It transpired that Google App Engine is remarkably easy to use.  On Mac OS X I just needed to install the App Engine Launcher, configure it to use TextMate as an editor, hit ‘Edit’, and get to work.  You pick an unused app name, and it gets deployed to [your_name].appspot.com for free.  Oh, how do you deploy?  Hit ‘Deploy’.

So it’s perfect for setting up a little application to do something very small (although, if your situation demands, you can also run the Django web framework on there, or indeed anything that’s either in Python or in a language that can be compiled into Java bytecode).  With just a handful of lines of YAML you can set up cron tasks (i.e. run scripts every, say, 9am on a weekday, or whathaveyou), which work by making an HTTP request to a given place at the allotted time.  If you’re just making a little Twitter bot, it couldn’t be much simpler.

So the idea was to create a little mash-up of the @big_ben_clock Twitter account (which just goes ‘BONG BONG’ etc every hour, like any good bell in a tower would – except on New Year’s Eve when it went ‘BANG! BANG!’) with the idea expressed with eloquent laconicity on Do Websites Need To Look Exactly The Same On Every Browser.  As Andy Clarke put it recently, “The real question is not should web sites look the same in every browser but can they? The answer is no. Live with that, move on.”  The unchallenged assumption, borne of the practice of web designers getting design sign-off from clients on the basis of static visuals alone, is something that, in 2010, is one of the biggest factors holding back quality web production.  It needs to be a truism that there is no expectation that interfaces should be the same everywhere – and rather, ultimately, that interfaces are already or will soon be actively expected by users to mould themselves to the context in which they appear.  So this Big Ben is a web standards advocate, who wears a blue beanie, in reference to the cover of Jeffrey Zeldman’s ‘Designing with Web Standards‘.

So, @beanie_big_ben was born:

You get the idea.

The only mild issue faced getting this up was the lack of available syntax for defining when cron scripts can run.  You can run them every hour, but there’s no way of determining which minute of that hour they’ll run.  So I went the route of running them every minute, and having the script just return early without making an API call if it isn’t within a minute on the hour.  Sounds inefficient, but we’re only talking the dinkiest little script.

This also unearthed the probable reason why @big_ben_clock sometimes neglects to BONG in particular hours – it probably works on Google App Engine too.  If you set a cron script to execute every minute on the App Engine, it’ll actually execute about every 62 seconds, which means that about every 30 hours it won’t execute within the minute after the hour, but just miss either side of it.  Tant pis, or even bof, as the French may say.

Beanie Big Ben also can do requests, if you have a burning web issue for him to pronounce upon.  (Either leave one in the comments or tweet him up.) Twelve times “IT DEPENDS” can just fit under the 140 characters – he’s not afraid to fence-sit!  Let him know.

The article could be described as “breath-takingly” lazy if, as readers of the UK press, we hadn’t become so inured to such distortion.  So here’s the criticism:

1. “Police leaders” are intimated, yet only one (Calum Steele, the general secretary of the Scottish Police Federation) is directly quoted;
2. The MyPolice website doesn’t exist yet, so how can it be pre-judged?
3. They are firmly against the site becoming a place for vendettas against individual officers – i.e. the intentions are not “shop-a-cop”;
4. Calum Steele’s words, apparently a “warning”, are not in fact unreasonable.  One would expect that anyone in his position would be properly questioning about the efficacy of such an initiative, but so would the MyPolice team be themselves.  The project is difficult, but it can provide benefits to people, so that’s OK;
5. The only substantive assertion in the piece, being Mr Steele’s one that the police service already have (sufficient, one would infer) ways for the public to express dissatisfaction, is not challenged, or offered up for debate.

It could be futile to complain about the slant the titling of this piece gives to the rest of the content.  The attempt at ostensible balance in reporting, without the journalistic organ being honest about their own voice (do they support initiatives like MyPolice?  what part of their remit are they serving by giving it coverage?), is surely a refuge of the lazy.  Mr Steele was clearly called up for a quote about something he couldn’t possibly have been adequately briefed about.  Clearly journalists are under pressure to conjure up a sense of controversy in their pieces even if it does mean distortion.  We often complain that news organisations never report about good, positive things, and it’s unclear whether this happens because we have so little trust in public institutions that we can’t agree who represents that which is socially good and positive, or just because we love lazily allow our attention to meander to the aura of controversy.

I suppose established newspapers are having some difficulty justifying themselves to themselves and their customers, what with the onslaught of the new-form read/write Web.  It’s curious, however, how this can happen when the newspapers already have one up over the interhorde – they still have the ability to make people, lots of people, read about stuff they wouldn’t otherwise come across, in the public sphere.  Time was, I could write a letter to a newspaper, and one of my grandmas, even, would notice and give her opinion (an opinion that was contrary to mine, because I was writing to the Daily Mail just to rile ‘em).  Nowadays, rare is the medium where that might happen, and unfortunately I’m all out of grandmothers.

