Forgot your password?
typodupeerror

Comment: Re:Cloud needs server huggers (Score 1) 409

by Vanders (#47013021) Attached to: Don't Be a Server Hugger! (Video)

Should nobody be hugging THOSE servers either?

As a former cloud administrator: no. When you have 2000 physical servers, why do you care that 50 of them are currently broken? Why would I care that the hard drive failed in one and I had to re-install it (with an identical image and configuration to the other 1999 servers)

Hell, we had servers that never worked from the day they were delivered and no one gave a shit: it went on the backlog for the DC guys to diagnose and RMA. Some of them got fixed after 6 months.

Comment: Re:Feels Dated (Score 5, Interesting) 435

by Vanders (#46879675) Attached to: C++ and the STL 12 Years Later: What Do You Think Now?
I'm one of these irritating DevOps types.

The Dev side of me loves Ruby. It's a nice language, it's powerful, the standard library is nicely complete and there are Gems for pretty much everything I could ever need.

The Ops side of me hates Ruby. Managing all those Gems on any given server is just horrible, rbenv & rvm need to die in a fire, there are a apparently one hundred different ways to run an application and proxy requests to it, and of courses Gems exist outside of the system package manager and that's always bad.

Comment: Re:Someone doesn't understand devops. (Score 1) 226

by Vanders (#46765635) Attached to: How 'DevOps' Is Killing the Developer

While I agree that some developers are cavalier with rules, consideration of resources is fundamental to writing software

There have been a number of occasions where I've had to say things like "No, you can't have 10 VMWare instances with 1TB disk and 140GB of RAM each. Because the VMWare cluster doesn't have the resources available, that's why." and "If you'd asked, you'd already know we don't have 2 DL380's with 192GB of RAM and 4TB of RAID1 disk in each datacenter. No I know you 'need' it, but it doesn't exist."

Usually the conversation then has to diverge into an overview of the concept of capacity planning and horizontal scalability.

Thankfully those kinds of conversations are rare these days.

Comment: Re:whine (Score 2) 226

by Vanders (#46765477) Attached to: How 'DevOps' Is Killing the Developer

Developers don't know how to run a production environment.

Yes. That's the problem that DevOps attempts to solve. You're supposed to have both "Developers who do Ops" and "Ops guys who develop" in one team to do "DevOps".

If you're working in a place that's done "We'll just get the developers to do Operations" then they're doing it wrong.

Comment: Re:That's funny (Score 2) 149

by Vanders (#46665547) Attached to: TCP/IP Might Have Been Secure From the Start If Not For the NSA

The packet-switching technology was military in origin - they were seeking a new form of communication network that could continue to operate without downtime in the face of massive physical damage, like cities being nuked. Academia soon adopted the technology, and the early internet culture came from there.

No. Wrong. Stop perpetuating this myth. Please, go read Where Wizards Stay Up Late

The vague concept of packet switching was developed simultaneously both by a British Post Office engineer (which is where we get the term Packet Switching) and a RAND researcher (which is where we get this ridiculous myth). However at no point did ARPA care about building the network to survive a nuclear war; it just happened that packet switching was a good way to make maximum use of the AT&T provided switched circuits that created the backbone.

Comment: Re:Thanks for peptuating (Score 2) 164

by Vanders (#46588667) Attached to: Peter Molyneux: Working For Microsoft Is Like Taking Antidepressants
I took anti-depressants for three years. My first course was Fluoxetine (Prozac to my US cousins): within a week I was a zombie. I would sit and stare at walls, or out of windows, until someone snapped me out of it.

After a month of this I was moved to Citalopram. This seemed better; there was less staring at walls, certainly. I spent over two years on Citalopram.

Then one day I stopped. It was kind of an accident; it was Easter weekend, I wasn't paying attention and ran out without a prescription to get more. So I ended up going cold turkey, which is the thing you're really not supposed to do with any SSRI.

I can tell you now, within three days I felt like I had woken up from a trance. I didn't realise it at the time, but the Citalopram made me feel like I was wrapped in cotton wool and wearing ray-bans. The feeling was exactly like the feeling of a dental aesthetic wearing off, except all over. I hadn't noticed because I'd come from Fluoxetine, which was even worse. So I thought I was onto a good thing with the Citalopram.

So please don't go around calling it a "myth". For some people, SSRI's really do have that kind of effect. In fact I suspect it's more pre-valiant than people realise, either because of long term use, because people think it's "normal", or because most people come down gradually and never really notice.

Oh and for all that, I really do believe I was much better off taking the SSRI's at the time than I would have been without them.

Comment: Re:Good if they succeed. (Score 1) 132

by Vanders (#46413901) Attached to: Oregon Withholding $25.6M From Oracle Over Health Website Woes
Why not? Do you think Oracle paid their developers and managers in promises and unicorn farts? Why couldn't Oregon simply have paid for those developers & managers themselves?

The point of hiring contractors is that they're supposed to bring instant expertise to a project. If they don't actually do that, why bother with the extra expense of the middle man?

"Engineering without management is art." -- Jeff Johnson

Working...