Posts tagged ‘errors’


Windows Unreliability Monitor

27.01.2018

Microsoft should rename Windows’ Reliability Monitor to Unreliability Monitor.
   This isn’t too unusual for Windows 10, is it?

   I’ve put Oracle Virtualbox and Cyberlink Power2Go back on, because it’s becoming more apparent that Windows 10 is incompatible with my hard drives in certain circumstances. It’s always when a drive (including a phone set up as a drive) is accessed that the system BSODs. It may also be a USB incompatibility. To be on the safe side, I have unplugged one of the two external drives I use.
   The Microsoft technician has finally given up and asked I do a clean install. As if I have the time—the last time I did that was on an Imac: it took days to get all the OS X updates and the software up and running again. Bwv848 at Bleeping Computer is, like me, determined. I’ll do a memtest (their latest suggestion) when I get a chance.
   Just another day using Windows 10 then.

PS.: Since the post: as my settings window would not come up (another fault of Creators fall), I deleted everything out of C:\Users\[username]\AppData\Local\Packages\Microsoft.MicrosoftEdge_8wekyb3d8bbwe. That was solved. I also went to Intel to download SetupChipset.exe. Not saying these are solutions to the original cause, and I was largely away from the computer for Sunday. However, I have a real suspicion that, because the computer often BSODs when Explorer (or something relying up on it) is open, there are hard-drive drivers that are failing despite, according to Device Manager, being up to date. One of the modules regularly affected is ntdll.dll, something the Reliability Monitor revealed.—JY

Tags: , , , , , , ,
Posted in technology, USA | 1 Comment »


Another program rendered incompatible with Windows 10’s fall Creators update

26.01.2018

It’s fast becoming apparent that Windows 10’s fall Creators update is a lemon, just like the original Windows 10.
   As those of you who have followed my posts know, my PC began BSODing multiple times daily, on average. There were brief interludes (it went for three days without a BSOD once, and yesterday it only BSODed once) but these (now) anomalies don’t really diminish my ‘three to six per day’ claim I made earlier by much.
   And it’s all to do with drivers. I won’t repeat earlier posts but the result was that drivers that came with Mozy, McAfee, Malwarebytes and Oracle Virtualbox caused these. In Mozy’s case, it was an old one. Same with McAfee, the remnants of a program that even their removal tool could not take out. Malwarebytes didn’t even show up in the installed programs’ list, and required another program. In Virtualbox’s case, there were both old and new drivers. They all had to be removed, in most cases manually, because removal procedures don’t seem to take them out. This is a failing, I believe.
   But with all these drivers gone, I still had a BSOD this morning. Four before lunch. The culprit this time was a CLVirtualDrive.sys driver that came with Cyberlink Power2Go, which came bundled when I replaced by DVD burner last year.
   And Cyberlink knows something is wrong with this driver. On December 13, two days after I began getting BSODs, it issued a patch for its latest version. Of course, it leaves those of us with older versions in the lurch, and I was surprised to find that the one it had issued for mine (years old) wouldn’t even run because I was on a bundled OEM edition.
   I’m crying foul. If your program is causing BSODs, then I feel it’s your responsibility to help us out. It shouldn’t matter if it’s a trial version, because this is a window into your business. This signals that Cyberlink doesn’t really want to offer a simple download to prevent users from losing hours each day to fixing their computers, even when they’re partly to blame for the problems.
   Let me say this publicly now: if any of our fonts cause system crashes like this, contact me and I will provide you with fresh copies with which you can upgrade your computer.
   I’m removing Power2Go as I write. It’s superfluous anyway: I only use it because it came as part of the bundle. Windows’ default burning works well enough for me.
   But there’s one thing that Cyberlink’s pages have confirmed: the fall Creators update has problems and it seems to me that it is incompatible with many earlier Windows drivers. We can lay a lot of these problems at Microsoft’s feet. Indeed, based on my experience, you could go far as to say that Windows 10 is now incompatible with many Windows programs.
   That’s all well and good if you have a new computer and the latest software, but what of those of us with older ones who will, invariably, have older drivers or upgraded from older systems?
   Are we now reaching an era where computing is divided between the haves and have-nots? It’s not as though decent new computers at the shops have got any cheaper of late.

Next part: click here.

