Re: My Vintage Dream PC

You are overpaying for your serial interface boards.

Electric stove? Probably full of SCR phase controllers, which not only make huge spikes, they modulate the RF reflectivity of your house wiring.

But it's almost impossible to cook with electricity. Gas rocks.

John

Yikes, only a programmer could be wrong on that sort of scale.

John

One shuttle launch was scrubbed because none of the three guidance computers agreed. They'd messed up the time-sync thing.

John

For certain very, very weird definitions of "bug."

When manufacturing needs it to build something.

Now, you may ask, how does manufacturing know that they need to build something? When they get a purchase order.

John

Exactly. And when times are tough, like for instance now, those customers send whatever business they have to the people that give them the best stuff and the least grief over the years.

Bugs = grief.

John

They may be the real Manchurian Candidates.

John

Anything is write-once if you only write it once. We could use hard drives, and physically archive them, and never re-write them. At around $80 for a 250G drive, one a week, that's $4000 a year, not absurd for company backup. But the DVDs cost more like $2 a week.

Claiming? I have a BSEE and I do engineering.

Go play with a few DVMs and learn something.

How can I tell that a file is missing or incorrect? Suppose accounting does some absurd transaction, or just closes out the month wrong, or the accounting software messes up. We might not know for weeks or months. By that time, all your rotating backups have bad files, and the good month-old baseline would be forever lost.

If no files were ever lost or corrupted, we wouldn't need to do backups.

If they were

Errors can't "propagate" onto write-once DVDs that are in baggies in The Cave.

I'm an engineer. If something is affordable and works, my time is better spent designing new products.

Any application or any person can screw up or delete a file. RAID doesn't help that... it just makes a bad file less likely to get lost.

We have a simple backup system that works and is bulletproof. You don't like it because you think you know everything and you're AlwaysWrong.

John

And what if a current file has become damaged, or changed such as to become wrong? In a rolling backup scheme, the old, correct files are lost forever.

Losing company history is insane.

John

We have a networked company file server (another identical HP box!) that's full of RAID drives. But if I want everything backed up, it still needs to be stored somewhere.

John

Then give yourself about two milliseconds to calculate the odds of that ever happening, dismiss the dumb twit's premise, and move on. Have a nice day, dumb twit.

You really do not understand the difference between a corrupt file on a file system, and a file that you or another one of your retard crew filled with bad info, and then wrote (saved). That FILE will NOT be corrupt, dumbass. The contents of the file are the USER's responsibility. Any errant data in that file is YOUR BABY. The file itself will read, open, and save just fine from the file system's POV. I cannot believe that you are so thick that you do not understand what you are backing up for, or what it is capable of.

The backup is a tool against hard drive failure, not retard crew failure, dumbshit.

Sorry, dumbass. I know how hard it is for you to follow and keep up with modern times. Are you still running Windows 95? I would not be surprised.

A single, 1.5TB drive can hold 50 of your 30GB daily backups. 50 times your $2 DVD is 100 dollars, which is not far off the reliable hard drive price, idiot. That is uncompressed. It would hold far more with compression.

Hell I could beat the system you have with the DOS based Zip utility.

So seven drives would hold 350 days worth of backup before compression alone. A five drive five day week schema would be even cheaper.

The reason you put them on days is so you can zoom in on any given day if you are hunting a file. It makes it that much easier than a chrono index on each drive.

You are at least ten years behind this mark, and miles below me, Johnny.

You gotta get in to get out.

Idiot! The "NAS chassis" IS a "Network Attached Storage" solution, dingledorf! The NAS device IS the backup device.

If it's wrong, I define it to be corrupt. And I need to recover the correct version.

The contents of the file are the USER's

I can't believe you want to lose correct files and over-write all the backups with wrong files, whatever the source of the wrongless.

You are playing with words, and I am protecting valuable data.

A PCB layout or a program source file may be worth many tens of kilobucks. I don't want to lose that for any reason.

And Windows sometimes trashes the innards of files without telling you that it did it.

John

You just number the things by hand. In real VCS systems the whole release contents can be tied to one configuration specification. When the product is 10+M lines of code, asics, assps, tens of FPGAs etc. the releases consist of quite many variables. And then expand that to a installed base of tens of years.

How about customer specific releases etc. ;)

The VCS runs in current platforms, and needs to be maintained. So VCS must be up to date, and migrated to new systems if needed etc. It is not some static system that is once built.

The version of the system must be unique and tracaeable to full set of source (and tools). Getting 10+ year old tools to run might be fun tough ;)

VCS management people, whose job is to keep it healthy and running.

The is no bug free design, if it is not trivial very small design. Even HW designs that have been manufactured for years in volume might have bugs that are uncovered by some third party component process change etc. For an asic 1 bug/10000 lines of code is quite normal bug density during its lifetime, for software the density is higer.

When there are thousands of changes the readme does not work. Release notes are just highlight of changes. I think we are approaching this from different directions, I'm thinking of designs that have hundreds of designers (hw+sw+testing etc.) you are coming from the small end.

--Kim

Odds are high. Example: system is brought up with the incorrect date/time.

/BAH

Bullshit, you lying, retarded bastard.

You are about as lost as it gets.