Why we should learn to stop worrying and love legacy – Fujitsu's UK head In the UK, IT Godzilla Fujitsu is perhaps best known for its unwieldy public sector contracts, being responsible for running a sizeable chunk of the government's legacy technology. Indeed most of its UK and Ireland revenue has historically come from the public sector, some 70 per cent at the beginning of the last Parliament in …
I've seen a number of reports where replacing any occurrence of 'digital' with 'fairy dust' would have no discernible effect on the content.
It seems to be a catch-all for how people with no knowledge or understanding persuade others (with even less) to give them lots of money.
I wish there was an antibiotic for cynicism. I need a large dose.
'If you mentally block out the word "digitalising" and replace it with "computerising" then it all works so much better.'
That doesn't work here.
Consider:
'the company is also winning "digital" contracts, too – including the digitisation of a number of train companies' front-end ticketing systems.' Aren't these systems already computerised and, therefore digital?
or:
"we have a lot of demand to digitalise the front end of those legacy systems – rather than full legacy modernisation."
and especially:
"Moran believes digital and legacy are not mutually exclusive."
In order to make sense of this one has to assume that either these systems aren't already running on digital computers or that she has some previously unknown meaning of "digital" in mind.
Did she get parachuted in from some entirely different industry?
How dare anyone imply that there's any value in legacy code, or that old systems worked as well as new ones. Everything should be written in javascript these days, should only work on tablets, and a new version should be released every 3 minutes.
As far as I'm concerned, any app that didn't get borked by the removal of left-pad from NPM last week has no place in the modern world.
"Everything should be written in javascript these days, should only work on tablets, and a new version should be released every 3 minutes."
That just about sums up my work this week. I wonder if my wife knows she's married to a modern, digital man. I wonder if she cares. (It's 12:43. What do you think? Too early?)
[W]e shouldn't necessarily assume something is irrelevant because it is old.
Lest we become magpie developers.
More succinctly, in the words of William Inge...
There are two kinds of fools: one says "This is old, therefore it is good"; the other says, "This is new, therefore it is better."
I work for an out-sourcing company doing support and development of "legacy" systems for various customers. The most modern technology used is 15 years old. When the customers finally decide to replace the whole thing with something agile and "Digital", I could be a bit stuffed. The customers think me and my employer are slow-moving dinosaurs. Because I work for multiple customers, it's unlikely TUPE would be applicable. I spend at least 10 hours a week of my own time self-training in modern technologies, which are almost completely unrelated to the stuff I use at work, and this is likely to continue. I don't see how this is sustainable.
heheh, you're not doing it right then.... :p
Tell that to all old school Dev's who came out of retirement to work on Y2K. The clever ones earned more in three years than Young Padowans earn in twenty.
If I'm still alive ~2038 I know how I'll be topping up my pension.
"The customers think me and my employer are slow-moving dinosaurs."
Here is the nub of your issue, the image your employer projects to it's customers. I suggest you need to take this up with your employer because unless they address this, they will not be asked to bid for the development of the replacement system...
I suspect Fujitsu, among others, are realising that the downsizing fad of the late 1990's didn't actually solve the 'legacy' software issue. Yes, it got rid of a lot of mainframes and the code and DBMS got refreshed, but there is much in business and government that doesn't change particularly fast and so this 'new' software has quietly done it's job and 20 years have gone by...
What amuses me is that a projection I made in 1997, about the viability of the mainframe platform was largely correct, namely, the commodity Unix/Windows platform with it's rapid rates of innovation and change was inappropriate for these long lived systems; there is a real business reason why the IBM mainframe maintained software compatibility back to the 1960's...
Trees smothered the bushes that sheltered their sapling from grazers. The bushes smothered the grass that protected their soil from erosion. Even the grass didn't start the soil it is standing in.
There never was any mercy. There never was any sanity. Completely succeeding in what they are driven to do guarantees their own extinction. Everything depends on something escaping the encroaching shadow somehow, somewhere, to return, repair and reboot the system after each crash.
And yes, there have been many, many 'successful' self-extinctions. And will be many more.
So Run, you fools..
(because there's never enough shelter..)
One item overlooked with legacy code is often it powers such mundane systems such as payroll. These are systems that must work regularly and often. Replacing this working code with new code risks chaos for several weeks while the bugs are worked out.
But there is a point when so much cruft and dead code has accumulated that making modifications risks bring the entire, creaking structure down. The trick is to recognize when one is reaching this point.
Dilbert! Perfect! How much legacy spaghetti code churns about in the computer farms of the finance and airline industries? They, along with government, were the first to use computers in a big way, and they are stuck with all that old COBOL. And maybe assembly language subroutines that nobody can decipher.
" they are stuck with all that old COBOL."
The world has moved on, legacy these days also includes C, Java, Javascript, VB, Ruby on Rails etc. ie. any programming language/environment that was popular in the 1990's and early 2000's...
But saying that, a payroll system implemented using COBOL is probably easier to maintain than the same system implemented purely in C or VB...
"Reports also suggested the supplier won its £700m legal dispute with the NHS in 2014 over the cancelled National Programme for IT contract. "
There's such a thing as a "pyrhic victory", even when the spec was incompetently set in the first place.
Taking on a contract you know can't possibly work as specified, then charging enormous overages making it work beyond the (non-working) spec is one way to make a short term profit - but when someone else comes in, starts over on the resulting clusterfuck and does a working system for half your final costs (or 1/10 in some cases I've seen), you're not going to win any repeat business from the organisation concerned - or new business from anyone else watching from the sidelines - especially if you take your customer to court for payments on a system which didn't actually work and had to be binned. Such things get attention and you can only do stuff with a handshake on a golf course (or by using backhanders(*)) so many times.
(*) I'm not accusing Fujitsu of this, but I've seen a large number of contracts won by the company that does the most wining and dining of senior management, instead of on technical and cost grounds.
But gubmint bureaucrats change regularly, get promoted, retire, move sideways... So writing the deal in the club house or over cigars and scotch brings with it time to tweak the contract, 'develop' and spread the hours needed to drive revenue through the roof and fund your lawyers to drag it out after it goes belly-up.
Your developers wrap the turds in an extra layer of glitter and run from the stink that threatens to consume their sanity. The pollies swap batons at the election, the bureaucrats get promotions for sending confidential data to 'accessible' clouds and adding a digital microsite and mibile app. You win the lawsuit and collect more free taxpayer money over something that no one has time for anymore, cares about, or actually understands.
Everyone who matters wins!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