However, they squander this advantage with brainless one-dimensional have-your-say sections, and forget that the job of journalism is sometimes to be didactic, or to be icily objective, or to stay above the celebrity fray, or to root out new stories that will help me make sense of the world we live in, and won’t just divide us up and make us ignorant.  The UK is about to let into government a party that doesn’t have substantive policies, that is just as bad as the incumbents only more so.  Public life is sick if people aren’t seriously questioning how this can be.

So the MyPolice project has all these tensions it needs to work through and resolve.  There’s the issue of how, exactly, they do go about making sure that negative comments don’t become ex judice crusades against particular named officers.  They need to make sure that all participants are paid back in kind by their involvement – a complainant needs to know that they will be considered seriously, and the police need to know that any effort they expend in engaging with people through the site is worthwhile in terms of making themselves understood.  An interested visitor needs to be able find out something interesting about their area so that they can remember that the site can be of value to them should they ever need it.

MyPolice seem to be going about this the correct way – through interviews, finding out the situations in which such an online service may be useful to people, and working out what positive outcomes might come from involvement in this online forum.

I might have used such a service in the past.  In the summer of 2005, a few weeks after the G8 summit at Gleneagles (at which time, you may remember, a large number of people in Scotland were arrested and held arbitrarily under terrorism-related charges), a friend of mine was arrested in central Glasgow by a couple of officers on a power trip.  I witnessed the whole thing, and afterwards looked into complaining about the conduct of the officers in question as it had fallen far short of the kind of professional behaviour one would, with hope, expect from the police.  The procedure was, essentially, that you would write to the officers’ superior, and they would write you a letter providing an explanation of their behaviour.  If you weren’t satisfied with this response, you could take it even higher.  Never, though, was there recourse to any kind of independent body (Scotland doesn’t have a body like England’s IPCC) – the police would always have the opportunity to protect their own.  The procedure seemed to be written from a position, essentially, of arrogance: one was to infer that, for any given action by a set of police officers, there would be a reasonable explanation that could be presented, which you could either accept (in which case you went away quietly) or reject (in which case you have to go through the rigmarole of re-reporting the incident to a higher-up with no guarantee of a reasonable outcome that wasn’t a tokenistic fob-off).

I didn’t even start with it.  I would have engaged with MyPolice, had such a thing existed.

What’s less clear, though, is what I might have got from it.  The central thing it might have provided would be the site’s central service – to provide a place where a complaint can be made, in public, about particular incidents that the police can then respond to, again in public.  It may also have been the case that other people had had similar experiences with the police, and so I might have been given more information, through the prism of the experience of other people, about what kind of recourse I had available to me.

Here’s what I wanted, though: I wanted an outcome.  I wanted the officers in question to be punished for misconduct.  I didn’t want their names mentioned on a public forum, but I did want them, as individuals, to be reprimanded appropriately for their actions in arbitrarily arresting my friend.  If possible, I wanted the option of a suit for wrongful arrest.  I didn’t want the kind of camaraderie one might find on a consumer website, where one might find consolation that everyone else is just as screwed over by a particular private utility company as you are.  I believe in a good police service that doesn’t tolerate two officers arresting a young woman because they can, and that doesn’t tolerate their buddies down at the station lying to me (and her) about the whole thing.  I’d hope that a site like MyPolice could make some worthwhile progress in convincing me that if this were to happen again, I’d have access to a complaints mechanism that wasn’t in place just to fob me off and make me go away.  That’s fine (well… it’s not fine, but that’s a story for another day) for a gas company or a bank – I can just switch companies – but I have no appetite for condoning private vigilante armies!

The success of the MyPolice project will be in the details, and in the re-thinking of existing models for online participation to make something that really works for the finite set of use cases where the service could really make a difference to people’s lives.  I wish them luck.  The Sunday Times should, too!

There’s also an excellent suggestion for a piece of blog functionality that can help orient the reader of a comment thread that may have become unruly by dint of sheer length – with the deliciously idiomatic name of a “mile-marker”, this would be a kind of meta-comment that would sum up the comments so far with the intention of making it far easier for the would-be commenter to jump in.  The mile-marker would either be written by the author themselves, in which case it could include some implicit commentary (i.e. picking out an individual comment in the mile-marker would intimate appreciation), or crowdsourced out (though it was unclear whether this would work!).

This would all probably work pretty well on Mr Santa Maria’s blog (and I would encourage him to try it out – seems to me it sits well with his other idiosycratic touches such as artistic direction for individual posts) – he’s a well-known web designer who gets many well-thought-out comments per post.  Online commentary, though, exists in many places, in many fora, and it’s unruly almost everywhere.  Before we can really come to any answers that may be expected to work to corral them (and I think devices like mile-markers will have their part), it’s important to cast our eyes over why this might be.

