« first day (2012 days earlier)      last day (3008 days later) » 

2:00 PM
@Bob apparently some universities are starting to treat employees and students the same in terms of the IT systems they deploy, sharing common email and other such systems, so while it may make sense to have remote wipe for employees, it's nonsense to do that for students.
 
Bob
Windows Phone is more likely to get it first, tbh. If it doesn't already have it.
 
those systems need role-based, configurable levels of security... although it wouldn't surprise me if they already have that but the hapless sysadmins just left everything enabled for everyone instead of customizing the amount of control it needs for students
 
@allquixotic Yup
Some staff even refuse to use it because of the remote killswitch, despite management promising "We won't use it unless you tell us to, honest!".
Most just use IMAP as a workaround and lose the calendar functionality.
 
I bet the management would seriously consider using the killswitch if they decided to terminate an employee involuntarily when they happened not to be immediately available (out to lunch, etc) for fear that they'll get wind of it and do something malignant before they lose access
 
Bob
That's why it all needs to be laid out in a written contract.
 
2:05 PM
Cutting out their AD account is supposed to immediately cut off their access.
If they have to kill the phone it's connecting with, I'd worry about their access via other platforms (e.g. desktop, doors, etc.)
 
does it support offline mode of any sort?
 
Bob
doors lol
 
doors? O_O
 
Bob
@allquixotic why would it matter?
they could just as well have printed their emails
that's an "offline mode"
 
I've heard "doors" referenced as a software application at every company I've ever worked for, but I've never seen it, never heard a description of what it does
also I'm pretty sure there are multiple different "DOORS" applications at different places I've worked
 
Bob
2:08 PM
huh, I thought he meant literal, physical doors
as in, walk in the door and start trashing things
 
oh lol
well most places of employ have an out of shape person with a gun that's probably not loaded (and even if it were, they might not know how to fire it without missing) to threaten people to not trash things, when they tell them while they're physically there that they're being terminated
 
Bob
they do?
is that an american thing?
 
possibly ;p
 
Bob
having an actual physical guard is rare here, to say the least
you'd have a receptionist. and locked (carded) doors. that's about it.
 
I imagine universities might summon the campus police if you have a staff member refusing to leave or trashing places, but at corporations, I've always seen at least one, often two guards at a desk
 
2:11 PM
Same. We have occasional security patrols around campus but that's it.
 
for some of the higher security places I've been, they definitely know how to use their guns, and they definitely are loaded
 
University campuses here tend to be fairly open anyhow.
 
but the typical low-sec folks are glorified receptionists
 
Bob
yea, you have uni security, but I've not seen guards at the door
 
Most of the private security folk aren't even allowed to touch customers/students in this country, and certainly aren't allowed guns.
 
Bob
2:12 PM
generally if a physical altercation starts, the police gets called
 
just think of it this way: American receptionists have guns, even if they seem to be dressed normally.... just assume every American receptionist you meet is packing heat
 
Bob
if a gun is involved, someone already failed somewhere along the line. also, that's very unlikely here.
 
Nonetheless I know of people (e.g. me) who have had door (swipe access) revoked while on long periods of leave.
@allquixotic s/receptionist//g
 
lol
 
> just assume every American you meet is packing heat
FTFY ^
Mind you our place was relatively high-tec and I had relatively high access (e.g. 24/7 unsupervised swipe access to the datacentres)
So I'd just wait for security to go to sleep, or use the back door, or any of the side/loading access doors.
 
2:14 PM
s/high access/ability to cart off half a million pounds of HPC equipment
 
Bob
lol
 
@allquixotic That's only the beginning
Another million or so of run-of-the-mill servers, tape libraries, storage arrays, and 10GbE network core(s)
 
you sound like you work at Amazon
 
I believe the wireless network controllers cost about £70,000 each, discounted from £120,000 RRP
We had five of them.
 
Bob
"discounted"
that's when they slap any price tag they want on em
 
2:16 PM
I wonder what their production costs are for a unit that expensive
 
Well yeah, hence why they never list prices.
 
more than a few thousand, I'd think
 
Probably fairly low. Most of it's R&D and licensing costs.
 
