Posts by Pete 2

A better mousetrap

> someone coming in with a completely blank sheet, able to write a clean and efficient system from the ground up,

All very fine, in theory. However a lot of the baggage that banking systems have are the fixes, workarounds, tweaks and efficiency hacks that had to be added to the original clean sheet bank system in order to get it to work securely, according to all the regulations, without stiffing customers' 0.0001 pennies from rounding and able to "talk" to all the other banking systems: which don't quite work in practice the way their paper specifications say they should.

How long would it take to get from a brand new, clean system to one that is able to operate in the same environment as all the mature (albeit rat's-nest coded) systems we have today.

And on another note: how would you persuade customers to deposit all their "hard earned" with you. Customers need to have confidence in their banks and all the annoying little glitches that would be found during (say) the first few years of operation wouldn't add to the customers' feeling that their cash was safe. And when you add to that the sinking feeling that a customer would get when receiving their introductory letter:

Dear Mr. Worstall,

Please tell all your friends about us

Doesn't really work, does it?

Sieze the opportunity

> "We're not perfect but we have got a lot better than we [ Virgin Media ] were,"

Not too surprising, since they have do so much scope for improvement.

Personally the only complaint I have about Sky broadband is the godawful help desk. However, once you can get past or short circuit that, their actual techies have been both helpful and reliable.

Units?

> Success can only be measured in time

Now that's impossible! (dot com)

I wonder how much success she's had with this venture our money? Would it be a couple of months, or a millisecond?

Although investing her own money is a bit of a win-win since any losses just reduce the amount of tax she (or more likely: some company that her accountant has magicked into existence as a container for this venture) has to pay, whereas any profit would go directly back to her.

Swings and roundabouts

It's an old cartoon, older than the internet, but still sums up the world of IT just as well as when I first saw this 30 <ahem> years ago.

http://www.businessballs.com/treeswing.htm

Apropos the question the governor or the governed

The answer is: both. IT has a parasitic relationship with "the business". Neither side can exist without the other, which is probably why both sides hold each other in such disdain.

Honesty is the best policy

One of the inter-generational differences (whether it's the metaphorical cart the or horse in this argument I don't have the insight to say) is attitudes towards honesty: telling the truth - whatever "truth" turns out to be.

One of the many throwaway lines in the BBC sitcom Just Good Friends (other than the innoculation classic: "didn't you feel a prick?" "Well, Penn, I was a bit embarassed") was Penny's admonishing Vince for telling lies and getting the reply "That's because I'm a liar" and after that the whole topic was dropped and the story moved on. Vince tells lies: it's no big deal.

These days, it seems, that "being a liar" doesn't mean telling people that it's "black" when really it's "blue". But that lying now means not volunteering every single piece of relevant, or otherwise, information at the earliest opportunity - what used to be called "blabbing". Worse than that: being caught lying (telling actual black is blue untruths) is portrayed as being such a heinous crime that the very thought of it is, well, unthinkable. So much for "ooops, you got me" and brushing it aside.

In such an environment, where information is offered freely to or even: forced on a person, how can the generation brought up thinking that way act any differently? They have it drummed into them that giving information is mandatory. That withholding it is "sinful". That you must tell the truth all the time, to everyone, no matter what the consequences for anyone involved.

So when a big, bad, internet giant says "tell us everything", they blab. Age, height, weight, most intimate thoughts, address, list of 50 closest friends, political leanings, shoe size, suspicions about classmates, inside leg measurement, which musician they'd most like to have sex with, hair colour, medical history, favourite programmes and fantasies about their teachers. Hell, even PV-ing is less intrusive.

Whether this is considered "being honest", or is simple naivety, or some sort of catharsis we'll never know. The good thing is that all this data gets lost in the mix along with the billions of other personal records. Fortunately for all these data-givers, so little of it is either interesting or relevant.

A quadcopter by any other name

Flying a quadcopter in Texas?? Isn't that known colloquially as a "target"

Intelligence on the internet?

The ability to imitate a 13 y/o boy is a good goal - if you're a 12 y/o boy.

However, I feel that if Alan Turing was alive today, and looked at the traffic on the world's most popular social media sites, he would (quite naturally) assume that they were test-beds for AI's and that there was still a long, long way to go before any of them appeared even faintly human.

If this result tells us anything, it's that a test devised right at the dawn of the IT era, before there was any experience of AI to draw on, is too limited to be useful. Just as we don't believe that aircraft imitate birds (even though they both fly), we shouldn't consider this anything like a computer imitating a person.

The Magic Roundabout

> manufacturers to launch a new range of connected Freeview HD televisions and boxes which consumers will be able to buy in store

Gottit! A gizmo to make TV watchers buy yet another gadget. Given the failure of 3D (which never actually was 3D, just a trick of perspective) and curved screens - surely the dumbest idea so far this century (really: tellies that only show an undistorted picture if you happen to be the one person sitting in the sweet spot AND which gathers all the reflections from everywhere in the room and beams them directly into your eyes?). Given those failures brilliant marketing strategies, the TV industry desperately needs to get us all buying more crap.

