816 posts • joined 11 Jan 2008
Re: Google Drones vs. Facebook Drones?
You won't be if they tumble to earth onto your head…
Re: Yahoo and DMARC
Well, I've read through all the comments above, but not one of them provided an actual solution (or even a helpful suggestion) to the problem I'm facing.
Well, about the only thing you might be able to do, is set up your list server to re-write the From address and other details so it looks as if your server is sending it rather than the other party.
I'm not sure what software you're using, I know Mailman has this feature (from the documentation)
Setting this variable to Yes causes the From: address to be replaced by the list address so that policies like ADSP or DMARC can be applied.
This variable allows you to turn on some simple anonymizing features of Mailman. When you set this option to Yes, Mailman will remove or replace the From:, Sender:, and Reply-To: fields of any message posted to the list.
Note that this option is simply an aid for anonymization, it doesn’t guarantee it. For example, a poster’s identity could be evident in their signature, or in other mail headers, or even in the style of the content of the message. There’s little Mailman can do about this kind of identity leakage.
What you can achieve will depend on the software you use.
Re: What legit email admin ...
And then when they move from that ISP to another ISP - they should have to remember every account they have with every website, and log in to each one and tell them what their new email address will be right?
Yep, and that assumes the website lets you change an email address. Some naïvely assume your email address never changes.
My father still has to keep an account open at our old ISP because a few sites still use his old email address and he can't change it.
Re: @ Stuart Longland (was: What legit email admin ...)
Yes, I run my own email servers. My systems talk to most of TehIntraWebTubes. But I avoid multi-national, billion dollar marketing companies with a passion. Talking to them is more trouble than it is worth. It's not "elitist", it's pragmatic.
As a side-note, you might want to get that chip on your shoulder looked at.
Yes, I run two mail servers as well. I run my own personal one, and I run one for a small company of about 20 employees.
There'd be a big yike from the group if I started to block free email services as many of them use Gmail and the like for personal email addresses. A fair few of our clients also use these services.
Sure, you don't trust them. I've got other reasons for not using them. People trust these services, who are we to judge?
Re: What legit email admin ...
... doesn't bounce toys like yahoo, gmail, hotmail, etc. in the first place?
Someone who isn't elitist enough to expect people to run their own mail servers.
Re: Missed the auction details...
…with a recorded message from mother nature to call back later?
Re: None of this changes anything
A sorry but NTFS HAD file permissions etc. from day one. And was part of NT from at least 3.51. So it always had a well developed permission/restriction setting. MS-DOS and the UI-Add ons (Win3.x, 9x) had not.
Indeed, I just attached a blank disk to a VM and fired up an image of Windows NT 3.1 here. Aside from NT crapping itself when it came face-to-face with a Ivy-Bridge Core i5, I managed to format a disk as NTFS and have a gander. So yes, not that I doubted you on this one, I can confirm that ACLs were present in Windows NT 3.1 which, to my recollection was the first release.
(And it's been 20 years already!)
One thing I pointed out before, on Windows, you've got nothing but ACLs for permissions management. There is no other option.
To compare with Unix, yes, 12 permission bits and a UID and GID field is rather quaint, and yes, ACLs can do that and more. But, it's widely understood and well supported. If you need more advanced features, setfacl and getfacl are there but most of us will stick with good ol'e chmod and chown as they do "good enough". :-)
Re: None of this changes anything
Transfering settings etc. between boxes - why? If you are in a "roaming" environment there is a central software repository with standard installations or a "default image" anyway (and that HAS all the settings since it is basically a backup) and the rest is done by the server storing your relevant profiles. If it is for backup - you can easily do that. The days of "tweaking the registry" died somewhere around the "days of the DOS extenders". OTOH the "guess where config file x is on distribution y" or even the "horde of config files" on Solaris are a PITA for most users
Mmmm yes, because everyone's got their computer on a Windows Active Directory domain, even the home users! (Ignoring the BYOD crowd who buy a laptop with the Home or Standard editions, then expect domain access, I have two of these to support at work.)
Probably the worst bit about the registry is that there's all kinds of magic buried in there about what hardware is in the machine, and if that doesn't match up, all hell breaks loose. Linux actually barely remembers what it had from one boot to the next: it auto-probes much of it.