Bob
production probably isn't that expensive.
more for R&D
but their pricing probably isn't even based on that
 
After all you can buy the same unit for £20,000 and then spend £100,000 on optional firmware upgrades.
 
Bob
2:17 PM
it's just 'what can the customer afford, what are they willing to pay, and what do our competitors charge'
 
well yeah, once you get to enterprise level it's all about not leaving any money on the table
 
Bob
> Manage users with Office 365 Admin on mobile
O_O
 
PHBs love to be able to fire people while riding comfortably on a train or airplane or while driving (the law says "texting while driving", not "firing while driving", officer!)
 
I had to Google PHB
!!google phb
Aww.
The first result was PHB Ethical Beauty
 
not sure what happened to cavil, lemme check
 
Bob
2:20 PM
> The address can be EX, X.500, X.400, MSMail, CcMail, Lotus Notes, NovellGroupWise, EUM Proxy address, and free text.
o.O
 
@allquixotic Bob said it died a while back and you hadn't had time to fix him yet
(Hence I thought you already knew)
 
I kinda did, but thought the auto restart script would bring him back
 
Bob
Oh, I thought you disabled it.
...why do I have a SIP address
 
> When he first appeared,[3] the Boss was very cruel and uncaring (shocking people with electric belts or wanting them to work 178 hours a week, although there are only 168 hours in a week — he expected the employees' families to contribute a few hours[4]). He showed few obvious signs of cluelessness. However, as the series continues, he became less malicious and more incompetent. He still show signs of this by referring to his department as his empire.[5]
> Receptionists in Stockholm, Sweden
 
node run-headless.js                                                                                             s
/home/sean/dev/SO-ChatBot/node_modules/nightmare/node_modules/phantom/node_modules/dnode/node_modules/weak/node_modules/bindings/bindings.js:8
3
        throw e
        ^

Error: Module version mismatch. Expected 47, got 14.
    at Error (native)
    at Object.Module._extensions..node (module.js:440:18)
    at Module.load (module.js:357:32)
    at Function.Module._load (module.js:314:12)
 
2:25 PM
> Error: Module version mismatch. Expected 47, got 14.
Sounds deceptively simple >_>
 
I have no idea how to fix that ... npm update / npm uninstall && npm install didn't fix it
 
@allquixotic What I managed to get out of that was... Your name is Sean?
 
@qasdfdsaq Hey! They have those fancy swivel-mounts for their monitors and the really swanky adjustable chairs that only the bosses get where I work!
 
@allquixotic Yeah, exactly. They're spoilt in Sweden.
 
what do the bosses get in Sweden? foot massages?
 
2:28 PM
Attractive blonde receptionists, obviously.
 
@qasdfdsaq Context?
 
Bob
@allquixotic /node_modules/nightmare/node_modules/phantom/node_modules/dnode/node_modules/we‌​ak/node_modules/bindings/bindings.js
That nesting...
 
The pointy-haired boss (often abbreviated to just PHB or "The Boss") is Dilbert's boss in the Dilbert comic strip. He is notable for his micromanagement, gross incompetence and obliviousness to his surroundings, yet somehow retains power in the workplace. In the Dilbert TV series, in which he is voiced by comedian Larry Miller, the character is notably smarter (although still quite stupid and inept) and more openly corrupt. He is also parodied in Bee Movie as Dean Buzzwell, also voiced by Larry Miller. Mr Perkins in Despicable Me is visually based on him. His motto is Anything I Don't Understand...
 
@Bob T/node_modules/h/node_modules/a/node_modules/t/node_modules/ /node_modules/n/node_modules/e/node_modules/s/node_modules/t/node_modules/i/node‌​_modules/n/node_modules/g/node_modules/./node_modules/./node_modules/.
 
@qasdfdsaq Oh, I see.
 
2:32 PM
> Dependencies and assumptions:
– Customer will provide suitable power and cooling to run the configured system
– Full access to be provided to the data-centre as required by installation team
No YOU provide the datacenter! That's what cloud is supposed to be!
 
@qasdfdsaq Please stop trying to out people.
 
@DavidPostill Hmm?
 
@allquixotic beaten with twigs.
 
