If you would like to read the next part of this article series please go to Interview: Maintaining Legacy Software (Part 2).
Sometimes the most practical (or even only possible) choice for a business is to continue running software applications or even operating systems long after they have passed out-of-lifecycle. The reasons for this are varied but usually it boils down to the fact that the goal of any business is to continue making a profit and not to enrich software!
While our own Deb Shinder is currently writing a series of articles called The Risk of Running Obsolete Software over on our sister site WindowSecurity.com, the operative word here is “risk”. Yes, it’s a risk (especially in the area of security) to continue running obsolete software that is no longer supported by the vendor. But business is also about evaluating and taking risks, because without risk there usually isn’t much in the way of profit happening.
So to provide some counterpoint for Deb’s excellent series of articles I’ve here interviewed Michael L. Hallsted, a long-time IT pro who has been working with computers since the late 1970s and who currently is the “part-time IT guy” for a small manufacturing company.
The Interview (Part 1)
MITCH: Michael thanks for taking time out from your busy job as an IT pro to let us interview you for our website.
Michael: Hello Mitch and thank you for the opportunity to answer your questions. Hopefully, this might prove useful, somehow, to someone, somewhere.
MITCH: Michael can you start by giving us a bit of background about your job at the company you work for?
Michael: My company was founded back in the 1940s, and over the years, they have specialized in the fabrication and manufacturing of agricultural products: tillage equipment, bed shapers, incorporators, shredders, mulchers. The main product was the tomato harvester. By the 1990s, they had produced and shipped tomato harvesters all around the world. We are also a small company, maybe 110 employees by the time I joined in the late 1990s. We have one location where all manufacturing, assembly, and painting of the tomato harvester took place. Since they had been manufacturing tomato harvesters for 40 years, they had a good idea what the actual cost was of building one (around a quarter million dollars), but they could not tell you, exactly, the cost and time associated with each step of the process. Part of this desire, to figure it all out, was due to our main competitor at the time, FMC, a global corporation and huge defense contractor, who also made tomato harvesters for some reason. [Side note: Around this same time, the people in charge of the tomato harvester division at FMC, got tired of being ignored, bought that division from FMC, and started their own company.] So, to figure it all out, they decided to spend close to three hundred thousand dollars on MRP software and a server for it. The MRP company told them the software needed one NetWare 5 server or we could buy three NT4 servers to handle their software and our company network. Well, that was a no brainer, NetWare it was. Basically, they bankrolled the company to figure out the true cost of their tomato harvester. Also, because of their close ties to agriculture and the tomato growing season, every year was feast and famine, tied to the growing and harvest seasons.
Other notes: When I was hired, there were maybe ten computers total there, for the engineers and office personnel, all running either DOS or Win95, using NetWare 4, and the entire LAN used BNC cabling. By the end of my first year, I helped them acquire and setup 50 computer and 5 barcode workstations, which easily added another twenty thousand dollars to what they already spent. We also hand rolled all of our new network cabling on our 23 acre lot, both fiber and cat5, because the cable installer told my boss, to his face, that he could not do it right, it would not work, and he did not want to come out to fix our mess. Well, my boss laid out and installed the entire cabling system himself, then got a different guy to help him with the termination points on the fiber runners, and it hums along perfectly to this day. The only problem we ever had with cabling was due to the machine shop erroneously wrapping the extension cord of a large piece of machinery around an exposed section of network cable. Every time the machine was used, it sucked up so much power that it cut off the network connection to the engineers. That was a head scratcher because it was a sporadic problem, until I thought about visually inspecting the cable run and noticed the extension cord.
Anyway, the first year of using the MRP software involved intensive employee training. Then after another 2 years, the company ended up selling the tomato harvester product line, lock stock and barrel, to our main competitor. Why? I never did ask, and I am not going to hazard a guess. The result, though, was a small company became a tiny one. We went from 110 employees down to 9 employees. I still had my job, but now I had to figure out what to do with a sudden abundance of fairly new computers. This also meant that replacing their really expensive MRP software might not happen within my lifetime. I stayed on full time for a few more months, stacked and cataloged all of the extra computers in spare offices, then started working part time, one day a week, fixing the miscellaneous computer problems and maintaining backups for them. And 15 years later, that’s how I spend my Saturdays.
MITCH: Why do you think some businesses keep on running legacy software for years, sometimes even beyond the supported lifetime of such software?
Michael: Besides the issue of money and being able to afford the cost of new software and equipment without bankrupting the company? Or the amount of time and energy needed to retrain employees on the new system and then dealing with the errors and mistakes they might make while learning? Or overcoming the psychological resistance of the employees that these changes bring, impacting their desire and willingness to embrace the new system? Basically, to me, businesses run legacy software because it is a proven system that works. It fills their needs and does everything that they might need or want.
