Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?

Comment Re: Issue is more complicated (Score 1) 922

You in any industry that cannot be discussed by me in polite terms. But you know that, and I'm not ready quite yet to tell you to find respectable or honest work. It's mostly not illegal, and of not proven illegal then I will not further condemn you.

I don't doubt it's an unforgiving place. The price of failure is tangible.

Comment Re: Simpleish (Score 1) 249

Don't get me started on turning Mag Card I/O into terminals. The selectors and solenoids were fairly straightforward, and the transmit was sufficient for keyboard data.

The 5218, I think that's the model, was a Selectric Printer for the DisplayWriter. Letter quality, reliable, not too noisy. The 5219(?) Daisy printer could not be repurposed to print from PCs , but you could run MS-DOS 2.11 on the DisplayWriter. Just needed the 8" floppies.

Many Selectric models were controlled electrically, Composers and Executives, Mag Tape (MT/ST) and Mag Card (MC/ST), several system printers for old mini systems, and they used a variety of interfaces interfaces.

The Selectric selected the character on the by its tilt and rotate position. Capital and shifted characters were on reach hemisphere of the ball. Impression was accomplished by driving the entire element carrier on a pivot towards the paper and platen, lifting the ribbon in so that it was pressed against the paper. Correcting, when implemented, lifted an adhesive tape in place of the ribbon, lifting the not-yet fully adhered Robin material from the paper. Functions such as escapement (moving the carrier to the right {or left for Hebrew, Aramaic?, etc}), paper feed or indexing, carrier return, tabulation, etc. Were handled by dedicated mechanisms, and were v electrified to be actuated by solenoids. Reed switches etc would signal the state of the machine. Printing was driven by the operational shaft, a half turn per character.

Selectrics are indeed remarkable machines. Frustrating to service, but worth the effort.

The Electronic 50, 60, 75 are advanced Selectric IIIs also electrified. Faster and not as reliable, the electronics advanced features. Before that, the Memory 50 and 100 used a tape loop to store data, functioning much like MC/ST. After the Electronics, it was WheelWriters. Not nearly as mechanical. Soon after this, PCs and WordPerfect took over.

People are still repairing Selectrics, and still converting the electrified ones for various uses. Certainly a good substitute for an ASR/KSR-33, the original single-element printer. I can refer you to some Selectric groups.

Comment Simpleish (Score 1) 249

0. Many old IDE drives failed when the stepper drivers got flaky and when hot would crash the heads. I put them on long cables, stuffed them in a freezer compartment, and they would usually live long enough to be backed up. Seagates did this some, but Maxtors were the worst.

1. When I was asked to install a 750MB drive in an old Novell 2.15c server, it took some thinking to figure out how many disk buffers would be needed to access the drive reliably. The customer asked me to leave the DCB attached, with both 20MB drives still spinning and serving data. Doing this without powering down the DCB and drives? Priceless.

2. Going back to the same server and replacing the thinnet NIC with a gig Ethernet NIC a year later. If you configure enough packet buffers, it works... We used a number Novell had never tried.

3. A few years after this, the DCB is still running, and they call me back to install a pair of 320GB drives. More buffers. Add-on zero slot SCSI RAID controller in RAID 1 mode. Linked the driver as required, the manufacturer did this hack and wrote the driver floppy for me, as NetWare 2.15c was EOL'd at least 10 years before this card was produced. They get credit for that hack. Keeping the DCB drive spinning so they don't stop and stick? Priceless. Figuring out the LBC-CHS mapping that allowed the server use all the space? Scary. It was non intuitive, but fixed thanks to a friend who does octal math in his head to 8 places. He's weird.

4. When GroupWise 4.x wasn't quite patched up, you had some Korean jerk reflecting Yahoo addresses off it as SMTP postmaster error replies to spam the world with Yahoo Mail addresses. The server I serviced could send 200-250 million in a weekend until the disk filled with the errors. Fix was to set the MX record for the customer to a server at their ISP, teach GroupWise it's SMTP gateway was that machine, and let it properly refuse the incoming mail. GW was patched a while later for this and the big security hold that this was not part of. That ran for about 8 years. Finally Exchange worked well enough for that customer to switch.

5. There was the Novell era where we did so many weird hacks to overcome corrupted volumes, funky network routing, and Novell's figuring out the IDE driver was causing the clock to lose time, forcing us to install NTP and eventually the whole NAMP stack. Apache on Novell was not my hack.

The Selectric stuff wouldn't interest anyone here.

Comment Re:Issue is more complicated (Score 1) 922

Actually, cut to the chase.

If the Kernel team's goal is the best work they can do, is abusive behavior helping achieve that goal? If a locker-room mentality works, so be it, but you will narrow your pool of available talent. Will that get you the best work you can get?

If the Kernel team is not interested in doing the best work they can do, then it's up to their leader, Linus from the sounds of it, to decide if that's the way he wants it to be.

And again, it's Linus' world there. Play or not, yours and his choice. Results will prove out.

Uncertain fortune is thoroughly mastered by the equity of the calculation. - Blaise Pascal