> You need to npm remove ... and npm install .... Basically you need to rebuild the package that is causing the error in newer node version.
Are uninstall and remove are equivalent?
 
rm -rf node_modules worked nicely as an alternative to remove
 
2:39 PM
Just set up my new work desktop and this chat looks sad with the default theme :'(
 
works now
 
@OliverSalzburg There are... themes? o_0
I've always thought this was all there was
 
Or @rlemon's more extensive extension: github.com/rlemon/se-chat-dark-theme-plus
Heheh, "extensive extension"
 
Aww what... your "Old woman honks at the cat" message isn't in search history :-(
 
Yuck. I really hate dark themes. They should burn in hell! ;)
 
2:42 PM
._.
 
Bob
"Microsoft Outlook has stopped working"
... -_-
 
The dark theme could sit quite well next to my 4th monitor, which is usually off, hence black.
 
Bob
Oh. Adobe Contribute plugin. bah.
 
Which reminds me to install f.lux!
 
Dark theme also could work well on my OLED display, but that's thinking too far ahead
 
2:44 PM
A study by Dr. Lauren Scharff and student Alyson Hill (more information) of Stephen F. Austin State University shows the expected results – black text on a white background is the best rated combination for readability:

…the most readable color combination is black text on white background; overall, there is a stronger preference for any combination containing black. The two least readable combinations were red on green and fuchsia on blue. White on blue and red on yellow were ranked fairly high, while green on yellow and white on fuchsia were ranked fairly low. All others fell somewhere b
 
I don't use dark themes under the assumption that they improve readability
 
@DavidPostill Sure, an 18 year old study before the advent of modern displays, browsers, and font rendering technologies.
 
There should be adaptive themes that change depending on ambient light!
 
@DavidPostill I find dark themes more readable
 
I mean just look at that text aliasing...
Also note that while white-on-black was lower than black-on-white, it was still higher than almost every other combination.
 
2:49 PM
Personally I like some choice ;p
 
Also seems they didn't test grey-on-white, which is what a good proportion of "black-on-white" web actually is
Also seems they didn't test any other colours on a black background other than white or grey.
@JourneymanGeek: Kung Hei Fat Choy
Is the Cantonese equivalent.
 
As spoken by an actual Chinese speaking friend of mine on Facebook.
 
The main dialect here's hokkien
 
Seems you were right all along and whatever my parents taught me was a load of rubbish.
 
2:59 PM
;p
buuuut
the government basically discourages dialect use
 
Bob
ok, lol
I've got simulated subaddressing, via a hackish catchall
 
low (same?) contrast Solarized is really kind on the eyes. There was a point that I had nearly all text editors and terminal-based stuff (bash dirs, tmux, midnight commander, et al) all solarized, but it gets into a confusing mess, as I tried adjusting the terminal colours so that everything still was coloured how I would expect in non-customised servers, which I have to remote into all the time.

I intend to go back and try it all again, but methodically this time.
 
Oh god not Sunday
Thanks Radio 1.
 
3:16 PM
@paradroid I love solarized
 
@allquixotic Yeah, it's great when everything is set up how you like it, but can get messy and confusing.
 
Solar palettes look real l33t in PuTTY/SSH windows
 
That's all I ever used it on, plus text editors
GUI ones I mean
well, there's nothing else you can use it on really
 
See those are what I'd think of as price ranges for "High end, midrange, and low-end" laptops
 
Bob
Yay! Email works, with receive catchall and send as distribution groups.
Convoluted, but it works.
...maybe I should've gone with GApps
 
3:31 PM
lol
 
Does anyone know whether "Antimalware Service Executable" is a part of Windows Defender?
 
All this hassle just to get... email?
@RahulBasu Yes, I do know.
And yes it is.
 
Bob
@qasdfdsaq Email on my own domain, yea.
 
okay, thanks...
is it supposed to be using half a gig of RAM?
 
