Fix it with duct tape – a note on technical dept

Last week I recorded an episode of Architecture Corner on the subject of technical dept.

Take a look at:

As you’ll see, technical dept is creeping slowly into your realms. You won’t notice anything to begin with; it’s a bit like the dust in the corner or the paint that is slowly starting to fall off. You get used to it and finds ways around it.

Posted in Legacy, LeMo

Are your buying habits for sale?

Auction Tsukiji fishmarket

Auction Tsukiji fishmarket

The RadioShack bankruptcy have led to a situation where customer information, that were promised never to leave the company, are now up for sale. Last week, I recorded a video about this and other privacy related issues; e.g. how is the wearables revolution going on drive privacy issues? Will you even be able to control the usage and will you – at all – be able to control the collection of information? The convenience of using a wrist band will overshadow the privacy concerns; at least until the public is made aware of the consequences.


The basic consequence

You become the product. And that’s fine as long as you take an active decision to become the product. Do you?


Passive collection without consent

Today, Burger King in Sweden were exposed to collect their customer’s car registration numbers. and, to make things worse, with no option to opt out. When can we expect information about who visited which restaurant at what exact time to be available to the highest bidder? Of course, the outfit will tell us this information is for internal, market planning purposes only, or that there is no way of tracking it back to customers, but by now, there have been too many examples of information leakage, based on breakage or based on monetizing decision to believe that to be true.


Posted in LeMo

Open source licenses means free and nothing to consider? Think again

There’s a worrying trend amongst corporations taking up open source solutions as part of their enterprise portfolio. Open source penetration is of course great, but I’ve seen some examples where companies simply didn’t do their home work when it comes to the license part. As with any license, be it proprietary, commercial or open source – you need to read it! And you need to comply with it. Complying with open source licenses is not that hard; most of the time it’s simpler to comply to than with commercial counterparts. But I’ve seen many examples where companies as part of their core system are using a “personal, free for non-commercial use” open source product  – where the commercial one would cost them €50! So it’s not so much about saving money, but simply thinking open source is exactly the same as no cost, in all situations.


The most common trap comes to extending or modifying open source solutions. Most often there’s an obligation to share your add-on, improvement or tweak; an obligation not always observed. But think of it. You got the solution for free. The only thing you need to “pay back” is to also make your additions available for free. Simple as that. But if you don’t, there’s a growing risk in you being sued for non-compliance with the license. This could have different consequences, one being you have to pay, another being you need to stop using the open source component as part of your solution. This could have anything from minor to all-encompassing effects; perhaps a total rewrite of your solution?


So, following an open source license is not hard or expensive; just make sure you make your home work. Have a strategy for using open source software; it’s enough with a one-pager – but don’t leave it to the individual developer to handle.

I’ve recorded a video talking about this subject.  Joakim - from AC#4 Open Source discussion


Posted in Chronicals, Open Source

#WASI – Wearables + Augmented Self + Implants

When is reality?

What’s science fiction and what’s reality? Or rather – when is something science fiction and when does it become real reality? RFID implants is something most people wouldn’t even dream of, but already in 1998 professor Kevin Warwick  had a small glass tube with an RFID chip operated into his arm. Most people would still view that as something very far out, but last week office workers at an office hotel in Stockholm were offered the same; to unlock and open doors, to start a printer, etc.

Real-time translation and image overloading, where e.g. a sign is detected, translated and an overloaded image with a translated text has been mentioned and prototyped for some time, but the time from Sci Fi to reality have been really short. One alternative available since January 2015 is within the Google Translate app.

Automatic translation with direct image overloading. Soon in a wearable on you? Or in you?

Automatic translation with direct image overloading. Soon in a wearable on you? Or in you?

(Image credit

This app also supports bidirectional audio translation, yet another invention from Sci Fi literature, in the form of the Babel Fish from the Hitch Hiker’s Guide to the Galaxy book series.


Right now, wearables is beyond hype and going mainstream; but one for one there’s not too much in it, the real advantages comes from the ecosystems surrounding the wearables. And the ecosystems are in deer need of standardisation. But what’s needed for it to grow?

Augmented Self

Perhaps when the wearables goes beyond being just a gadget and actually improves things; or improves you? When they trigger a better behaviour for e.g. physical training; measuring pulse and position is good for some, including myself, but for it to go beyond the early adopters and addressing the early majority, it needs to be simplified down to a level where it’s totally un-intrusive and automatic. But in order for it to be fully un-intrusive it might need to be physically very intrusive.


Like in implants; what more intrusion could you think of? Yet some people in Stockholm are “chipped” since some time. Beyond the pain and awkwardness the real doubts are dealing with privacy. If I have a chip in my hand to open a door, that chip could of course be read, analysed and quantified by whoever have the equipment.


So what’s the next level of “personal computing”? Perhaps when the computing goes totally personal, as in being a part of you. Science Fiction? Yes, certainly now, but don’t expect it to stay Sci Fi for too long. Privacy issues have to be dealt with, but it haven’t stopped technology roll-outs that dig deep into people’s privacy yet. Likely it will just gradually go to the point where the Internet of Things include you, literally.

