Monthly Archives: November 2013

Useless counting

I was just walking out of Security at London Heathrow and passed a small kiosk with the question “What was your experience at Security today?” and four buttons with various smileys. No travelers pressed any buttons, but passing security staff pressed the “I’m very happy” button.

security_happy

I’m guessing that security staff are measured on the happiness of their “customers”. And any traveler with experience of London Heathrow security would find every occurrence of “very happy” spurious.

If you are taking actions or basing remuneration based on  information you collect remotely, are you sure what you are collecting?

Smarter than humans

Google are beginning to achieve results from their work on image recognition with neural networks. In effect, the computers have learned to recognize a few objects better than humans can. And interestingly, the humans have only programmed the learning method, but the computers have worked the recognition themselves.

With ever-increasings amounts of data and a lack of human data scientists, we’ll end up with a mountain of computer-generated data that only computers can make sense of.

We live in interesting times.

Virtual Developer Day on ADF

Oracle is regularly putting on virtual conferences called “Virtual Developer Days” with online presentations and the possibility to ask questions in real time. The current one on ADF will be running in European time tomorrow, November 26th from 10am to 1pm Central European Time.

I’ll be presenting “Top 10 tips for successful ADF projects” on the Oracle ACE track at 10.30am – if you see the presentation live, you have the option to ask me questions on chat. There are many other great presentations in four tracks – see the agenda and sign up.

It’s 10 PM. Do you know where your servers are?

Back in the 60’s and 70’s, news announcers would intone in a serious voice: “It’s 10 PM. Do you know where your children are?”

Awareness is a good thing. And if you are in some way responsible for the servers in your organization, you want to know where your servers are. And whether they are up to date with patches and security software. One IT admin found 30 unmanaged servers – Internet-facing, unpatched machines. And he was working for a serious organization that actually performed scans. I have been with big customers who had to resort to mass emails like “Does anyone use server xyz.company.com? It will be shut down in three days unless someone takes ownership.”

It’s not enough to have one measurement from your management console, because that will only show the properly registered and managed servers. You need a separate tool to actually scan you network to find all the rogue servers. Do you perform regular scans to determine which servers you are running?

Your site knocked offline for $3/hour?

According to security researchers, the price of a Distributed Denial of Service (DDoS) attack is down to between three and five dollars per hour – or a week-long attach for just $400.

Here in Denmark, someone wanting to prove a point spent about $10 to knock a piece of critical infrastructure out, temporarily keeping 5 million Danes our of their home banking accounts.

A DDoS attack is a classic example of asymmetric warfare – it costs very little to perpetrate an attack and it might cost you a lot of expensive network administrator time to fend it off. Unless, of course, you prepare in advance. You need to configure your network components correctly, and you need to put a contingency plan in place so that you can handle a situation with minimal disruption to the business. Prepare now – this will get worse.

If you can dream it…

A new startup called Coin has just started offering a credit-card sized electronic gizmo that will store any number of credit cards. In theory, you will be using your Coin card just like any other credit card or loyalty card that uses a magnetic stripe.

There are already thousands of people all over the Internet arguing why this can’t possibly work. But that’s completely besides the point.

The important point that Coin illustrates is that the Internet and crowdfunding allows any disruptive idea a hearing in the marketplace. If they can persuade enough people to pre-order one, it will become reality. It might flop. Or it might be the next great thing.

You don’t need to do market research to figure out what people want  … you just have to wait and see what nonexistent products the people are raising millions for all on their own.

If you can dream it, crowdfunding and flexible manufacturing can allow you to build it.

I’ve ordered mine.

Illiterates discussing a book

I heard part of the congressional hearing on the Obamacare debacle, and one commentator made the acidic remark “It was like watching people who could neither read nor write discuss a book.”

If your business does not understand the tradeoffs and complexities in your IT landscape, you need to educate them. In order to do so, you need to provide them with simple analogies that allow them to relate to the architectural choices and problems IT face. Do you speak User? You should.

Stop listening to your users

Here is a way you might build a IT system: The users tell you what they want, and you build what they asked for. Sometimes you get it right, especially if you have good communication with end users and an iterative approach. Sometimes you get it wrong, especially if you have no communication with end users and a year-long waterfall project.

Here’s another way to build an IT system: You observe the users and figure out what they need.

Steve Jobs had it right when he famously said “A lot of times, people don’t know what they want until you show it to them.” For those of us without Steve’s almost infallible intuition about what people want, actually watching users will provide much more information and give you a much higher chance of building something that actually helps real users.

Information Overload

I just needed some content from the Oracle Wiki. As a major enterprise software vendor, Oracle surely has a whole team managing their web presence – but apparently nobody is looking after their Wiki. It has more than a hundred content areas (called “spaces” in Oracle terminology), ranging from whole product areas like MySQL to highly technical detail topics like “CommSuite6RRSnapshot”.

Every aspect of your web presence reflects on your organization – do you have any of these chaotic, unmanaged content repositories with your company name on them?