@Bob All your comments over the past few days are suddenly beginning to make sense now :-P
 
Bob
3:33 PM
-_-
 
haha
 
Bob
I've always had the namecheap forwarder but never bothered to set up proper outgoing email
Been planning to DIY it for a while but then I figured I didn't want to bother with maintaining an email server
 
@RahulBasu Uhh... it varies, but probably not?
Mine's using about 200MB on Windows 7 and 95MB on Windows 8.1
 
@qasdfdsaq I'm on win 10 and currently, it has gone down to 426 MB from 513MB... :(
 
@RahulBasu Restart the service?
Interesting, AMD Carrizo APUs have cut system idle power down to ~2.7w. Suddenly seem a lot more worth considering on a laptop
 
Bob
3:40 PM
?!?!
Outlook 2016 can only send as a distribution group
Outlook on Android can send as an alias
wtf.
Microsoft, get your shit together.
 
lol
@qasdfdsaq now its on 80MB
thanks...
 
Heh. Nice.
TBH I probably wouldn't notice if it did that on my home system. I haven't checked my RAM usage since I built it.
I occasionally have Aquasuite eat up >3GB of RAM and crash but even that doesn't affect the overall system in any noticeable way.
 
Bob
3:55 PM
Ok, so that's email off my looooong todo list. What's next...
Install Win10. Ohhhh boy.
actually I need to set up some rules on my gmail account too :\
 
Oh wow, AMD Carrizo is actually a huge improvement over Kaveri
10% more performance at 20% lower power
 
4:51 PM
So if they increase their TDP to 150W, they might get as much performance as Ivy Bridge's 65W? ;)
 
That's why I'm worried the 40% increase in IPC mooted for Zen won't be enough.
 
> Hi Sean,

Thank you for taking the time to share this information with us!

Can you do me one more favor and pick a shirt from the Avenger Storage Facility? We've just restocked with loot, but don't worry it's all been cleaned. We're pretty sure dish soap gets out alien blood?
2K is not responsible for any allergic reactions upon encountering Sectoid fluids

I'll just need to know the shirt you choose with the size and shipping address, and I'll get that out to you.

All good things,

Crystal P.
I'm getting a T-shirt for submitting info about how I resolved my problem with my USB RNDIS connection
 
@allquixotic 2K games and USB RNDIS? o_0
 
@qasdfdsaq Yes. TL;DR: There are two versions of the Microsoft USB RNDIS driver: One written for the XP / XP64 codebase using NDIS 5.x, and one written for Vista+ using NDIS version 6.x. Either the OS actively ships both versions, or some third-party software package dumps the 5.x version on you regardless of what version of Windows you're running.
Some API somewhere within the networking stack of Windows that does something to do with broadcasting or network discovery (it's not LLDP; I already tried disabling that) totally fucks up and blocks the main thread of whatever application calls it when you're using the NDIS 5.x driver.
XCOM 2, the Network Connections control panel applet, and anything else that uses this API will be totally unusable with this defect until you force Windows to use the NDIS 6.x driver.
 
5:02 PM
Oh interesting
 
Whatever application or driver installs the 5.x version (probably something to do with Android, maybe Samsung's Android drivers, just a guess) makes it the primary / default driver.
So when you plug in a USB RNDIS device it auto-loads the 5.x version
 
I need to keep that in mind next time I'm debugging network hangs on my machine.
There's a few obnoxious hangs I previously had under W7 that basically drove me into upgrading to Win 8.1. I wonder if my huge pile of legacy drivers (Including, for certain, several NDIS 5 and USB network drivers) had something to do with it.
 
@qasdfdsaq So here's what happened: Microsoft has continued to update the NDIS internals within the kernel core and low-level subsystems of the OS, but out of fear of breaking devices that don't have active driver updates from the OEM, they retain backwards compat with horrible old versions of NDIS drivers.
 
Is this blocking thing universal to all NDIS 5 drivers?
 
The backwards compat is not perfect.
@qasdfdsaq I wouldn't know. It may be. It's worth testing, if you have a physical network device of any sort that has old drivers. You can look in the INF to see if it's NDIS 5.
 
5:04 PM
Mine even say (NDIS 5) in the device manager description.
 
ahh
 
Well currently I think I'm using NDIS 6.2 drivers. But IIRC I had to use the 5.x ones to get teaming to work on my onboard NICs on my old motherboard.
 