Think about what happens if a motherboard dies and you have to move the hard drive across to another computer? Most versions of Unix/Linux are happy once you update /etc/fstab and the boot-loader: they'll boot up to the point you can then fix everything else up (such as video). Windows 7+ are not particularly happy, Windows XP even less so.
For the user's part on Unix/Linux, the "hordes of config files" that pertain to a user usually live in their home directory. Often the file begins with a dot, so copy across all the .dot files and you'll probably get what you're after in 90% of cases.
Just like in Windows, if you get their profile directory, you can probably shift that without too many problems.
However if you just want to move the settings for one application… if all the settings are in the registry, then you're in for fun! On Linux you can probably go hunt for a file called .foorc in $HOME and hit paydirt.
DLL Versioning is a problem with all dynamically linking OS, Unix/Linux is no better there. Requires lib x version a.b.c is a problem. Windows is smart enought not to kill newer versions. And multiple .NET versions can exist on one box
Mmmmm, but when app A installs foo.dll version 1, and app B installs foo.dll version 2, is NTFS smart enough to work out that when app A asks for "foo.dll", it wants version 1 and not version 2?
I know ISO9660 does versioning, but even then I don't think that'd work that well.
On Linux, the library would be installed with a name something like: libfoo.so.1 and libfoo.so.2. No naming clash there.
"Posix" is a combination of API definitions and tool definitions. And the API are abstracted away in the compiler libraries so I do not care (much). Non UI / non Appserver software is portable if necessary. More complex stuff is either in a language that does not care (Java, PHP), runs in an app-server or needs UI changes anyway (Try porting a Swing based UI to Android, join the "Nuke Mountain View" fan club)
POSIX is more than API, it also describes the shell commands and syntax. I can write a shell script in the classical Bourne shell, and be certain it will mostly do the same thing across different Unix systems.
It's for this reason, autoconf emits scripts in Bourne shell.
Porting such things to Windows is a pain because amongst other things, there's no 'make', there's no 'sh', and so I've got to maintain a whole special build system just to support Windows. Yes, there's newer alternatives like cmake, but there's a lot of existing software that predate it and for whom, GNU autoconf works fine.
Doing x for the sake of x (Real time is a good example) is stupid showmanship. RT, even more hard(time guaranteed) RT is extremly difficult to program and absolutely unneeded for 99.9 percent of the users. "Close enough" time shedules (That Unix and Windows can do) are more than enough. And those who do need RT - need a certified system and often are restricted to "use THIS". Been there, wrote the Fortran/Dicol and Step5 code.
Hey, you brought up real-time. :-) You asked what the need was, I gave you examples of (soft) real-time systems. I also acknowledged that hard real-time was a more specialised case, and pointed out that some versions of Unix do that too.
As for ACL: There is some support for the older, withdrawn version that can be switched on/used. Under current Windows versions - they are on. Always.
Yep, and most Linux distributions I've used have it turned on too. Under Unix-like systems, you may use it, but there's a simpler alternative that works well enough for most uses. Under Windows you have no choice.
Re: None of this changes anything
Q: What's the problem with the current (W7 and better) registry?
Fragile, difficult to manage, overly complicated, difficult to update without a running instance of Windows.
When it works, it works reasonably well, except some developers use it as a dumping ground for all kinds of crap that doesn't belong there.
But if your Windows installation goes belly up and you need to change a registry setting, you're in for hell. Contrast this with the nearest Linux equivalent I can think of: gconf, or the MacOS X equivalent, plist files, both of which are basically XML files stored with a strict naming scheme.
So anyone with sufficient knowledge can edit the appropriate file with a text editor to correct problems, a text editor that might be running from a LiveCD — this is more difficult to achieve in Windows with the registry.
Q: What's the problem with the current (W7 and better) DLLs?
Nothing with DLLs themselves, the biggest problem is Windows package management. MSI is a step forward, but there's still a lot of cowboy installers out there that just ram stuff in where ever it seems to fit.
