over 500 entries in it now. Because of its focus on digital projects, it doesn’t attempt to link to every Classics-related website – for example, it doesn’t cover Classics departments or most journals – but it does include the sort of resources which were once found on Intute.

I don’t know whether similar sites exist for other subjects. But it would seem that in the case of Classics at least, we have reverted to the earlier model of subject gateway – created by volunteer experts and with basic functionality – as being more sustainable than the relatively short life of sites such as Intute.

(* I don’t use ‘folksonomy’ perjoratively – traditional classification schemes such as UDC or DDC would struggle with some of the categories which are useful for digital resources in classics. A topic for another post!)

The Premdictor

My sons and I decided that we wanted to run a Premier League prediction competition this year. We decided on a system where you got 2 points for putting a team in the right place in the table, and 1 point if your prediction differed by one place in either direction. The advantage of doing this with the Premier League is that, at least until the latter stages of the FA Cup, the teams play rounds of matches, at the end of which every team has played the same number of games. (We are all Reading fans and a postponed match has caused our team to have a spuriously low position in the table for several weeks, now remedied.)

A parent with coding skills can put something like this online in a short time. I have written a script that grabs the current table off a Web page and parses out the team names in their current order, together with the number of matches played. Sorting this out probably took most of my time. This order is then compared against four lots of predictions (my husband has joined in, with a random order of teams). The current table, number of matches (so one can see whether the current round is complete) and all the predictions are displayed as columns in a table. Teams in exactly the right place are in a box with a red background; those one place off are in a pink one. Below the table the current score for each predicted table is given, with a breakdown of which predictions have contributed to it.

After each round of matches I note the scores and winner of that round in a spreadsheet. I could have automated that process too, but I decided not to put in all the extra effort. Perhaps at a later stage. A pattern is beginning to emerge. The random predictions tend to score about 3. The non-random ones have gradually improved as teams have gravitated to their natural place in the league, except that the most accurate prediction of all so far was mine after the first round of matches: 13. And predictions can fluctuate a lot between rounds. After round 14 I was the winner with an accuracy of 8; but one round later I scored only 4.

The non-random predictions have some agreement among themselves. We all thought Everton would be 7th; this seemed most unlikely for a while, but they’ve now risen to 10th. We all put West Ham two places below them, which is way off at the moment. And Leicester in 11th which is rather better as things stand.

Online forums: blessing or curse?

A recent* thread on an online forum leads me to think about the dangers of such forums – or are they blessings in disguise? A communication from a Cathedral to a visiting choir implied a significant change in what such choirs would be permitted to do. An online discussion of the implications of its content caused consternation at that Cathedral. It turned out that this communication had left out a couple of relevant points, and the original poster ended the discussion by posting a correction.

I understand how organisations must fear social media and forums with their power to spread information, possibly altering it in the process. But there are benefits for them. Firstly, an awareness that the content of communications may quickly become public may cause greater care in the wording of those communications and on how they are delivered; in my experience Cathedrals can be quite casual about this. It may mean fewer messages coming from ‘on high’ with the originator of a decision getting someone else to deliver the actual message, with the attendant risk of distortion, misunderstanding and lack of transparency. Better communications all round can only be a help.

Secondly, the internet is making the sort of discussions members of visiting choirs have always had among themselves (there is a large overlap in membership) more visible. In other words, it is now possible to eavesdrop on these discussions, scotch false rumours, correct inaccuracies, and even reflect on the actions which sparked off those discussions. These points seem relevant to a wider world outside the rather restricted one of Cathedral visiting choirs.

* I have backdated this thread so as not to reveal any identities, which is not the point of this post.