Yeah. If you recall, Microsoft has declared consistent Kernel ABI "compatibility" back from the ancient old days all the way up to Windows 10. However, that compatibility has huge limitations in areas like graphics, networking, and printing.
That means you can technically load a .sys file from Windows NT 4.0 into Windows 10 32-bit, but it probably won't work in practice.
XP 64 drivers are not too likely to work on Windows 10 64-bit, either.
With Windows 10, you can't use any driver that doesn't support at least WDDM 1.0, because there's no such thing as running Windows 10 in non-DWM mode. If you have a desktop, then it is composited.
 
They mostly don't. Except for maybe printers.
Though as far as I recall XP-64 drivers were mostly Server 2003 64-bit drivers.
 
Well, yeah. XP 64 was the client implementation of 64-bit Server 2003. Same codebase.
 
5:08 PM
And Server 2003 64-bit being a "fork" of NT 5.1 that never went anywhere
Whereas Windows Vista and Windows 7 (NT 6/6.1) were continuations of the 32-bit NT 5.1 XP codebase
 
Vista was probably the most significant OS overhaul Windows has received since the introduction of the NT kernel itself.
Mandatory Integrity Control, the DWM, WDDM, NDIS 6, Session isolation, ...
 
Although I could be wrong, since XP-64 enforced driver signing like Vista, whereas XP-32 didn't
 
A lot of people hated Vista, but then again, it was a fairly rough implementation of ideas that formed the core of Windows ever since, and haven't drastically changed, other than the UI appearance layer and some enhancements to WDDM to accommodate new GPU architectures, and very incremental NDIS improvements.
It's almost like, if an OS has revolutionary engineering, then its UI tends to suck.
KDE 4.0. Revolutionary engineering, horrible UI. Took multiple releases for it to be usable.
 
I liked Vista because it had improved remote desktop, and the block I/O layer was massively improved for my RAID arrays.
(Incidentally that also paved the way for SSD compatibility later on, which few people realise)
 
I liked it too, even its UI. I was using Vista on my workstation at work (mandated by gov't) as recently as 2012.
Gov't actually moved from XP to Vista to 7, unlike some corps that skipped Vista entirely.
 
5:12 PM
Half the problem were OEMs building cheap crappy PCs with poor hardware and consumers buying cheap crappy PCs with poor hardware.
 
Yeah... you don't want to run DWM on an Intel 965GM or other such trash IGPs
 
Vista on a 64-bit dual-core desktop with 8GB RAM = Fine
Vista on a 32-bit first-generation Atom with 512MB RAM = Kill me now.
And yeah, my dGPU probably helped too
 
Rumor has it that one of my clients (big org.) is moving their client systems to Windows 10 and has already moved much of their Windows-based infrastructure to Windows Server 2012 R2 and the latest SQL Server and AD... I'm a little surprised at how rapidly they're moving on that
For comparison, company I work for is sticking with Windows 7 until more-or-less the end of security updates for it
(They did the same thing with XP)
 
Our organisation is doing limited rollouts of Windows 10 on the desktop with a view to full deployment later in the year.
Basically they're pushing all I.T. staff to run W10 to iron out the kinks. We skilled W8/8.1
 
yeah, skipping Windows 8 / 8.1 seems to be popular...
 
5:16 PM
As was skipping Vista in the enterprise.
 
there was brief talk about them evaluating W8/8.1, but that fell by the wayside immediately when W10 release date was announced, I think
 
Most UK universities (including all the ones I've worked for or with) just went XP => 7 => 10
 
my company went XP => 7 => (eventually) whatever the latest is when 7 goes out of support
Client went XP => Vista => 7 => 10 in calendar year 2016
 
Some companies just like being on the latest I guess
 
some only upgrade when Microsoft all but forces them to ;p
 
5:17 PM
I'd suspect game developers in particular want to be developing and testing on the latest platform at all times
 
developing and testing are two very different things -- they might develop on Windows 10 with beefy rigs for expedience, but their testing organization really should be working off a very diverse set of machines with different OSes
unless you're developing a super high-end game that you don't care if it falls down on even mid-grade machines (Star Citizen, I'm looking at you), you should be testing on fairly crusty old OSes and machines for a playable experience
 
Lol
Few games scale as well, apparently GTA V is one of the better ones.
 
even high-end AAAs like XCOM 2 should be testing on Windows 7 (Vista...?!?!), particularly since they ship both 32-bit and 64-bit binaries.
 