Consider a newspaper article – say, on the Guardian’s Comment Is Free section – where the body of the article is written by a staff journalist, yet the focus of the article would be deemed to be the comments (in this case using a platform provided by Pluck).  The journalist may or may not join in with the comments themselves – indeed, the piece may be a syndicated spiel by Noam Chomsky, say, a fellow who couldn’t be counted on to deign to give feedback to comments wherever his original piece may wind up.  Even if the journalists comment, it’s rare that they are speaking as authors, and the presentation of their comments doesn’t privilege them as such – much of the time they are sucked into chumming with the commenting contingent, backpedalling from the original article content (which they have no freedom to edit anyway).

There’s a tension here, which is that the commenting system on a site like Comment Is Free exists as a rootless, amorphous mess of pseudonymous commenters ready to latch onto any article that is published and bend it to its will, not because the commenters necessarily have anything to say, or are interested in anything so much as a conversation, but because the subtext of a site like Comment Is Free is that any article that falls in the forest without a comment makes no sound.  It’s there on the Guardian’s front page – an article’s importance is not to be determined by the content of the article itself, but by the Geiger counter read-out of the comment radiation that surrounds it.

There is the question: what is the motivation of a commenter on such a public webspace, where commentary is almost entirely pseudonymous, and where Pluck’s system (probably by omission) discourages any actual conversation?

There is the central tension: when someone writes a comment against any piece of web content with clear authorship, is the commenter speaking to the author directly, or merely performing a kind of existential venting?  Or is it the case that comments will all fall somewhere on that scale?

There is another tension: when someone writes a comment, are they writing that comment with sole reference to the original piece of content, or are they sensitive to the context of there being intervening comments?  Or is there another scale, and where on the scale is the correct/conscientious place for the commenter to be?  (Here’s why the mile-marker idea shines: it is both unreasonable to expect a commenter to read a preceding set of 384 comments, and reasonable for a blog author who is interesting in fostering something that be called a conversation to expect a commenter to at least be up to speed on where the conversation is at the point in time they come to comment as the understanding of the topic in hand will hopefully have moved on from that implicit in the original article – so the mile-marker allows the commenter to quickly go some way to reconciling these two states so as not to embarrass themself when they write the comment.)

Another tension: between whom is the conversation?  Is it between the blog author and the combined body of the readership, so more or less a Q&A?  Is it between a group of people who already self-identify as a group, with their own established shared discourse, their own means of naming themselves, etc?  If so, where does the author fit in – hapless supervisor?  Babysitter?  Careless troll-feeder?  What does this mean for the nature of the conversation?

Another possible source of tension: is the author of the piece of web content the controller of the entire forum for commentary, and are they clearly able to exercise this control?  Or is the editorial control for the commentary the responsibility of someone else (as in the case of a newspaper article)?

Why is commentary allowed?  Is it a pretence at democracy, or is the author actually interested in things commenters might say, as opposed to the weight of commenters who may say it?  (Democracy ain’t qualitative.)  Is the author really interested in furthering a discussion that is greater than themselves?

How do commenters identify themselves?  To what degree are they projecting their own egos in a given situation?  (This could, possibly, be measured by determining whether a commenter is saying anything factually new or ideologically different to anything else that had theretofore been expressed in a particular comment thread.)

How do comments end – some kind of realisation by someone that makes all other comments moot?  Or do they just fizzle out?  (Don’t they pretty much always just fizzle out?)

What about an article that’s a year old but still relevant – what does it mean when a commenter writes a comment then?  Does the consideration that the author probably won’t see any comment against such an old article mean anything?  What about the future, when it’ll be common to come upon blog posts that are fifty years old yet still with commentary open?  When does the notion of interactivity die?

It’s clear that in order to rein comments in, this stuff cannot be left open to chance.

Re: the mile-marker idea, I had thought of the concept of a comment chorus, whereby comments didn’t comprise consequential posts after the article, but instead comprised a wiki that subsequent commenters could edit (with, naturally, an undo admin function!), the sum-total of which would be a kind of meta-article that would show “what we’ve learnt”.  Commenters’ names would be listed, but would be separated from their individual contributions to the chorus.  Consensus, if one could be reached, would be left like a residue for all to see, and see quickly.

I think the matter of identity in online commentary is actually the elephant in the room of this discussion, and I’ll write about that in a subsequent post.

So, comments? (Or comments about comments?)

Molly Holzschlag let it be known a couple of weeks ago, on Twitter, that best practice involves not using underscores because of “irregular implementations”, and also mentions in neighbouring tweets that underscores were originally illegal in CSS2 (and had to be escaped in CSS1), but that they were made legal in the errata.

