Not a Data Centre tale, but I recently had a client come to me to upgrade from desktop to laptop, as her desktop had finally given up the ghost. I advised her on a suitable laptop and undertook to get all her data across, to run under OpenOffice 3. All went well, until it came to her word processor documents, which turned out to be wps files from Works for Windows 3, circa 1995. She had never needed to upgrade before, because the kit she had covered all her needs. Isn't that the way it should be? Change only when what you have doesn't do the job, or could be done more quickly/simply/cheaply?

BTW, converting Works 3 wp files to doc or rtf is not an easy thing, anymore, with very few converters going back before Works 6 (including current MS offerings). In case anybody else has the same problem, you can massage them through Ability Office -- as I eventually discovered.

Legacy is just what we use to call yesterday's big fashionable idea that failed to deliver only to be replaced by the latest IT bandwagon that will disappoint in turn. No doubt we will never learn.

The biggest problem I have encountered is a business problem.

The business commissions an app to fulfill a function in the business.

Say for example they buy an email system for 500 users.

Then 3 years later they buy another company, add 250 people to the existing email system but critically they don't upgrade the kit that original email system is running on.

Then they have the gall to bitch and bellyache that the system is inadequate for the businesses needs, despite their IT people telling them over and over that it won't cope.

IT people hate legacy systems because they generally cause more support calls per user than any other app. Many of them are also specific to a department and despite efforts, that dept INSISTS on sticking with their outdated, cluncky, non-company-standard app.

We are all for apps that work as advertised and don't add to our workload, at least until the business, through lack of investment, allows the app to become obsolete.

Hey, if your data centre is still alive, running COBOL etc, with the economy going as it has been, your employer is a winner surrounded by losers. Had you blown all that investment by trying to rewrite it for whatever was hot in 1990, trying again in 2002, and so forth, you might now be amongst the losers.

Some respect for legacy, please!

my old high school was a treasure trove! bbc micros, acorn computers (with a HDD Platter!) and huge plotters hmmmm!

As long as something does what it's supposed to, it's fine.

The real headaches only begin when you have to migrate away from ancient, no-longer-supported proprietary systems using proprietary and undocumented file formats. Just because a file ends in ".doc", for instance, doesn't necessarily mean that the latest edition of Microsoft Word will open it, let alone render it correctly. Proprietary software vendors have to change their save formats from time to time, otherwise users would have no reason to upgrade to the latest versions. (Of course, that's not the only reason save formats change: sometimes a radically new feature needs a radically new storage format. But at least documented file formats make it much easier to create migration tools if the need ever arises.)

At least with mainframes and dumb terminals, you could usually use a PC to simulate a user performing lookups and capture the data -- though if the replacement system was proprietary and closed, you might have a job getting the recovered data into it.

If you stick to open, documented data formats (and you've got IT staff with at least some rudimentary awareness of scripting languages) then you will never, ever have migration issues when new software and hardware come along. An awk one-liner or a few lines of Perl are all you need.

Users of proprietary, closed systems, on the other hand, may well find themselves retyping lots of data by hand one day soon. Strangely, almost nobody will regard this as being less than acceptable.

...that will become legacy after it gets deployed?"

Er, surely the answer to that is 'everything' (eventually)?

I see your 1995 Works files and raise you my own collection of Mac Word 5.1 (copyright 1992) docs, which NeoOffice won't acknowledge (though I don't think I've tried Pages... hmm...) - fortunately I still have an OS9.1 machine I can open them on, must get round to converting them all...

We have a "legacy" system. It was written in the early 80's. There are only two guys left at the company who wrote it and one is about to retire so the auditors are making us move away to a better supported product. The problem is, the current product has eveolved. It's been tweaked, streamlined and perfected. Granted, it may not look pretty. It still runs off a DOS interface using numbered menus to access the functions. But it does exactly what WE want and in the way WE want it. Some of it's features are years ahead of other rivals.

We've been eveluating competitor products for nearly a decade but none come close to the current product in terms of fitting OUR needs. Basically, we are going to have to buy an off-the-shelf product, take a huge hit in terms of functionality and, therefore, productivity and then spend several years trying to catch up to where we are today by writing out own plugins and interfaces.

No Steven Jones, Legacy is the yesterday's big idea that WORKED, otherwise it would not have hung around long enough to have gained the label!

Back around 1976, I took some early 8080A microcomputer programming classes, I'll never forget what happened at the first day's lecture. The professor described how he was a former engineer for Univac, and how as a hobby, he still maintained the last working Univac II system still in active use. It was used by a local company to calculate payroll, not a terribly complex task by the standards of the modern era, although it was a huge breakthrough back in the late 1950s. Of course he was baiting the students, and one of them asked the obligatory question, "why would the company keep such a dinosaur in production, when you could do the same job on one of these new microcomputers?" The professor thundered back, "because the old computer is already PAID FOR!"

I thought about that incident a lot, especially in 1980 when I sold that company a small microcomputer with payroll software.