But if you want to take advantage of the latest DirectX then some developers are too cheap to build both a DX10/11 and a DX9 codepath
 
though the hardware they use (chiefly of interest, the GPU) is up to the engine folks and how aggressively they push texture quality, etc
 
5:20 PM
Sure, bigger companies would have the resources to basically build two game engines, but smaller ones don't.
Incidentally I wonder when KSP 1.1 will be out (Moving to Unity 5 engine)
 
Well... I don't think it's unreasonable to expect users to be running an OS newer than XP, no matter what kind of game you're developing. Vista+ is pretty much a given. For indie / casual, you definitely want to support 32-bit, DX9, and IGPs from 2010 and later.
For simulation, you want to have fairly modest base requirements but perhaps all but Iris Pro IGPs are out of the picture.
 
You'd be surprised what modern IGPs are capable of
 
For AAA FPS and Star Citizen (it's in a category of its own, really), you basically mandate a dGPU from the past 4 years. Low detail if it's a low-end part, but should still get reasonable FPS.
For MMOs you drop that back to supporting dGPUs from 6-7 years ago, and new IGPs.
 
The 6w IGP on my 15w Skylake-U CPU runs GTA V, Star Trek Online, Battlefield 4, KSP, and just about everything else I've thrown at it
 
telegram's api is pretty nice for building bots.
 
5:22 PM
@qasdfdsaq what FPS and quality levels, though?
 
been playing around with it the last few hours.
 
Star Citizen is really the modern reincarnation of Crysis
AKA build something that no current hardware can run on maximum.
I kinda wished all games still used similar development models, instead of pandering to the lowest common denominator/console. I miss having to upgrade my PC to play the latest game on maximum
 
@qasdfdsaq Yeah. I think Star Citizen is going to require minimum HD7970 to run playably at 1080p on low detail. For awesomeness you'll want Nvidia Pascal's top-end part. For VR or 2K/4K or multi-monitor you'll want four Pascal top-end parts.
 
@allquixotic Mostly medium/low at 30fps
Eve Online, Call of Duty 6, Team Fortress 2, and Star Trek Online all ran on my 2012 37w Sandy Bridge IGP too, but pretty poorly.
 
The benchmarks of even the the latest Skylake "Iris" part (no Iris Pro 580 benchmarks out, unfortunately) are pretty terrible.
It's basically a GeForce 745.
 
5:25 PM
It's basically a Geforce 940M
Which incidentally is the dGPU I had in my new laptop that I returned, and similar to the Surface Book.
 
Wake me up when they are offering a 970M on a CPU and I'll pay attention :P (so will Nvidia too, I imagine)
 
Give it 6 months.
 
you think Kaby Lake is going to have that much better of a iGPU than Skylake?
 
Truth is the iGPU in a 15w Skylake-U part has 3x the performance of a iGPU in a 40w Sandy Bridge part.
And the 15w iGPU in a 28w Skylake + Iris part performs as well as a 35w NVidia dGPU
 
So when are they going to truly scale up? I'll take a 95W+ desktop CPU part with a CPU perf on par with the i7-3770K (or faster, if they can manage), and a hugely overblown IGP that competes with modern dGPUs. That would be awesome.
 
5:27 PM
AKA Intel's 14nm process improvement is giving double the performance per watt of NVidia's latest 28nm dGPU.
 
I'd also probably need liquid cooling, but eh.
 
So scaling that up, a 35/45w laptop part with a 30w iGPU on current/next gen processes (possibly in Zen) would be pretty decent.
 
I think the number I'm looking for at videocardbenchmark.net is 5000, before I start going :O wow, IGPs are rockin' fast
Right now Iris Pro 580 is probably going to hit under 3000.
5000 is a good ways slower than a HD7970, but definitely in what we consider "dGPU territory" today
 
I dunno what the figures on videocard benchmarks are given in
But the scores seem to be very similar to 3DMark FireStrike
 
Magic PassMark Units ™ '_
;)*
 
5:30 PM
Good afternoon :)
 
And yeah, 2000-3000 is roughly where the Iris Pro 580 will be at
Mind you we're getting into weird territory here.
In order to get the performance of a 100w dGPU, even with Intel's process advantage would be looking at a 50w iGPU.
And mainstream desktop CPUs have a 65w-all-in TDP with about 20-30w for the best Iris Pro iGPUs
If you want to shove a 50w iGPU in a chip you're looking at a 70w-100w CPU which has no place in a laptop.
 
