from the memory banks …

I posted this in a conference on CIX in October 1998, and was reminded in an IRC conversation with <lj user="syllopsium" this morning.

and I thought I’d seen it all . . .

after a conversation with a certain Dane (that was landsmand of this parish), I was encouraged to tell you about the place I’m currently contracting at. It’s an ex British Steel Company site, consisting of about six manufacturing units
operating under a holding company. A multi-domained NT server site, with a few odd machines included (when I say “odd”, consider that a mission-critical app is running on OS/2 v1.3). At least, when I say /operating,/ I’m amazed that they ontinue to function at all. In no particular order, we have:

1) Internet e-mail (via MS-Mail – no Exchange here) which relies on somebody instigating a manual dial to Demon at 8.00 a.m., 12 noon, and 4 p.m. (this is the one that made Henrik collapse in hysteria)

2) across the two companies with which I’m involved, two NT domains, incorporating 6 NT servers. One of which doesn’t have much on it, and can’t browse the network – no, I don’t know why.. The /meatiest/ is one running 3 huge SQL databases – P120, 160-ish Mb RAM (of which SQL is grabbing +/-115Mb). This is, astonishingly, generating huge amounts of error messages . . .

the PDC is a P100. The BDC is a P75. All of the servers have had their swapfiles reduced, as they’re short on disk space.

3) I was called down to the Lab today, because their IBM Ruggedwriter (what?) had stopped printing. It’s running off a 386 with 4Mb, DOS 5.0 and DOS Shell (I kid you not), which hooks straight into an HP3000. The printer’s printhead is buggered – maintenance? Don’t make me laugh. And no other wide carriage dot matrix printer on-site. So the lab stops printing reports. This is a 365/24 site, right?

4) there’s an IBM RS6000 on the network. An old one. Not Y2K compliant. They’ve bought a new one. From Computercentre (oh dear). It’s been on-site for 6 months. Its OS is non-compliant. Its SQL is non-compliant. Its applications are non-compliant. The purveyors of its bespoke software now (as of yesterday) say that they won’t be in a position to supply updated versions of their software for another three months. And this for a company whose lead times are in excess of 12 months (count the months until year 2000).

5) after a meeting with the MD yesterday, I discover he’s not prepared to expend any more money on the WAN/LAN, unless *I* can come up with quantifiable figures. e.g. how much traffic /should/ the network carry over a given bit of wire, and what targets we’re looking to hit with any changes we make. He’s “fed up with seat of the pants stuff”. He says that IT is the same as any other process in the purchasing loop – we must provide documented targets, and then prove that what we’ve done meets those targets. And there’s no network analysis software whatsoever in place, and he won’t sanction any expenditure.

I could go on, but it’s the weekend – who needs it?

and I’m on a contract with a one-month rollover . . .