MITCH: Which kinds of businesses are especially prone to such practices? How widespread is this kind of thing in different industries?
Michael: I don’t have much experience outside of manufacturing, but I do imagine that manufacturing and industrial businesses are prone to this due to the cost of the machinery involved and the software that was written to go along with that particular piece of equipment. To get a feel for this, all one has to do is visit local businesses and notice the type of technology in use. The blood donation center in my area still uses dot matrix printers for their forms. Sitting in my dentist’s chair, they have WinXP running their x-ray machine and appointment software. Watching the waitress at a local restaurant input my food order in their system, it looked exactly like an old DOS based menu system, which put a smile on my face. And speaking of DOS, a local car repair center has been using DOS based software forever to track the work they do. They are excellent mechanics, which is why you go to them to fix your car. They have all the latest technology to fix the new cars, but for their office, they like their old software. Also, while visiting a local MyUPS store, the owner looked perplexed, I asked him the problem, and the alarm sensors for his front door were not functioning right. He inherited this particular system that used WinXP and needed a serial adapter to interface between the software and the wiring to the front door sensors. He had bought a new serial adaptor, but the software and the computer did not recognize it, and UPS phone support was unable to help him. I offered to look at it. According to device manager, the serial adapter just needed a driver, XP offered to go out and find one, it actually found one, installed it, and life was good again for the store owner.
MITCH: What are some of the legacy operating systems, platforms and programs that your company is still running, and why are you still using each of them?
Michael: We have one NetWare Server (v6.5) running our MRP software, Visual Mfg. v6.1.3, as NLMs, using ipx/spx as the communication protocol. During installation of Visual Mfg on the workstations, 16-bit database drivers are installed. This meant that 32-bit WinXP is the last operating system we could ‘natively’ use. We use MS office97 and MS office 2000 for word documents and spreadsheets. I use a win2000 computer to run the daily backups. I use a win95 computer to run the database maintenance scripts in SQLTalk, because on win95, SQLTalk will display the number of seconds each line of the script takes. With WinXP, the number of seconds is not reported back in SQLTalk. And our solution for receiving faxes and printing them on our networked printer: 10 years ago, in 2006, we bought an aged Macintosh PowerPC G3 [ 300MHz ] that was hacked by its former owner to run OS X Tiger, v10.4.7. It came with an equally aged US Robotics serial modem. Starting with OS X Tiger, Apple included basic fax functionality. We just use it to receive and print faxes, but it works great for that, and is still going strong 10 years later. The original apple monitor gave up the ghost awhile back, but I found a 3rd party video adapter with built-in toggle switches that allows you to specify a certain frequency, video resolution, and area of the world, enabling you to use any modern monitor with older models of Macintoshes.
MITCH: Let’s take a closer look at your company’s continuing use of NetWare. What do you do if you need support for your NetWare environment? i.e. what do you do if something breaks? I think I heard for example that Micro Focus (which acquired Novell) hasn’t released any new service packs for NetWare since 2009.
Michael: Support?!? That’s what Google is for. But seriously, I learned a long time ago that once you mention the word NetWare, people who had an interest in helping, stop returning your emails. Basically, if something would ever break, it is up to me to figure out how to fix it. Fortunately, all we need is a very basic NetWare server, nothing fancy, since all it has to do is run the MRP software. And [knocking on wood] it has run solid and true for us over the last 17 years. The only issues we have ever had, have been hardware related, but that is what mirrored server hard drives are for.
MITCH: I just checked the Novell website and notice it says their Extended Support period for NetWare 6.5 will be running out at the end of this year, see this link. So it looks like even Novell is pushing its last remaining customers to upgrade, in this case from Netware to their Open Enterprise Server. How will that affect your company’s network?
Michael: Well, it won’t affect us at all, since we need NetWare to run the NLMs for our MRP software. So, there’s not a lot of choice for us, and as long as we can still buy older Dell servers, we’re good to go.
MITCH: In general doesn’t the high cost of extended support contracts for legacy or out-of-lifecycle software tend to make it more cost-effective to upgrade to newer software versions? For example does the extended support contract for your NetWare take a big bite out of the IT budget for your company?
Michael: Well… in general, I would probably agree with you. For us, we haven’t paid for support contracts for almost 15 years now. Support contracts can cost a lot of money and stopped making sense for us a long time ago, due to the changes the company has gone through over the last 15 years. Budget? That brought a smile to my face… I have never had an actual budget to work with. If I could demonstrate a need for certain software or hardware, it was never a problem. Otherwise, I had to make due. Open-source software can be real handy at times.
To Be Continued
Stay tuned for the second half of this interview to be published soon right here on WindowsNetworking.com!
If you would like to read the next part of this article series please go to Interview: Maintaining Legacy Software (Part 2).
