“I felt positively embarrassed to get my gear bag out when a flock of trolleys was nearby”
After a stint as an in-patient at a London hospital, Steve has a few radical ideas about the use of on-trolley PCs
Ihave recently discovered that I am not a hospital person. Partly this was caused by visiting a sick relative, but mostly it was brought on by a few days being fixed up and then cared for on a ward crammed with patients, nurses, doctors, families and hangers-on, porters, catering staff – the food was pretty good, incidentally – and the occasional scribbler in suffering. Actually, it can’t have been so bad, because I was utterly hypnotised by what the nurses were using as an IT system.
Each nurse had a trolley – nothing new there, you might say. Except that these trolleys were topped off by a 24in widescreen monitor and adorned with a massive steel shelf, just at arm height, where the keyboard, mouse and sundry paper printouts were to be found. Oh, and a little barcode reader, to pick up the data on the wristbands of their patients.
I felt positively embarrassed to get my gear bag out when a flock of trolleys was somewhere nearby; for a bit of in-patient chic I had brought my Toshiba WT-20c
Windows 10 tablet. This is an unusually broad and cleverly designed separable, with a gorgeous hi-res screen and a little hidden pen nestled in the base. I couldn’t find the accompanying keyboard, which was a shame, but on the other hand, as a naked tablet, I paid £55 a unit.
And it works. Looking at the nurses with their Daleklike trolley, I was forced to wonder, what would be so terrible about making use of a smaller, more personal device and display? Sure, there was a need to connect some peripherals like the barcode reader, but the payoff in improved mobility and battery power (most of the trolley’s cargo seemed to be regular 240V mains powered, so the only way these things were truly mobile was with some kind of on-board battery to drive the big screen) was surely worth it.
My shame at losing out in the “size matters” stakes was rapidly dissipated by the first ward round, when the senior doctor toured the ward with a gaggle of students, and an outer ring of trolley-pulling nurses. They actually had proper trolley jams whenever more than three nurses were clumped together to hear the senior man’s words. It seemed absurd to me that the ward was almost crammed full of display screens already, on separate blood-pressuresampling devices or as pay-per-view TVs hanging over every bed. With a few minutes of hackery, I could drive such devices as secondary screens from my humble Tosh, via Miracast services in Windows 10 and 11.
With a little more work, I could have a Tosh-sized device on a shoulder strap, almost unobtrusively sitting a few inches below the armpit of the user.
Was I the only person there, I thought, who remembers the days of “wearables”? To me, the information on the 24in screens the carers were toting could be divided between VR glasses or goggles for the practitioner, and the aforementioned Miracastdriven general-purpose screens up on the walls. I’m not proposing a new use for all those face-covering helmet style VR systems: I’m harking back to some of the less intrusive efforts, which look like a regular pair of specs with the lenses either left out entirely, or present as a thin translucent film.
Bedside manner depends quite a lot on eye contact, so dressing up your nursing staff as robot zombies is probably worse than having them accompanied by a seven-foot tall Dalek drug dispensary.
The real meaning of “business continuity”
A lot has already been written about the global service outage on a lot of cloud services at the end of July. It’s unfortunate that the finger has been pointed at Microsoft when the confessions from management and workers at CloudStrike have been frank and public. No mysteries here, no cabal of shadowy hackers to blame: by far the biggest interruption to service came from a fairly simple mistake in operational support. The basic problem was how convoluted and difficult it proved to get the various affected machines and apps back in a working state.
I suffered directly from this outage: it took out the NHS systems that stitch together GP prescription issuers and sundry prescription-handling pharmacies. This caused my GP’s practice pharmacist to hand-write (oh, the horror!) a prescription, have it personally signed by the duty GP, and then call my partner to come and physically collect it. We made jokes about “back to the 1970s” until my partner turned up at the pharmacist’s to find that the young man on duty had probably not dealt with noninternet delivered prescriptions before. Incredibly, he treated the lump of green paper with ballpoint pen scribbled on it as a suspicious article, and eventually handed over only one of the four items on it.
This was not, by anybody’s definition “business continuity”.
Not for the supplier, not for the doctor, and certainly not for me. We have become used to the idea that BC and DR (aka disaster recovery) are matters strictly for the IT team, so that whatever the tech issue it will be solved and service restored as soon as possible.
Since that definition crystallised, the business has permeated the technology just as much as the technology has permeated the business. Eventually, if you follow the siren’s call, you and your customers will be drawn into an invisible web of dependencies, designed by app and service programmers whose world does not include the possibility of catastrophe.
The CrowdStrike story contains some considerable complexities, and a few parts where the censor’s big black pen will descend, lest we give away too many secrets to hackers. And Jon ( see p110) and Davey ( see p118) have covered it in some detail in their columns. So I’m going to offer you a simpler case study.
It concerns an old client, whose accounting team had a separate developer on contract, and a remit to enable the company website to take orders that fed back into the accounting database, from which invoices were made. What these guys wanted was an open incoming access, from the web servers to the accounts system, so that any time a customer clicked ‘Buy’, in would come the triggers to produce and post them an invoice. Simples, they cried. Over our dead bodies, we replied.