Posted in Chronicals, WASI, Wearables

API Economy – a lever to free your core systems



Core systems are often seen as an effective blocker to innovation and introduction of new technology. With a shifted mind-set you can break the tedious define business requirement, formalise, prioritise, design, reprioritise, implement, test, deploy-chain that can take months or years. If you introduce multiple velocity layers in your system landscape you can allow fast innovation and still make sure your back-end systems will have enough time to ensure data and process quality.



Open API, Open Data Lake and other concepts are at play, forming the API Economy.

More on this as the Capgemini CTO blog.

Posted in Agile, Digital, EA, Innovation, Legacy

Why it’s a good idea to deploy 10 times per day

Legacy landscapes and systems are pestered by many things, one being the very long release cycles. Looking at some of the Internet age success companies, deployments don’t happen a few times per year, but rather many times per day. This of course builds on a set of pre-requirements but also opens the door to actually benefiting from e.g. cloud deployment (whether private or public).

Himeji-lo castel. Maintained with a cyclic rebuild, over and over again.

Himeji-lo castel. Example of Continuous Improvement in the real world.

Last week I did a lecture at NFi Utbildning where I discussed the mechanisms behind legacy systems problems and how a DevOps approach could be on piece of the puzzle in avoiding building future legacy problems.
The presentation is here (and as always with my presentations, they’re quite a lot more understandable when heard live…)

Posted in Agile, EA, LeMo

Remaining successful in the Industry 4.0 age

Last week I was speaking at Capgemini’s annual Architecture Week, this time for manufacturing clients. My reasoning was around how you should think when planning for disruption – in essence, how do you as an Enterprise Architect plan for the unplannable?
Three topics:

  • Differentiated velocity layers (aka. Dual Speed IT)
  • OpenAPI and Open Data Lake = API Economy
  • Third Platform

A bit more to come; here’s my presentation:

Posted in LeMo

So, when do you start your modernisation?

NOVA System

NOVA System, my first “real” computer

There are a couple of things that almost always are true when I get in touch with an organisation with modernisation needs:

  • It concerns the very core of both the IT landscape and the business operation
  • The system(s) have been there for ages, very often it consists of system(s) from the first generation
  • The systems are very often monolithic, i.e. it’s in effect one single system with more or less separate modules (often less)
  • Very few people know the system and they are about to retire soon. Or have retired and are now billing outrageous hourly rates to keep on doing what they’ve done the last 20 years or so, only slower.
  • The cost of ownership is high and rising and the ability to deliver new solutions is low and sinking

Yea, and one more thing – my clients have always already tried to replace the system 2-3 times. Or 8.

When I come in it’s always one of two issues triggering:

  • Technology obsolescence. The vendor have finally pulled the plug on the platform, OS, language, database or the like is no longer on extended life support.
  • Cost is skyrocketing and someone compared it with the subjective business value; e.g. a new CIO or CEO enters and see it for what it actually is.

It’s interesting to note that these IT organisations have always been very aware of the problem and have, in various ways, tried to mend and handle it, but they’ve never been able to show the long-term effect of not handling the problem correctly to the business owners. Small things like a mainframe OS upgrade become huge both from cost and risk perspective if you wait 15 years. Sure, the short-term business value of an OS upgrade is zero; but you build a huge technical dept that can eject to a cost level where it swamps your IT budget, inhibiting any other change or improvement for one or several years. Years where you have to tell your business owners, “sorry, but those bottom line improvement will have to wait”.

Bottom line: Don’t wait with those infrastructure or platforms upgrades. They will just become more expensive and risky. Don’t wait with that compiler upgrade, or that database engine renewal. Make sure your IT development goes fairly hand-in-hand with Moore’s Law, not opposite to it like the organisations I describe above. Because one day, you’ll need to modernise the application on top of the infrastructure and having an up to date infrastructure make life much easier.




Posted in Chronicals, LeMo

Do you have a systems decomissioning departement?

Most companies have a systems development department; it could be called IT department, business development, digital innovation or whatever. But do you have a systems decommissioning department? Very few organisations do, and that’s one of the main reasons old systems stay around, for ages.

Here are some slides from a presentation I did back in 2011.

Posted in EA, Legacy

Retirement home next?

This is a serious and upsetting title, I know – but I see many organisations that need to do some serious planning for their future. Not all have yet understood it, but I’ve come across major Swedish organisations, companies and authorities, that have been caught in a severe trap in that their core IT systems can only be maintained by a few, elderly people. Now – I don’t have anything against elderly people per se, but I cannot understand e.g. a major bank permitting the IT department to have one single person, that actually would have retired three years ago, being the single person on this planet who actually know their core system. You know, that system that will stop all business if it doesn’t work and that would cause a bankruptcy if failing a couple of days. My colleague Sander Hoogendoorn described a similar situation in his blog called  Alzheimer Architecture. My client, who eventually understood they couldn’t demand a single 70 years old person continue this endeavour, employed a younger person to take over. That’s right – one person. Still only one person. They’ve taken away the acute Alzheimer threat, but rest assured, it’ll come back.

Now – do you see this situation in your organisation? I’m sure you don’t have business critical systems dependent on a single person’s memory; a memory that perhaps soon is sagging?


First published 2010-10-20 on

Posted in Chronicals, EA, Legacy, LeMo