I found myself thinking about gimmick movies last week, after watching Source Code, and being struck by a dog-that-didn't-bark realisation. There just aren't that many of them. What I had in mind were movies which messed around with the narrative conventions of commercial film, as Source Code does by repeatedly looping back through the same eight minutes in time. I wasn't thinking about avant-garde film, in which pretty much anything goes, and I wasn't really thinking about art film (or video) either, which usually accepts that people won't be watching from beginning to end. I had in mind the kind of films that you go to the cinema to see, and expect to unroll in a conventional, sequential form. And it seemed interesting to me that the experiments were so few and far between.

Take real-time or single-take filming, for example, the latest example of which hits our screens this week in the shape of the Uruguayan horror movie The Silent House, which – notionally at least – unfolds in one unbroken shot. In one respect a real-time film is the most obvious narrative experiment of them all, one that almost seems embedded into the technology of film – even if it took technology decades to catch up with the idea. The very earliest Lumière films were effectively all real-time productions, even if the technology of the day didn't allow them to be very long ones – and there's a sense in every subsequent real-time film that some kind of purifying process is being undertaken. It's a discipline, after all, not an indulgence. On a spectrum that runs from natural to artificial, or from the raw to the cooked, the single-take film is felt to sit much closer to some notion of primal innocence.

It's a sophisticated illusion, of course. When Hitchcock filmed Patrick Hamilton's stage-play Rope, attempting to persuade cinema-goers that they were watching a single unbroken shot, he had to resort to all sorts of tricks to get round the limitations of current film technology (concealing an edit in a foreground pan across a character's back, for example) and the set was a masterpiece of contrived landscape. Mike Figgis's Timecode – as far as I'm aware the next mainstream real-time movie to be made (52 years later) – had it slightly easier, in that his running time exactly matched the capacity of the digital cameras he was using. But the fact that he used four of them simultaneously and had to orchestrate interlocking scenes in real-world locations more than made up for that. "Four cameras. One take. No edits. Real time", read one of the poster taglines, a promise of simplicity that didn't mention 15 consecutive days of identical shoots.

I haven't seen Aleksandr Sokurov's Russian Ark, a 2002 film which charts a labyrinthine course through Russian history and the interior of the Winter Palace in St Petersburg but which also made an implicit promise of distillation in one of its poster tag lines: "2000 cast members, 3 orchestras, 33 rooms, 300 years, ALL IN ONE TAKE", it read. And now we have The Silent House, which sets out to terrify in real time; no moments of reprieve here, one imagines, when the sun comes up for 15 minutes and you can take a breather from fright before it all starts up again.

You can see that the frequency graph on real-time movies has turned sharply upwards; virtually none in the first 90 years of the medium and then in the last 10 years a relative boom. Given what the technology allows I don't doubt that there will be more experiments, but I think audiences and even directors are likely to come to the same conclusion that Hitchcock did back in 1948 and decide that it is "an experiment that didn't work out". Why? Well, possibly because of this paradox – that in eschewing editing, the camera and its movements become unignorable. It becomes a character on screen, never directly alluded to but always there, in between you and what you're watching. The edits in a conventional movie, you realise, aren't just a convenience when it comes to telling an extended story. They are also the place in which the camera can most successfully be hidden away so that we forget it altogether. One of cinema's innate instincts is to pretend that it doesn't exist at all, which is why films that insist on the fact are so rare.

Mona Lisa mystery still raising a wry smile

Silvano Vinceti is at it again. In December last year, I wrote about the Italian "art researcher" (frequently described as the chairman of Italy's National Committee for Cultural Heritage, which sounds extremely grand and official but actually isn't): "I won't risk a prediction about exactly how far we will penetrate into 2011 before his next sensation," I wrote, "but I bet it won't be long." A matter of months, in turns out, since this week Vinceti announced that he's planning to dig up an Italian noblewoman called Lisa Gherardini and reconstruct her facial features "to put an end to a centuries-old dispute". In the past, Vinceti has also proposed digging up Leonardo himself, to check on the theory that the 'Mona Lisa' was actually a disguised self-portrait, and he recently claimed that Leonardo painted a secret cipher into the eyes of the 'Mona Lisa', one which was only detectable under high magnification (rather raising the question of how Leonardo got it there in the first place). Any theory will do, it seems, as long is it attracts attention. And since the 'Mona Lisa' is the Paris Hilton of visual art, Mr Vinceti nearly always gets it. I can't help wondering what he will he do with himself if he actually does put an end to the 500-year debate about the sitter's identity? Then again since he also recently hypothesised that Caravaggio was poisoned by lead in the paint that he used (another cause for exhumation) it seems unlikely that he'd just go quiet.

James Joyce, licence to thrill

It's intriguing to see that James Jones' heirs have given permission for a revised edition of his 1951 novel From Here to Eternity, which will include references to homosexuality struck out of the original edition as well as language that the original publisher, Scribner's, felt was unacceptable at the time. Jones resented the censorship but had to cave in to prevailing attitudes. The first thing I thought was that someone might re-issue Norman Mailer's The Naked and the Dead without its irritating adenoidal concession to contemporary niceties, which soon becomes fuggin' irritating for a modern reader (though since Mailer retained the fugs for a 50th-anniversary edition, it's possible that permission wouldn't be forthcoming). The second thought was to wonder what would happen to the history of 20th-century literature if contemporary liberties could be backdated. If James Joyce wrote Ulysses now, and sexual and linguistic licence was all that had changed, would it be significantly different to the novel he did publish? Would Lawrence have had another crack at Sons and Lovers and Women in Love, de-euphemising their more passionate scenes? More to the point, what would happen to Molly Bloom's great torrent of thought if Joyce could have come straight out and said what she was up to? The losses of retrospective liberty might be as great as the gains.

t.sutcliffe@independent.co.uk