The only authoritative article I can find on the subject, and the one I think is being referenced here, was written by Eric Meyer in March 2001. It mentions that Netscape Navigator 4.x simply didn’t read underscores in class names, and neither did Opera 3.x-5.x. IE4.x upwards, more by luck than judgement, have recognised them fine.

So, unless somebody can stop me, I’m going to use them. Case, if not closed, then ajar.

There was one occasion where I went to HMV to buy a CD. Standing outside the doors, I could see aisle upon aisle, rack upon rack – every CD known to mankind is there for me to buy. Over the decades my taste in music has broadened; I like all sorts of things. There was just this knowledge that whatever I buy, I’ll get it home, I’ll put it on and I’ll be disappointed. Later that night, I got home, got the children to bed, started doing my emails. I’d read about Napster, but I hadn’t actually used it. I started imagining that every piece of recorded music is out there – I could click on my computer and have it in my hand. I knew that whatever I got on to the hard drive, when I listened to it, it wouldn’t be what I wanted. From that point, I started seeing recorded music as a kind of genre in itself. Somehow, all recorded music, no matter where it is in the world, even if it’s from the pre-recorded music era, once it had been sucked into being recorded music and you’re listening to a two-dimensional thing, that can be listened to anywhere any time while you’re doing almost anything. In my head it’s all become this one thing that’s fast draining of meaning. Once an artform loses its meaning, it no longer has any real value.

Noticing a theme emerging?

OK, the “watch for mistrakes!” is a tiny bit too unartfully kooky, but that’s forgiveable.  What really impresses me is the help text given for the password field:

We hate bizarre password rules. Keep it over 3 characters, but otherwise feel free to make it as secure or insecure as you want.

This is great!  So many misconceptions fly about around the nature of a “strong” password.  Web application developers take it upon themselves to “educate” the unwashed masses of the benefits of picking a hard-to-guess password, and increasingly force you to pick “at least one number” or “at least one non-alphanumeric character”.  So people think that pa$$w0rd is a significantly stronger password than password, just because Facebook tells them so.  What’s missed is that once passwords get to a certain level of complexity, they become impossible to remember.  So the user is forced to reset their password every time they use the service, or note it down somewhere, or have it emailed to them.  It opens up further points of insecurity.  So jam isn’t going to keep even a basic brute-force attack especially busy, but jammertime might! Plus, you can remember it. Plus, it’s up to you, and you as a user are not forced to abrogate your own responsibility to keep your account access secure by adhering to someone else’s nannying over-simplifying AJAX widget.

So I was pleased RescueTime wears its opinions on its sleeves, and doesn’t take me for an idiot. Because, well, we share enemies, and I suppose that could be the start of a beautiful relationship!

Oh, and while they’re at it, another type of thread that is due a markdown:

Newbie:

I’ve got this problem. I can’t figure it out! I’ve tried this and this and that, and no worky!

“RTFM” nerd:

Listen carefully. Type ‘ifgumbo -err –include-suppositories’ into the command line and post what you get. You know what a command line is, right?

Newbie:

*screeds and screeds of stdout garbage you have to page down ten times to get through*

…

*tumbleweed* *it’s at this point you realise the thread is from 2003*

I have an idea about alternative culture: if there’s something interesting going on, leave it alone. Don’t whack it on YouTube, don’t blog about it, don’t hire a PR and get a piece in the Guardian Guide. And never invite a marketing executive to enjoy the ‘experience’. Just go there, have fun, go home, shut up. Mass communication is all very well, but stumbling across culture is important too. Special is good.

I suspect more and more people are going to have this impulse, and a growing band of people are going to struggle against being automatically mediated at every turn.

So the specific gripe is when a designer actually goes to the effort of capturing text rendered on the page and creates a reflection from that, ignoring the fact different OSs and browsers have a different array of fonts, and that text is, well, text, and shouldn’t be treated as a graphic.

Life then gets far worse when you change the text but neglect to re-generate the graphic. I suppose that would be a hassle, right? I missed the boat on getting a screenshot of this, but in the earlier part of this year, the Microsoft home page was making exactly this mistake in its footer: the text itself said “2008 Microsoft”, but the reflected text put the year as 2006. They seem, now, to have noticed (after a period of at least 12 months!), and have obfuscated the reflected text so you can barely make out the lettering. Good, trailblazing move, Microsoft!

So it’s a disappointment, considering that jQuery is my most favourite of JavaScript libraries, to find this same error on the jQuery UI download page:

Especially as there’s a neat fade-to-orange effect there, which needs this image:

It’s a shame such a neat effect is spoilt by a lack of attention to detail.

Re-generating an image every time you want to change your text is a hassle you can save yourself from quite easily – just don’t reflect the text in the first place!

Now, I look forward to the day when these types of reflections can be rendered by CSS. How cool would it be if you could select the text and see the selected text in the reflection? Ersatz approximations of this, however, look ugly, are unmaintainable, and should be nixed.