wumbernang

Realistically everything is mostly shitty architecture and nearly on fire all the time.

The best critique is surviving production and constant iteration and tackling limits and problems as they become apparent.

If you see a problem, research solutions and hit the cycle again. Sometimes things don't work.

There's not much science in it I'm afraid.

The mid-range HP stuff is pretty reliable. We had a DL380p Gen7 survive the switch underneath it catching fire. Had zero chassis failures on about 500 nodes in the last 12 years as well. Lose disks and power supplies all the time and the odd Ethernet interface but nothing else.

Agree with ebay. I still look around for Sun Ultra kit now and then but the wife has other ideas because it's noisy and expensive to run.

Sounds reasonable but three points:

1. It's not anonymous. The ledger is enough to identify someone.

2. It's not 100% better than my current spreadsheet (one tab a month for the coming 6 months).

3. There's no motivation not to close it after a month I.e. no contract

No no no no. This is not right.

1. If you debug with this, you're debugging different code to what blew up in production for example. Subtle timing issues wiped out instantly and memory ballooning masked etc.

2. If you leave it in for prod, it's going to have masses of call overhead.

You can debug well with assertions, logging and unit tests. There is no need for this. I rarely have to spin up a debugger these days.

Interesting - thanks for posting.

He's waiting for a surgeon to have a meeting about him now. Not sure what the state of things is TBH.

It isn't happening until it has happened.

Also sure a compiler may drop but its useless without a pile of libs.

We had 11 years of uptime on a VAX cluster at a company I worked for in the late 1990s.

They took it down in 2001 to replace it with something that took up 2U of rack space, about 2kw less power and ran windows 2000.

I turned up in 2012 to replace it again with something cloudy and it had 11 years of uptime (well done NT!) again[1] so YMMV.

The cloud based version has gone down about 10 times (thanks Azure!).

[1] not a great position but this was on an isolated network with locked down everything so less of a problem than a normally networked system.

Transferred to another hospital overnight (this time a competent one). CT again. Found a small bleed and are currently waiting for a surgical opinion. He's in a shitty mood unsurprisingly!

Thanks for the uplift; appreciated.

Indeed. Thanks for the kind words.

Glad your friend got through it. Terribly traumatic but a good outcome looking at your other post. Happy to know someone who got through it against the odds.

Comedy timing for me.

My father was taken into hospital 4 hours ago after collapsing and turns out he's got a brain bleed. They stuck him in a CT straight away and are now doing a lumbar puncture to see if there is any blood in the spinal fluid. They don't know what has precisely happened but they suspect an aneurysm that went undetected. The mortality rate of this event is 50% in 30 days. He's had a long history of hypertension and a couple of surgeries to clean out arteries in his neck.

Not sure why I'm sitting here on HN but it was taking my mind off things. Fail :(

Anyway, moral of the story: Don't smoke and dint eat piles of shit; it'll get you one day.

Edit: Ordered the book as well now like an idiot. Scary tale when you're close to it but I find comfort in knowing things rather than ignorance.

Thanks for posting this.

Ah yes. A guy I know designed a simple voltmeter based on a uC and a 3 digit LED display for some radio equipment. Shipped the design out to china.

Look on ebay or aliexpress now...

Boy is he pissed about that.

Interestingly though the product that contained this ended up being built in the UK for only about 20% more. Shafted on import costs of components though.

I'd be surprised if they did and don't forget that it's serialized/deserialized from something which will be hanging around in the GC in the form of a memory backed stream or something too.

Basically, when an object goes out of scope, it isn't de-allocated instantly.

Immutable strings aren't standard; they're an implementation choice.

You're right and I glossed over most of it but the killers are concurrency, substitution and fixed deliverables. With concurrency, if you can prove that during the contracted period you had billable work outside the contract then that goes in your favour if you get nabbed and the contract is borderline. With substitution, the contract should state that you can provide a substitute on demand so you're contracting out services and yourself. The third is fixed deliverables; that's always a killer. Rolling contracts as non distinct labour are how to get shot.

For ref I had a small battle on this front and concurrency won it for me.

TBH get a good solicitor and get them to take a look at the contract paperwork. That's better advice :)