Tags: , , , , , , , , ,
Posted in technology, USA | 2 Comments »


Mozy driver could have been behind 100–200 BSODs since the Windows 10 Creators fall update was installed

17.01.2018

A post shared by Jack Yan 甄爵恩 (@jack.yan) on

Two very helpful people—bwv848 at Bleeping Computer and Sumit Dhiman at Microsoft—have taken me through the steps to figure out what was going on with my Windows 10 desktop computer, on which I’ve had between 100 and 200 BSODs since the Windows 10 Creators fall update arrived.
   Windows claimed that the error was a DRIVER_IRQL_NOT_LESS_OR_EQUAL in tcpip.sys, but we know that that wasn’t the cause of the crash.
   They had both got to the point where the Driver Verifier had to be run again. On the first attempt, the process had identified an Avira driver, although after removing and reinstalling the anti-virus program, the crashes continued. I had found other dodgy things in the Event Viewer, but solving them didn’t get rid of the BSODs.
   Now that I’m back from holiday—and with Windows 10 crashing one more time and costing me more work that hadn’t been backed up—I gave Driver Verifier one more go.
   I had been averse to it because of the crashes that resulted from it, and had a sense it would tell me the same thing it had in December.
   True to form, Windows wouldn’t even load and it BSODed during the boot. But this time, running Windbg on the dump file revealed something called mobk.sys (Mozy Change Monitor Filter Driver), part of a program called Mozy.
   I’ve never heard of Mozy, but it appears to be a back-up program. Checking my driver, it dates from April 2010 and was installed in 2012—around the time I bought the computer.
   It could well have been installed by me as part of a bundle, or by PB (the retailer).
   Mozy wasn’t helpful. They have a forum, but when you sign up to use it, you get to a page where they want to charge you US$109 for one of their plans. Personally, if I was making software, I’d want reports from people like me. It’s not as though the question was complex: I wanted to know if it made sense to delete the offending driver in safe mode, or maybe download a trial version of their program, then remove it, in the hope that the driver would be overwritten and deleted. It’s only been a couple of hours since I Tweeted them, so I don’t expect any replies till tomorrow.
   Rather than wait, I popped into safe mode and deleted mobk.sys from the system32\drivers folder.
   These errors are deeply frustrating and in direct contrast to the stability that my Imacs have exhibited. Even though I’ve tired of OS X, at least I wasn’t losing work because of three to six BSODs per day.
   The advice I can give to others is to create a system restore point, then run the Driver Verifier, and repeat the two processes until a culprit has been identified.
   There are a few silver linings to this: I got rid of certain software which might have been insecure, and the random resets were quite handy in “clearing” the PC sometimes when I was doing work on it remotely.
   I wonder what had changed in Windows between the spring and fall Creators updates that generated this very serious problem. I haven’t seen Windows crash this often since a dying laptop, on Vista, needed a fresh OS installation (it now runs Ubuntu). I’m still of the mind that Microsoft shipped a lemon, given that I’ve had no end of problems with this OS since it launched, from inconsistent behaviour (Windows 10 would originally be different each time it booted up, from Cortana settings to which keyboard it believed I was using), to very difficult updates (Anniversary took 11 attempts on this PC and never made it on to my laptop even after 40 attempts; it only updated to Creators because all other updates would fail).
   While I can understand that there was no way either Mozy or Microsoft could have checked on a 2010 driver for compatibility, and there are so many configurations of Windows out there, there’s still no escaping that Windows 10 could have shipped with fewer bugs. Happily, as it turned out, the troubleshooting procedures may have worked, even if things wound up taking a month.
   I’ll blog again if I’m wrong about Mozy.

PS. (January 18): After 24-plus hours with no crashes, I got another one, with the same message. Following my own advice, I ran the driver verifier again. Windbg pointed this time to Oracle Virtualbox. I intentionally ran an older version of this because since 2015, no newer version would work due to its hardening feature. Faced with no choice but to update, it had the same error which, finally, I traced to Mactype. This was the error, for those searching:

The virtual machine ‘Windows XP’ has terminated unexpectedly during startup with exit code -1073741819 (0xc0000005). More details may be available in ‘C:\Users\User\VirtualBox VMs\Windows XP\Logs\VBoxHardening.log’.

Result Code:
E_FAIL (0x80004005)
Component:
MachineWrap
Interface:
IMachine {85cd948e-a71f-4289-281e-0ca7ad48cd89}

The key is to insert these three lines into Mactype.ini:

[UnloadDll]
VirtualBox.exe
VBoxSvc.exe

   The error also picked up that there were McAfee drivers left behind from what should have been a full removal. I ran mcpr.exe, found in a thread with the ever-helpful Peter (Exbrit on the McAfee forums). Mcpr.exe did not remove the three drivers, so I took them out manually (despite this going against expert advice): mfeclnrk.sys, mfencbdc.sys and mfencrk.sys. There was also a driver from Malwarebytes, which I downloaded after expert advice in the wake of the damage done by Facebook and its forced download in 2016. Malwarebytes had to be removed with a program called mb-clean as it didn’t show up in the Windows 10 programs’ list.
   One important point: when the system restored itself after the latest crash, it appeared the old mobk.sys reinstalled itself into system32\drivers. I removed it again in safe mode. I’ve since created multiple restore points so hopefully none of the now-removed drivers resurface to cause problems again.
   I am very happy that I’m running the latest Virtualbox, too, since posting in 2015 resulted in no solid leads. It’s why I’m posting all of this stuff, in the hope others find it useful.—JY