Here’s the problem, we explained. The systems you’re passing through to make this connection all have their own configurations, which demand the occasional reboot. No mechanism was included to pass back a “stop” order to the web servers; transactions that were generated during a firewall or accounting server reboot were simply lost in the ether. No retry mechanism, no queues: just keep that link open 24/7 and we’ll be fine, old boy. Except when we came to the inevitable moment when a router reboot was mandated; then their blind-sent transactions would be blasted to an address that simply isn’t listening.
We came up with a different way of doing the job, which allowed us to do reboots of the kit we were looking after without losing track of the web-generated transaction queue. We did this by shifting the active party from the external web server pushing, to the internal accounting server, pulling. That overcame everyone’s objections, at the occasional cost of a couple of minutes’ delay from order to invoice.
Don’t take that simple example as a literal translation of the CrowdStrike/ Microsoft problem. It simply shows just how far apart the worlds of cloud developers and businesspeople can be, even at the simple stage of automating a process previously done with printed spreadsheets and yellow highlighters as the accounts team worked through that day’s orders. From a disaster recovery perspective, the accounts people were confident they could get the highlighter pens back out at the first hint of trouble, though they didn’t have a plan for the kind of widespread chaos across many contributing systems that arose from CrowdStrike’s operational error: there’s not much point scrabbling for the marker pens if your customers can’t raise a purchase in the first place.
I’m very tempted by the idea that the term “professional” should
“It shows how far apart the worlds of cloud developers and businesspeople can be”
be revised, to include the statement “someone who knows how to do the job without use of a computer”, and that’s not just me having a fit of pique thinking about my chemist and the refused prescription. As cloud services remove skills from our companies and normalise everyone’s working practices, we must be able to make innovative decisions while times are both hard, and changing fast – decisions that by their very nature are unexpected and new.
Preparing for eventualities whose descriptions and effects seem guarded by an opaque curtain of nerd-speak is a 21st century challenge for the modern businessman. For the IT professional like me, it’s a case of back to the old school, because it would be a very unusual contingency plan that didn’t include a few extra or upgraded machines in the office, synchronised in terms of the records they contain, but not in terms of the software used to access those records.
If your software guru is spluttering in outrage at such a revolutionary suggestion, I commend them to take a look at the database-sync abilities of FileMaker Pro, or the end-user toolkits around MySQL and lots of others. You might think that various SaaS services (and that includes Microsoft 365) take an approach to customer data that’s essentially no more reassuring than a pat on the head.
Video goes pro for amateurs
I hate social media advertising. Once, in the very early days, I was offered 25% off Puma attack helicopters in a social media advert. For me this set a benchmark in credibility (and incredibility), which helped for a while because it made me simply ignore any social media adverts, not even devoting mental resources to the “click/don’t click” decision.
This worked fine until I was in a bed on an open ward, hooked up by various tubes, very bored and dependent on my phone for entertainment. Up flashed an advert from Amazon, showcasing something that looked a bit like an over-fed iPad. I followed it and realised that what I was seeing was a live streaming, multi-camera “mixing desk”.
I followed the advert (I did say I was bored) and two things became apparent. One, I was still befuddled by the surgeon’s anaesthetics. Befuddled enough that I didn’t bookmark the page to present it to you here. Two: someone, somewhere in China, thinks that the demand for multicamera live streaming and recording devices is so widespread that they can charge over £1,000 for a little bit of hardware that does this job.
Streaming live video is a contentious topic. There are objections of a commercial nature and even political factors: live streamed video is associated with copyright infringement (filming the away match whose TV rights have already been sold) and even fringe political activities (fanatical followers ignore mainstream news media and live solely on a visual feast of streamed, real-time coverage by self-admitted partisan producers).
There’s also the vexed question of interactivity. One of these multicamera, tablet-format devices can present five distinct video feeds, leaving the viewer to decide which one gets the big window and which are relegated to the thumbnail rail. This doesn’t sound like a problem, until you sit down and have a chat with someone who’s worked in old-school media.
One of my hospital visitors is one such: a video media professional, very interested in the rules of production. She’s firmly in the club that says just because modern tech can receive five simultaneous feeds, there’s no guarantee that the audience can repeat the benefit of that technology.
Her research came from a slightly different time, when all the camera firms produced 360-degree panoramic cameras, leaving the audience to pan, tilt and zoom the data feed on the receiving viewer, rather than leaving those moves to the decisions of the film-maker. It turns out, she said, that people far prefer to be told a story, rather than having to find it for themselves.
That’s not the absolute death-knell for a five-camera pocket edit desk that costs you a bag of sand; even a basic face-to-face interview setup or Zoom call can gain in credibility with a variety of shots. Think long-shot panoramic views, close-ups on faces and mid-range shots that reveal things such as hand gestures, hair-flicking, nail-chewing and so forth.
A few years ago we would have thought that a specialised, dedicated tablet device doing this job was an unnecessary thing, a toy for the hobbyist. Suddenly, streaming has come of age, in ways we could not see coming even five years ago.
“Once, I was offered 25% off Puma attack helicopters in a social media advert”