What better than telling us that the smart TVs we've bought over the past 5 years are all now obsolete,as they won't have the software to receive this new service (and after the first and probably only software upgrade, won't be able to receive it in the future - after someone, somewhere changes something and we all have to whip out our wallets again).

The only question that the once bitten, twice shy consumers should be asking is "what guarantees are there that this platform, with it's multiple content providers, won't implode under the weight of it's own infighting in a few years time and we all have to shell out again, just to get back to watching all the TV that wasn't good enough to watch live when it was first broadcast?"

Personally, I give it up until just before the 2018 world cup. What a coincidence that would be?

Pro or con?

> sentences for attacks on computer systems fully reflect the damage they cause

How about requiring some of the responsibility for insecure and badly designed systems to fall on the heads of the people who design or (mis)manage them? One of the properties of being a "professional" person is taking responsibility for the quality of your work (on the basis that it is specialised and therefore beyond the grasp of "ordinary folk" to understand or critique) and that you have a duty of care towards those who place their trust in your services.

Maybe if people who's designs were so flawed - or who cut out every failsafe, oversight, procedural check or protection mechanism for reasons of cost or stupidity - that they amounted to criminal negligence were held to account there would be at least a start to getting some half-decent, secure software. What we have at the moment is like blaming the woodworms when your house falls down.

Britain's got secrets

> classified 3 levels above Top Secret

Looks like the good old Top Secret isn't quite so "top" any more.

A case of inflation or just using the "onion" strategy of having layer upon layer?

It does make you wonder, though. At what level are things really kept secret, so that only those who need to know, actually *are* the only ones who do know. And what is it that they seriously don't want us (or The Guardian) to know about.

We should be told!

Going off the rails

> All versions of Mint until 2016 will use exactly the same base as this version

Which means that everything will work fine ... until you change something. Just like with all Linux distros.

It also presumes that everything you could ever want is available through one of the "approved" repositories. And that the software version available there is fully up to date with both bug fixes and features.

If only that were true.

While most people who's daily usage is limited to surfing, playing videos, knocking out the occasional document and possibly spreading some sheet - then everything will be fine. But what happens once you stray off the path? If you "need" an application (say a wizzy new webcam app) that isn't on the list of blessed software - or only has a version several major releases behind. What then. You find a repository, add that to the list, download your new application and all it's dependencies.

And there's the problem ... You're no longer on an LTS release. Your dependencies could include many newer library versions, updates of other, dependent, programs and possibly even security fixes that your LTS doesn't recognise or support. if you dare to buy new hardware during the 2 years of LTS-ness, there' every likelyhood that you'll have to look above and beyond the official suppositories to get the support, drivers and libraries you need to get your new toy to play.

2 years is a long time in the Linux world, and things change quickly. What are the chances you'll want a change, too?

Linux on the postage stamp

> The 25 x 25mm embedded system

Electric motors revolutionarised (groan!) the world. They're in most powered appliances these days - but they're "invisible": just one of the components that we don't give a second thought to.

This little thingy, and its future generations could do the same.

While there are embedded microcontrollers that are just as ubiquitous as electric motors, they tend to be single operation, uncommunicative, non-networked, un(re)programmable devices that are indistinguishable from all the other chips, and are there simply to replace a handful of logic chips and keep costs lower.

However, these devices will be exposed to the outside world. They have networking capabilities and therefore can be accessed by anyone with a Wifi transceiver connected to a keyboard (whether you want them to connect, or not). Not only is that their strongest feature, it's also their biggest drawback. Apart from the hacking vulnerabilities, a dependency on a network link that can disappear, be reconfigured or suffer RFI is just another thing to go wrong. So while these could turn the IoT into reality (once the Mk 2 gets its power consumption down) I really hope that whoever is making these devices, or embedding them in their products, is gearing up to provide an extremely highly staffed help-desk.

Maybe that's where the real money is to be made?

What goes around ...

> This column is a repeat publication

We know that TV "stops" (well, stops broadcasting anything worthwhile) during the summer ... and for the month or two over christmas ... and at weekends and various random times throughout the year when the american broadcasters go for a lie down and screen sports instead of entertainment - and we have to stop too, in case we accidentally see something good first.

So in the spirit of repeats, let's have a few commentards repeat some tardy comments.

Here's a possibly prescient offering from October 2007: (filed under Olympic ticketing system crashes under demand)

Speed kills

> automation can also weed out errors that may creep in

Automation can also apply the same cocked-up commands to all your databases in the blink of an eye. At least with manual operations, you know something's wrong after you've destroyed the first production instance. It's then possible (maybe even desirable?) to stop and investigate before breaking all the others.

While I'm a great fan of automation (as previously discussed here), I do realise that it's a "force multiplier" rather than unconditionally good. It can help you do good or bad things much faster. However it's also prone to problems when someone changes something that knocks your automatoin off it's rails (maybe something as simple as changing a password - always hard to automate *and* keep secure) or "fixing" a misspelled table name.

However automation can free up ones time to allow for more goofing around on IT forums. Sometimes that even gets mistaken for work.