End result: the OS has no idea what package owns what file, and so when you go to install a package that conflicts, the installer will likely just overwrite the existing DLL with theirs, or the uninstaller will delete it breaking something else. Worst of all, the DLL version information is embedded in the file, rarely does it feature in the file name.
Contrast this with Linux: where the package manager is responsible for installing and uninstalling packages, will scream at you if you attempt to install two packages which both include the same file, and dynamic libraries are named by their version number (so multiple versions can be installed in parallel).
Q: How many client systems are real time? How many server ones?
Not hard real-time, but you try listening to music or doing anything multimedia-related without some kind of real-time support. Multimedia is a real-time application, if the decoder doesn't keep the buffer on the DAC filled up, the buffer drains, under-runs, then the user complains that the audio is crackly.
VoIP is where this is particularly felt, and is applicable to both server and client — the whole system has to operate to real-time constraints. Too bigger buffers, and the latency becomes untenable for regular conversation, too little, and you become vulnerable to buffer under-runs.
Q: And of those how many are HARD real time like the old ORG/M or (IIRC) OS/9?
There is a hard real-time fork of the Linux kernel. Solaris also supports hard real-time.
Q: What do I need RT for outside of specialist systems
See above. Hard real-time tends to be more for control systems, and thus tend to be more specialist in nature, however examples of soft real-time systems abound everywhere.
Q: What is the benefit of Posix?
Write-once-compile-everywhere. Where there are differences, the changes are small enough that simple #ifdefs in C code can handle it.
I recently did a project wherein I was porting a coal train weighbridge system from SCO OpenServer to Linux. The system itself was based on MacroView SCADA and used UUCP as a means for transmitting weighbridge reports via phone lines back to the central office for billing.
The field computers needed to talk to a particularly old weighbridge controller. I was able to take the source code of their old driver, tweak a few serial port settings in the code and compile it for Linux. The biggest changes were in the stty settings for the serial port, everything else more-or-less JustWorked™.
I hear one of their techs (with no Linux experience) was able to take a field computer out to site and get it going without assistance. A true plug-and-play system.
I shudder to think what hell I'd be in for if I had to port the thing to Windows.
Q: Does Linux support Access Control lists OOB?
It has done for some time. Do we use them? Most of us find that the old-style Unix permissions work well enough for our needs. ACLs are useful in edge cases, but are overcomplicated for the bulk of permissions requirements.
Q: Where is the equivalent of WSUS or ZenWorks? (And no, the Repositories run by "somebody else" are not)
Perhaps you'd like to elaborate on what WSUS and ZenWorks do that aren't present in the open-source offerings? Rather than expecting us to be Microsoft-experts as well as Unix ones.
My rough recollections of ZenWorks was that it allowed deployment of software to a network of machines: something addressed by the combination of repository managers (e.g. apt, yum, etc) and orchestration software (e.g. Ansible, puppet, chef, etc).
Re: Took them long enough...
Ignoring X11 for a moment… which has similarly been around almost as long as I've been alive.
Not everything was invented for the S/360 or came from IBM. Disk storage dated from IBM's 305 REMAC - Random Access Method of Accounting and Control machine - in 1956. COBOL - a staple of mainframe applications - dated from 1959 and was based on the work of ENIAC and UNIVAC programmer Grace Hopper.
That'd be Admiral Grace Hopper… it was in the US Navy that she became acquainted with the ways of computer programming.
Re: An accident waiting to happen? @ChrisBedford
Sounds a lot like Jacques Electronics-style PoE, which used to unconditionally feed 48V down the line. If your kit wasn't built to withstand it, it blew up.
Thankfully, their modern stuff uses the proper IEEE-sanctioned PoE standard.
Re: What about On-The-Go?
Update: Or alternative, express it in megabits, so USB 10, USB 480, USB 5000, USB 10000.
Re: What about On-The-Go?
So is this "USB 10" faster or slower than "USB 480"?
I'd suggest adding a SI-abbreviation, i.e. USB 5G (USB 3.0), USB 10G (USB 3.1), USB 480M (USB 2.0), USB 10M (USB 1.0).
Re: An accident waiting to happen?
