I have a client who is still using an application written for dBASE II on a Kaypro under CP/M. It’s now compiled with Clipper and uses dBASE III files, but the core source code hasn’t changed. The output is fed into other programs as pure ASCII text, so the graphics and formatting side can be done with any program they want.

I’m working on converting it to something more contemporary, but I’m not sure they’ll like it, because the office people are still typists and don’t like pointing devices.

In certain fields, like engineering, there are liability problems if they change things. You have to have confidence you are getting the right answers.

I did a contract in the late 90’s for a large utility organization. At the interview, I was asked if using DOS would be a problem for me. I laughed and said “That would be a pleasure!” To which the guy interviewing me (who turned out to be the chief engineer, their HR people couldn’t handle the interviews properly) said “Really? The last thee people said it couldn’t be done.” And I said “How old were they?” And we both laughed and I got the job. 🙂 It turned out to be very easy because I could reuse code I’d written over the years. 🙂 They still use it as far as I know.

Hard to believe he coded all that stuff over 15 years ago now and it’s still running and doing exactly what it’s supposed to be doing. A small city gets its water and hundreds of thousands of people get their power courtesy of 15 year old software running on simulations of 15 year old computers. Wow. Needless to say he prefers his newer software running on modern Windows variants, but these folks truly are taking “if it ain’t broke, don’t fix it” to the logical extremes :-).

I have written a lot of “core business” software, programs that are essential to the business, not set decoration. When it breaks because of an upgrade, I have to work until the problem is resolved, because the business is dead in the water until the software is running again.

Trying to convince people to avoid the “latest and greatest” is the biggest challenge there is when people have convinced themselves that they have to have the newest technology to show they are a forward thinking company.

If there is a security concern, I upgrade, otherwise, forget it. There are enough things that are broken and need fixing to tie up my time, I don’t need to go looking for trouble.

My personal web server runs Debian Stable. Obsolete, and reliable as dirt. At work I typically run whatever version of Apache ships with the version of Centos/RHEL that our software release is certified on. Red Hat has done a very good job with their RHEL updates of not breaking things, as good as Debian as long as you stick with one version of RHEL. Yeah, you’re running obsolete software for four years. And that’s a problem… why?

One of the things I have to keep on our China team about is *not* upgrading to the latest/greatest version of everything. For example, they ran into an issue where the version of Python provided with RHEL5 didn’t support something they wanted. They proposed upgrading to a newer version of Python. I put my foot down *hard* on that nonsense. The only thing we install are official packages and security patches from Red Hat. Period. End of discussion. And even those get carefully scrutinized — I have them look at every patch to see what is different between the old version and the new version to make sure it won’t break anything. Our software is certified to run on a specific OS, and while it might seem cool to upgrade pieces of that OS willy-nilly, the end result is nothing but mess and we don’t want to go there unless we have absolutely no choice (which has been the case sometimes — such as when I backported the RHEL4 bridge support to RHEL3 because one of our customers needed bridge support — but it isn’t something we do lightly). As someone who’s been doing this sh*t for over 25 years now I know the dangers of willy-nilly software upgrades from sad experience, but the newbies just out of college need some convincing with a limp herring to the face before they get it. So they just have to work around the limitations of the hoary old OS we’re using because that’s all they’re going to get. Heh.

Yeah, I’m a mean penguin sometimes :-). BTW, Linux sucks. The problem is, everything else sucks even more :-).

– Badtux the Linux Penguin

That isn’t my idea of blogging, i.e. something I do when I’m waiting for something else to finish, or I need to vent. If you blog to release pressure, having an aggravating platform is not the way to go, no matter how many bells and whistles might be in the box.

The responsibility is a PITA, but the control is nice.