Real or imaginary

> provide valid photo ID, such as a copy of their passport or driving licence

Strictly speaking, they are asking for an ID. Or something that looks like it's an ID. Or rather: a scan or photo of something that looks like an ID which may or may not have been doctored.

Whether the name or photo on the electronic copy of the (real or fake) document is actually the person (or their agent) who is asking for this take-down is not something that Google has any way of verifying. Unless they are planning to try mugshot matching against their collection of individuals in photos that someone may have correctly tagged.

One assumes that this is merely a deterrent, rather than any sort of checkable auditable "proof". And that if someone does decide to be a little bit naughty and asks for someone else's (maybe <shock!> not even an EU citizen's) URLs to be removed, then there's nothing really that can be done. Except for Google to email the falsely removed victim and ask "is this you?" Which then just starts the cycle of verfication and validation all over again. Without any absolute proof of anyone's actual identify ever being independently validated.

But it does give Google a nice little stash of supposedly government issued ID documents (gee, I hope they are kept safe!) submitted by people with something to "hide" and a pointer to what it is they want to keep quiet. It would be a real shame if that collection of IDs and hush-ups ever got out. Is this really privacy, or is it actually making things worse?

Nice but dim

The photos are good, but the biggest problem I have with my phone is the total inability to see the preview screen when outdoors and therefore to frame the shot.

Sure, you can stand there with the phone in one hand and use the other as a sun-shield, hoping you can press the photo button with your thumb whilst neither shaking the camera nor dropping it. Better would be a low tech solution: simply for the camera makers to drill a hole right through the phone's case: front to back that a photographer could peer through, just like "old fashioned" viewfinders. However the ultimate, which I'm still waiting for, would be a screen that combined the non-reflectivity of a black hole with the eyeball-searing brilliance of british sunshine (yeah, I'm kidding) so that you could actually see the image you were about the commit to posterity before you took it - and if the screen then provided DSLR-like data regarding what the sensor was seeing, being able to see that would be nice, too.

So while we still have screens that are too dim and too shiny to see in daylight, I can't see the need to upgrade. However many megapixels a new camera (that makes phone calls, too) can boast about.

Power corrupts

... and absolute power corrupts absolutely.

Let's face it: in the markets they operate, Amazon reigns supreme. Not merely dominant but in total control of the who, what, where, when and how much. So for a company in that position to say

"a legitimate negotiating tactic"

when the victim of it's negotiating is so powerless brings to mind all the "tactics" used by Standard Oil to suppress competition and maximise profits, at the expense of everyone it dealt with, 100 years ago.

Yesterdays problem

If you look at such esteemed techniques as ITIL you will see that they provide for a whole host of Capacity Planning processes. As it is with all methodologies that come from government sources (and are recommended by external consultancies), ITIL assumes that your organisation has an infinite amount of IT resources at its disposal. All of whom are flexible, helpful, knowledgeable, do what they're told, always have time for the odd meeting or 10 and accept the findings of others with respect and instant compliance.

However, capacity planning is based on the notion that computing kit is expensive compared to a person's time and that it makes sense to have teams of people poring over RMF reports and trying to squeeze that last MIPS out of the mainframe.

These days, nothing in IT (leaving aside the downtime on a major national retailer or bank - which is usually down to software cockups or hardware failures) is as expensive as the people who run it. Nothing at all. Hence the days when CP could "earn its pay" by forecasting growth, performance, response times, limitations, bottlenecks and when to add that next Gigabyte (yes, 1000 MB) of storage are long gone. At yer average cross-charge rate, slapping in another 1 TB disk is about as expensive as your senior team leader taking an hour for lunch. So when the question arises: should we add more memory, or more disk space, or more processors? then the cheapest solution is to just answer "yes" and get on with it.

Even if you don't need the capacity right that minute, you will do: soon enough. And SSDs have now removed (provided your systems architect has more than John Innes number 3 between the ears) the I-O limitations that used to plague database performance, even a few years ago. Though the inefficiencies of software production will soon use up all the benefits that SSDs, gigabit networks and 256GB of RAM can provide to servers.

Gets no argument from me

I did once (once!) find myself on Buzzfeed's website - not in a Google searchy sort of way, I just navigated there through ignorance (or was it cat-killing curiosity?). Yik! I felt the need to wash my hands, disinfect my computer and trash my monitor: as they must all have been contaminated by it's awfulness. A feeling I haven't had since I accidentally clicked on a Daily Mail link.

However the question of "who is to blame" is harder to answer. In the first instance, a lot of blame must fall on the editorial staff and their willingness to commission and publish the stuff they do. But, looking deeper, they wouldn't be on that Road to Hell if there weren't people willing to read it, or at least get caught by the clickbait-counters that determine the advertising revenue that will accrue.

One would hope that these things are merely an aberration of a still-experimental internet, and inexperienced users who haven't yet worked out what it is that they want from the web. Whether that will turn out to be the case, or if this is the sort of stuff that's here to stay and will grow (like Japanese Knotweed) is difficult to say. Luckily it is easily avoided and here at least: once bitten, twice shy.