Why don't you just advertise for someone to clone your "legacy" software -- probably on a Linux platform, if only because you can be fairly sure if it builds on Linux it'll build on anything -- and this time, demand the Source Code as part of the deal?

and you will be able to use your old hardware for more than just number crunching.

But back around 1999 (TaDum, TaDum), I got a call from someone who had been told I had IBM1401 experience. He had an _urgent_ need to reverse-engineer some 1401 binaries that had been running in a customer's workflow, under emulation, since, well, since that hadn't been a truly daft idea.

The latest release of the mainframe OS, required by the latest hardware upgrade, no longer (some 30 years after migration should have been complete) supported 1401 emulation.

After some discussion, he ended up just looking at the inputs and outputs and replacing those parts with Perl.

Yeah, when I hear "legacy" I think of it in terms of a Lovecraft story, where it refers to that hide-bound chest in the cellar of which it is best not to ask what sort of hide it is bound in.

It's not just in IT that individual departments insist on using idiosyncratic non company standard tech. All across the hospital where I work we use the same needle free connector for central venous lines. All except for the Intensive Care unit, where they have to use their own. So whenever patients get transfered out of the ICU, not only are they complicated and difficult (or they wouldn't have gone to the ICU in the first place), but they need the fiddly sterile procedure of changing the claves over to the ward standard.

Not even useful as boat anchors, except maybe for battleships.

Pen plotters - worth more than most cars in their day, built like brick sh1thouses, still work beautifully (made from METAL not creaky plastic shite we get nowdays) BUT they use serial ports (eh, whats that?) and HP (yep, those clowns) sites say - old model, not supported, no drivers beyond NT4.0

Nice.

Their year-old inkjet models don't properly support 64bit, which we've just upgraded to... grrr...... so now we have MORE legacy tech... the waranty ran out on 24/10/08 so I can't even drop it down the stairs to 'upgrade'

Maybe it's different in the US to the UK - I always hear how the US is about three years ahead in deploying new IT ideas - but here in the UK we don't seem to get rid of anything unless we really have to. When I was a young and naive graduate, I had this idea I would jump into a job with a company brimming with the latest technology, but my very first job entailed working on a system almost as old as I was!

The pre-Y2K period was a real eye-opener what with the number of dinosaur systems coming out of the woodwork, all having to be checked to see if they were Y2K compliant, the usual request being not to replace the old systems but to find ways to "temporarily" get round the problem - some of those dinosaurs I helped "temporarily" patch are still grinding away! The scare of my life was when I was working late with one of the contractors brought out of retirement (!) to decypher and write updates for in-house COBOL apps from the '70s, and - tired of my jokes about dinosaurs - the old fart faked a heart attack!

I'd be intersted to know from US readers if they see the same in corporates there, with re-use rather than replacement being king?

> But back around 1999 (TaDum, TaDum), I got a call from someone who had been told I had IBM1401 experience. He had an _urgent_ need to reverse-engineer some 1401 binaries that had been running in a customer's workflow, under emulation, since, well, since that hadn't been a truly daft idea.

> The latest release of the mainframe OS, required by the latest hardware upgrade, no longer (some 30 years after migration should have been complete) supported 1401 emulation.

Knowing IBM, what's the betting that this was heard around 1996:

"Have you seen our latest company-wide software and systems audit? They've done something about that last 1401 system - it's not listed any more."

"Really? I'll make some enquiries, make sure we're not treading on anyone's toes, but I'd say we can ditch 1401 emulation in the next upgrade."

---

Legacy *can* be the idea that worked. There are legacy systems that were tailor-made then, and still fit now. There are also ones that don't fit now. And also ones that were an expensive, troublesome "best bodge" back then, but could now run on a LAMP server, with web browser clients accessing it over SSL - if the people in charge could get their fingers out and make the changes.

That's a myth.

There are running, functional systems, and there are broken systems. Legacy stuff that is still running, and functional is (by definition) not broken.

All of the examples listed in the 17 (at this moment) comments above are MANAGEMENT errors, and have nothing to do with the functionality of the systems being discussed.

The biggest management blunder is not having a plan for the future.

The second is not standardizing on functional systems (including upgrade path).

The third is not properly training staff in the use of those systems.

I have 30 year old machine code (running in a PDP11 emulator on Linux) that one client of mine is happy as hell to have ... It helps him keep track of a portion of his business that would otherwise require three or four pairs of technically trained eyes. (It's a specialty greenhouse, if you're curious ... the guy grows orchids, and keeps track of humidity, temperature, light levels, soil moisture, pH, salt, etc.). It originally ran on a pair of PDP11s (redundant systems ... hardware was flaky 30 years ago), now it's on a pair of dead-screen Pentium laptops that cost a total of $50. His electricity bill dropped, and the UPS powers the laptops for a little longer than the PDP11.

Is legacy stuff inherently "bad"? Nope. Bad management is, though.