Updates and Obsolescence

[Bob_Y.]

In this post, the main question was about how to make an installer for an old system. As part of gathering information, there was a bit of discussion about obsolescence.

http://forums.ni.com/t5/LabVIEW/Create-installer-without-labview/m-p/2793556#M820187

Rather than further sidetracking that thread, I thought I would ask here: How old is too old to support or what should be the plans for replacement/updates of old systems (that still work)?

I look forward to the discussion.

Bob Young

[JÞB]

Nice re-direct Bob!

How old is too old?  If its not readilly comercially available from multiple vendors its too old!  

That said, there is a lot of room for customizing company policy to meet specific business case objectives.  THE MAJOR point being to HAVE a policy in place.  I recommend a 3-5 year review cycle for each program.  The review needs to have a deliverable.  A report that states what the obsolecence state of the project deliverables is today and what the impact on the busness is if project deliverables fail or are subjected to natural disaster.  Somewhere up the food chain some director can then make an informed discision on when to budget for upgrade.  This keeps shareholders happy to know that the company is planning to maintain infrastructure.

Assume: Project X that delivers a solution for 100% production test of widget Y that is 15% of the companies annual sales with a projected production lifetime of 35 years.  There are three replicates of Project X each operating at a 75% tempo (There is some required down time for calibration and other routine mainainence)  Loosing one critical component will impact shareholders.  This project may need annual review! (and a well stocked spares kit!)

On the other extreme Project A delivers a solution that is used for engineering evaluation of returned product.  Useful for your continuation engineers but it runs a batch of 6-10 items every 2 months (Congrats, your Production Standards are quite good)  You could let that go obsolete and wait to upgrade only after the lab co-ordinator can't find replacement parts on E-bay (He can keep an intern busy.)

KEY TAKEAWAY: Every component in every project will become obsolete!  Acknowedge this as a fact up front and think about why you bought the delivered solution in the first place.

[RobertoBozzolo]

How old is too old? As of my experience with certain customers, if it works then it's not too old!

I have a customer of mine that is still using old DOS-based equipments for end-of-production testing   . It's not that they don't know they are old, but they simply claim they do not have need/money to refresh them. We've refactored two equipments in the factory, but there are still three or four of them in original state.

Now this thread comes in handy: these very days I had to revive a very old machine that was left in a warehouse for 20 years (  ). I had to dig in my memory to revive informations about himem, edlin, clipper, the good old Basic 7... and I had to find in the office a PC with a floppy drive to restore the software from the backup to that machine.

A serious IT archeology task! 

[PaulG.]

I've seen robotic lines that are over 20 years old. This kind of equipment really shows it's age but a company would rather nickel and dime itself with service calls than go to their only other option: replace the entire line. Service contracts cost thousands. A new line would cost millions.

I don't think there are any real rules or principles regarding obsolescence. You can have all the meetings you want but the bottom line is that a company will replace an old system when the old system simply doesn't work any more and cannot be repaired. "If it ain't broke .."

[SnowMule]

I thought replacing machines with PCI for machines with PCI-E was a headache. 

Floppy drives?

[RobertoBozzolo]

I was Lucky I had a spare box of 1.44 MB disks left on a shelf! :smileyWink:

[Darin.K]

---

@RobertoBozzolo wrote:

I was Lucky I had a spare box of 1.44 MB disks left on a shelf! :smileyWink:

---

A couple of months ago I received a request from Amazon to answer a customer question since I was the most recent confirmed purchaser of a box of 5.25" diskettes.

[RavensFan]

So what question did the person have about 5 1/4" disks?

[Darin.K]

The picture and description differed on whether they were double density or not.

[JÞB]

---

@Darin.K wrote:
The picture and description differed on whether they were double density or not.

---

You know I have to ask.

WT($@#$) did you need them for? 

It sounds like a great example of failing to plan for obsolescense.  Would you care to share lessons learned?