Yeah, but if they build it, system integrators can start to sell "built-in gaming GPU, no dGPU needed!" to their customers.
Then Intel just needs a software development lifecycle comparable to Nvidia's Game-Ready Drivers :P
 
You're almost certainly looking at desktop-only CPUs in any case.
 
Seriously, though. Think about it. A new third competitor who's serious about the full-fat laptop and desktop gaming segments. What's a stronger motivator to get AMD and Nvidia to run faster on the treadmill? :D
 
Eh, it's hard enough to find a 12-13" laptop with a dGPU as it is, let alone a convertible one. There's absolutely nothing on the market that satisfies my needs in any combination.
I'd be happy enough having GTX 950M class performance regardless if it comes in a dGPU or an iGPU
 
5:37 PM
I'll have to make a youtube video with the Hitler "Downfall" scene (the popular one in the bunker with the generals) with the Nvidia CEO waving his hand going "Eh, IGPs aren't fast enough to compete with us yet." then "Mein Fuhrer... the Iris Pro 680 benchmarks faster than the GTX 960M."
 
Right now... they're pushing sub-920M class performance in iGPUs
 
"Everyone who's an Intel fanboy, get out."
2
 
As it stands, modern designs have trouble cooling 50w in a 14" chassis, especially a convertible one, and ones with over 30w TDP of power are already rare.
Which is weird seeing as my previous 13" convertible came with a 37w CPU and an optional dGPU.
Which is why I'm waiting for the next-gen products from AMD and NVidia.
Pascal/Polaris are basically going to double performance-per-watt.
And that basically means GTX 960M-level graphics becomes possible in a modern sub-14" convertible.
And that's getting close to Radeon 7850 performance.
Well OK, more like Radeon 7750
Which is a lot better than the Radeon 5750-class performance we get right now...
@allquixotic Those Magic Passmark Units ™ look a bit dodgy. Apparently the Fury X scores 15% lower than the GTX 980, whereas everywhere else puts the Fury X anywhere from roughly equal to 40% faster.
@allquixotic Also you have extremely high standards. The chips scoring over 5000 currently are 125-250w dedicated GPUs.
 
Love it!@
 
6:56 PM
Hi folks! FWIW, not too sure if posting a self-answered question like this is acceptable to the community, but I erred on the side of there being a consistent flow of hard drive failure requests here so perhaps the info presented is useful? If not, please let me know. And if it’s just not the type of thread this community wants, please feel free to delete the whole thing.
0
Q: If someone purchased a Seagate ST3000DM001 3TB hard drive and it failed sooner than expected, will Seagate replace the drive?

JakeGouldOne of Seagate’s current popular hard drive models is the ST3000DM001 3TB hard drive. It seems that many users have reported problems with the drives failing faster than expected causing data loss. What can be done by someone who purchased this drive to replace the drive?

 
@JakeGould Self-answered questions are accepted and encouraged. I've done it many times, and somehow I turned into a mod on another site!
As long as the information is there, and you would answer it as if it was someone else's question, I'd say it's good
 
Yeah, looks good to me
 
@JakeGould Oh dear. I think I have one of those. Should I be worried?
 
@DavidPostill Buy a Western Digital to get your data backed up, then talk to Seagate. I have no idea if that is worthwhile, but I never trusted Seagate when I worked at the computer shop selling both Seagates and Western Digitals.
 
!!Is the chatbot alive?
 
7:09 PM
Urgh.
 
It's a ST3000DM001-1E6166. Does that make any difference? BTW it is one of my 2 independent backup drives. All SMART values are OK.
 