Designer: "I know that the devices are supposed to negotiate this, but our device will always want 20 V so we can save money by having a 20 V only PSU."
User: "I need to charge my tablet. This plug will fit it..."
This has already happened. We have here two Digitech CB hand-held radios, both run on, and require, 12V. The newer one has a barrel connector, the older one has a mini-USB. We got car charger cables for them: the one with the mini-USB connector has a big fat warning not to plug it into USB equipment — a warning that was absent on the original power supply for said radio.
Re: Am I the only one ....
Yes. Your current phone charger is also capable of delivering 100 watts.
Bollocks it's meaningless… I'm yet to see a phone charger that can deliver 100 Joules in a second.
Re: Compulsory voting in Australia
so yes ,voting is compulsory but formal voting is not compulsory.
And a good thing too… because it'd be impossible to figure out whose vote was informal to issue the fine.
mmm, Ubuntu, Mint, Debian, OpenSuSE, Red Hat, Puppy - many of those are mainstream, and not one of them would allow the install of the Broadcom driver for the WiFi card in my netbook.
And if you want somewhere to point the finger of blame there, its Broadcom.
Broadcom for ages, officially refused to produce any kind of Linux driver, but instead insisted on people using NDISWrapper to load the Windows NDIS drivers in Linux. The only exception to this was the mipsel binaries for their SoCs used in many wireless routers.
Eventually a team did a clean-room reverse-engineer of the Broadcom driver, that's how we got the b43 driver. BUT, it needs the firmware loaded by the proprietary driver so it can squirt that into the wireless chip to make it work.
The Intel cards need this too, and Intel make the firmware available on their site under a license that permits redistribution. Thus if you've got an Intel wifi card, most LiveCDs already have the firmware and driver, and will JustWork™.
Not Broadcom. Instead, they make it neigh on impossible to get the firmware directly, so one must get a copy of the mipsel driver from a third-party site, and use a firmware extraction tool (b43-fwcutter) to extract it for the b43 kernel driver.
The good news is that some of the newer Broadcom chips, do have a GPLed driver, and Broadcom themselves are the ones pushing it, rather than pushing their own silly STA driver which is proprietary. The bad news is they haven't bothered to make right, the poor situation on their previous chips even though the open-source people have done 90% of the work already. So those of us who have older kit with the older chips, are left high-and-dry.
It was for this reason, I decided to not buy another MacBook. I'm using a late 2008-model MacBook which uses a BCM4322. After all the pain I had enduring the above, I decided my first laptop with a Broadcom WIFI chip would also be my last laptop with a Broadcom WIFI chip.
I bought a new laptop for work about 5 months ago, a Panasonic which had Intel WIFI, and now enjoy largely issue-free networking. The flakey Broadcom-based machine now lives at home where I can put up with flakiness.
They banned XOR (no. 566)? What would George Boole think!
Seems anyone named Dick (no. 488) or Al (no. 421) is banished. As is the gardening implement, the hoe (319). Personal disorganisers are banned too (no. 306).
And don't you Americans even think about gloating about your T1 (no. 706) Internet connections!
Re: Prior God-given art
Let's see you change the tips then…
Re: Probably intended for graphic artists
I have an Intuos 4 tablet here… had it a few years now, and yes, it has interchangeable nibs. I've worn a couple out already.
Re: Typewriter and Abacus are not running XP
I thought those things worked at 300 baud?
Apparently XP is ancient but Windows 2000 is fine.
Does this mean now …
… that I need to start putting sunscreen on my arse now?
My gripe is logging in and pressing "Update" does diddly squat for me as it then asks for a FTP/FTPS host.
I don't run FTP or FTPS on my host: I have SSH/SFTP/SCP for that. I have a heap of shell scripts that download, backup and unpack updates for each bit, but it's extra effort still, and so it doesn't get patched nearly as often as it should because of the above limitation.
This plugin, should be standard issue:
I know they define it earlier in the document as "aging detection circuits" but does anyone else find this talk of "analogue to digital converters" confusing?
Re: Yota phone
And you don't think the early LCDs were the same?
I remember trying to play a pinball game on my father's Pentium 120MHz laptop circa 1996 with its passive-matrix LCD.