"My computer bluescreens so here's a chkdsk log" is the computer equivalent of "The engine fault light came on so I went around and checked the tyre pressures"
 
@DavidPostill Read the details on my answer. The link is to a class action lawsuit and for all you know you might “win” the lottery of drive replacement with added $$$ for headaches.
 
@JakeGould Except I don't live in the US :/ I'm in the UK. Drive was bought in NL.
 
Kinda glad I got the WD Red rather than the Seagate now.
 
7:13 PM
@Mokubai Slightly higher than average failure rates for a particular, old model desktop drive is no reason to avoid a company for NAS drives forever
 
@qasdfdsaq True, and my last seagates (2 x 1TB disks) did last nearly 5 and a half years, but still... :P
 
I've got 24 Seagate drives sitting in a chassis behind me, including 19 NAS drives, and they're all working fine, despite being pushed well beyond their "rated workloads"
I've seen more WD drives fail than anyone else, though most Seagate "failures" I had are simply drives who don't like the environment they're placed in
Most Seagate SATA drives seem to have a strong dislike for any environment with shared power busses common in NAS, RAID, and hot-swap environments.
 
@qasdfdsaq So just most of the places where spinning rust is relevant these days?
 
@Mokubai Desktop drives designed to be used in desktop chassis... shock horror.
 
lol
 
7:19 PM
Their NAS and server range cope with it a lot better.
Although they still seem to have the same symptoms when pushed really far. It seems to be some kind of common design... strategy
 
@qasdfdsaq I actually curious why that would be the case, from an electronic standpoint... poorly specced power regulation? An earth loop which they worked around in other models?
 
Halfway torn between calling it a feature and a bug.
 
@qasdfdsaq That's good to hear. I will buy another drive anyway just to be safe.
 
Either way the drives I had lasted long enough and gave me a relatively gentle failure mode that gave me time to move everything I cared about. I've no real beef with Seagate.
 
@Mokubai I dunno. To some extent it could be a protection mechanism
i.e. some protection mechanism that locks the heads away in the case of a power dip or cutout is way more sensitive than everyone elses
 
7:24 PM
!!info
so, wtf...
 
While most drives have emergency shutdown/safing mechanisms, the Seagates seem to do it at the sign of the slightest fluctuation. And it's not logged properly or communicated to the host properly.
 
$ node run-headless.js
Need to authenticate
phantom stdout: TypeError: null is not an object (evaluating 'document.querySelector(selector).focus')

phantom stdout:   undefined:2
  :3

phantom stdout: TypeError: null is not an object (evaluating 'document.querySelector(selector).focus')

phantom stdout:   undefined:2
  :3

phantom stdout: TypeError: null is not an object (evaluating 'element.dispatchEvent')

phantom stdout:   undefined:5
  :6

Should be done loading stuff.
You are now in a REPL with the remote page. Have fun!
Fuck
 
7:38 PM
well because Netbeans' HG-based source control system sucks ass, I had to download 4.7 GB of data to my server in order to remove 3.7 GB of .hg directory commits in order to download 89.2 MB of tar.xz'ed source code (uncompressed about 1 GB)
Le sigh...
 
At least tell me you don't have a capped data plan...
 
on my server? I do, but it's an atrociously huge cap
locally? nah, unlimited Verizon LTE
 
Lol I couldn't find a nosecone I liked for my plane so I stuck a backwards jet engine on the front and set it to reverse thrust, so it propels it forward anyway
 
rofl. KSP?
 
Yup
Also helps moving my centre of mass forward, which I kinda need
 
7:56 PM
@DavidPostill Well, my thread at least can now serve as some place to share notes and info on experiences. It seems the issue affects drives bought around January 2012, so hey.
 
8:09 PM
@JakeGould Yes, I'm hoping I'm OK. Mine was bought Sep 2014
 
 
2 hours later…
9:49 PM
@JakeGould question is fine i suppose, though it talks about a warranty. That said, go dig up a HGST. Most reliable drives there are ;)
 
@JourneymanGeek Except for the Deathstar legacy...
Every manufacturer's had some bad drives at some point
 
@qasdfdsaq true but currently they have the best reputation of any of the brands there are
seagate has been shit for years
 
Aren't HGST part of WD now anyway?
 

« first day (2012 days earlier)      last day (3008 days later) »