I quickly gave up when I couldn't tell which of the 6 balls was the real one and which ones were ghosts.
Re: 2007 hardware obsolete?
If you get 5 years out of a laptop you're winning.
stuartl@portege ~ $ cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 5
model name : Pentium II (Deschutes)
stepping : 2
microcode : 0x16
cpu MHz : 299.951
cache size : 512 KB
fdiv_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pse36 mmx fxsr
bogomips : 599.90
clflush size : 32
cache_alignment : 32
address sizes : 36 bits physical, 32 bits virtual
stuartl@portege ~ $ free
total used free shared buffers cached
Mem: 154788 151980 2808 0 136 85444
-/+ buffers/cache: 66400 88388
Swap: 9775516 11376 9764140
stuartl@portege ~ $ uname -a
Linux portege 3.11.2-portege-dirty #3 PREEMPT Sun Oct 6 13:47:57 EST 2013 i686 Pentium II (Deschutes) GenuineIntel GNU/Linux
What's that machine do all day?
http://aprs.fi/info/a/VK4MSL-1 ← monitors VHF packet.
OS: Gentoo Linux i686
Applications: Xastir, mostly packet-related software, Taylor UUCP.
How do they know they got the same bits back?
Surely they'd be yet another copy?
Re: If they were serious about punishment
they'd give him a Windows
8CE machine instead.
Fixed that for you.
Can you recommend a router that runs windows?
No, but there are a couple that run VxWorks, and some of the Cisco kit runs eCos.
How many people would know what BNC connectors are for today?
And how many use them on a regular basis? RG-58 coax feeds I once used to hammer Ethernet frames down (at the blistering speed of 8Mbps) now cough up 100W PEP of RF to my antennas. A lot of my equipment uses SO-239, N-type and SMA connectors, but my standard RF connector of choice these days is the humble BNC as its cheap, good to 2.7GHz, easy to terminate, quick-connect and durable. Paul Neill and Carl Concelman knew what they were doing when they invented it.
DIN5 I see is notable by its absence, both in that it was used for MIDI, but also for the original PC keyboard port. They're a cheap and reasonably rugged connector in my experience.
IEC power is another notable omission. Show me a standard XT, AT or ATX desktop or server power supply without one.
Re: Fortunately, and gratifyingly
Nah, that'd just make it a stab in the dark.
Re: Link to the interesting bit of the video/128k RAM
Still a damn sight less than the 8MB RAM required to run Linux or the 256MB RAM to run Windows 7…
(And still less than the 2GB+ or so to make either of them useful.)
Link to the interesting bit of the video
I found the linked page wouldn't load properly (after a long time-out, I eventually got a page devoid of formatting and no video). However, I did see the link to YouTube, and managed to find the point of introduction:
Pretty remarkable what they managed to cram into 128kB RAM.
Do I see a potential glut of Debian devs in future?
I can see the recruiters on the Debian list having to beat off the newcomers looking for a free game with a rather large stick.
Yep, we can't have randoms retransmitting data. Best put websites like the Bureau of Meterology behind a paywall and prosecute anyone rebroadcasting the storm warnings for any reason without explicit written permission.
The information is obviously too important to be released to the general public, as a note on the CFA website saying residence of town ABC should leave and that residents of XYZ have left it too late now — may lead to too much confusion.
As you say, you can't have it both ways.
For the record, the govt. is perfectly correct here. This is safety information and there are time-tested methods of distributing public safety warnings. Methods that are work and are managed so that people don't get mixed/conflicting messages.
And increasingly, it's the Governments that are turning away from those time-tested methods.
These days, there's a big push to using the Internet. I say it has its place. Licensing of that data is a big issue, and it's an issue I face myself, as someone who provides emergency communications in such events.
Restrictive licensing prevents someone such as myself, sending a map of affected areas to a station in the area who is cut off from traditional network services. (i.e. via slow-scan television or packet radio instead of email or television) I suppose if taken literally, these licenses prevent me from telling someone: "Don't take some road, it's closed due to floods". This is not helpful.
Even if I was allowed to say what roads were closed or where the fire was located; by the time it has filtered through about 4 sets of ears, brains and mouths, who knows what will come out the other end? The old story of the military unit sending the message "Send re-inforcements, we're going to advance!" getting back to HQ as "Send three and four pence, we're going to a dance!", comes to mind.
Contrast this to say me, taking a screenshot of a webpage depicting the affected areas, a dump of a spreadsheet; transmitting that to someone out in the field and them printing off a few copies/relaying to others. They get a more-or-less verbatim copy of what was on the government authority's site.
The government need to decide if the public should have this information or not. If they're worried about it getting into the wrong hands, then they should keep it secret; and we'll collectively run around like headless chooks causing even more mayhem.
I say: let people re-broadcast the material. If presented to the public in any form, it will be rebroadcasted one way or the other, so it's pointless putting any means to stop it. Doing so does more harm than good.
Guilty as charged…
It's not so bad if you're away from crowds, but this phenomenon happened to me once whilst fiddling with a hand-held GPS.
linux.conf.au 2012 walking back to my accommodation having sourced dinner, I was putting some waypoints into the GPS with the intention of recording details of where one might get food on OpenStreetMap later.
Muggins, not paying attention to where he was walking, walks straight into a 60km/hr speed sign!
So yeah, doesn't matter what mode of transport, WATCH WHERE YOU'RE GOING; collisions happen at any non-zero speed.
Re: Help me out here
Actually no, that calculates as 130mV per cell, if the cell is 1cm² in area. We don't know how big they're making their cells and what the mass of the cells are.
I'd agree with you on the J/kg measurement, but I'm just working with what's mentioned here as I don't have access to the full document or the remaining information. There are factors we don't know, so it's not an apples-to-apples comparison.
Re: Help me out here
0.8mW per cm², a maximum current density of 6mA per cm² and an energy-storage density of 596Ah per kg.
To put this into perspective… I have some 12V (so 6 cell) sealed lead-acid batteries about the size of house bricks weighing in at around 1.5kg. 9Ah capacity, which equates to 6Ah/kg.
We don't know what the mass per cm² is for these new cells, but 596Ah/kg looks quite good.
Authorities are welcome to come search my laptop.
I do not employ disk encryption, and will happily provide them with an administrator password.
Their ability to drive the OS is not my problem however; being a personal computer, I set the machine up for my use, not theirs.
Re: If someone makes those updates available out on the web...
..expect to find a lot of "XP security updates" in warez and torrent sites....
And of course they'll be 100% risk free as they came from a warez site!
Call us when Mozilla have Firefox ported to asm.js…
It's more of a "mid-level" language, only one step above assembly language programming.
C used to be considered a "high-level" language. One line of C code may translate into several machine instructions. Unlike assembler in which one line of assembly is one machine instruction exactly.
I think that's where C gets its "high-level" status from. Granted, it's not as high level as today's modern languages, but it was very high level for its time.
Re: Not another smart phone?
There are some ruggedised handheld computers that would meet your spec.
That's the key though: in this form factor they're not considered "smart phones" but "handheld computers".
Many will run Windows CE or Windows Mobile, but there are some coming out with Android.
for example this. Not a clamshell design, but still pretty rugged.
You probably won't get a clamshell design because the hinge is seen to be vulnerable to damage. Otherwise though, there's a fair selection out there.
Re: OS of your choice
There are a few…
Sailfish, Tizen, Moblin…
Then there's the enthusiast route. Gentoo phone anyone? Unlikely to take the world by storm, but I'd imagine someone will be keen enough to try it.
Re: Cooking the Turkey
You mean Microsoft marketing was wanting to pitch the Surface as a turn-key solution and Engineering misread it as a Turkey solution?
Re: @ Stuart Longland: Wow.
A decade's worth?
10 years ago, OpenOffice struggled with many Microsoft Office documents. Microsoft's Office Open XML standard didn't exist, OpenDocument was not ratified as a standard, Ubuntu didn't exist, and hardware was at times, a hit and miss affair.
A lot has changed. That decade would have consisted of many pilot programs with various Linux distributions, numerous software package trials and probably many failures too. The key point is that they now believe they have succeeded, and I wish them well. I think we'll all be better off for it.