P.PS. (January 22): No crashes for three days, I update both the Microsoft and Bleeping Computer threads with the good news, and within nine minutes, bam! Oracle VM Virtualbox is to blame again, if the driver verifier is accurate. That was yesterday. Today, I attempted to remove the program from the Windows Control Panel. Merely removing it caused three BSODs for three attempts, literally within minutes of each other. I booted into safe mode once, tried to remove it (I couldn’t), then back to the regular mode. I was then able to remove Virtualbox. I have since reinstalled it—let’s see what happens next.—JY

P.P.PS. (January 23): Two BSODs this afternoon, still so very disappointed software is this unreliable today. Removing a networking driver from Virtualbox has made no difference. Same error as before. I haven’t re-run driver verifier, but I have now updated MacType to the latest version and double-checked the ini file changes are still there.—JY

P.P.P.PS. (January 24): MacType update did nothing. Bwv848 recommends removing Oracle Virtualbox altogether. I may have to do that, and reinstall it only when I need it, and see what happens. Sumit at Microsoft has given up for the time being.—JY

P.P.P.P.PS. (January 25): After one more crash despite some tweaking of the power options last night, I removed Oracle Virtualbox this morning. There were five remaining drivers that removal did not address, two from the latest version (VBoxNetAdp6.sys and VBoxNetLwf.sys) and three from the old one (VBoxNetAdp.sys, VBoxNetFlt.sys and VBoxUSB.sys). I manually removed them. No crashes since, though I will be interested to know if reinstalling, without any of the old drivers present, will make a difference.—JY

P.P.P.P.P.PS. (January 26): Got to its first crash by 11.45 a.m. Driver verifier now blames CLVirtualDrive.sys. Found one user on Virtualbox’s forum who had the DRIVER_IRQL_NOT_LESS_OR_EQUAL crash but the mod doesn’t like me helping out (very protective people, who don’t like their favourite software criticized). A system restore saw Oracle Virtualbox return, even though I made a restore point long after I deleted it. Let’s see what CLVirtualDrive.sys is. Four BSODs before noon. Man from Mozy got back to me—the first contact other than on Twitter—because they wound up spamming me and never responded to my original support question. Amazing how a few events—including Facebook’s forced download in 2016—have directly led to this time-wasting point in 2018.—JY

Enough postscripts. The next episode of the saga is here.

Tags: , , , , , , , , , , , , ,
Posted in technology, USA | 3 Comments »


Solving my BSODs with Windows 10 Creators fall update—it’s not the usual culprits

11.12.2017

Amazingly, Microsoft Windows 10 Creators fall update arrived last week on my desktop PC, and it took all of 25 minutes to do (running a Crucial 525 Gbyte SSD). (Add an extra 35 minutes for me to put my customizations back in.) This is in contrast to the Anniversary update, which took 11 attempts over many months, including one that bricked my desktop PC and necessitated repairs back at PB Technologies.
   However, I began getting regular BSODs, with the error message ‘Driver_IRQL_not_less_or_equal’ (all in caps), saying that tcpip.sys was the system file affected. An analysis of the minidump file using Windmp revealed that the cause was netio.sys (add ‘Netio!StreamInjectRequestsToStack+239’ if you want the full line).
   There were few people with a similar issue, though I can always count on people in the industry who help—usually it’s folks like Cyrus McEnnis, whom I have known since we were in the third form at Rongotai College, or Aaron Taylor, or, in this case, Hayden Kirk of Layer3, who pointed me in the right direction (that it was either hardware or drivers).
   First up, Windows Update isn’t any help, so let’s not waste any time there.
   Secondly, Device Manager was no help, either. Getting Windows to find updated drivers doesn’t necessarily result in the latest ones being downloaded. If the file that was crashing was tcpip.sys, then it does hint at something afoot with the TCP/IP, i.e. the networking.
   I couldn’t solve it through a virus scan, since a full one would never complete before I got another BSOD. (In fact, one BSOD knocked out Avira, and it had to be reinstalled.)
   It wasn’t Nvidia Control Panel, which was a regular culprit that people pointed to. I did remove and reinstall, just to be on the safe side, but that didn’t fix the problems.
   I had used the ‘Update driver’ option in the Device Manager for my network adapter, the Realtek PCIe GBE Family Controller #2, and while it did update, it wound up on version 1.
   Without much to lose, I decided to feed in the full name of the adapter to look for drivers. Realtek’s website took me here, where I selected the Win10 Auto Installation Program.
   This installed a driver that was version 10, and last updated on December 1, 2017, according to Realtek’s website (the driver is dated October 3, 2017).
   So far I’ve been BSOD-free, and things appear to have settled down.
   If you’re interested, I filed a bug report at Bleeping Computer, and my dump files are there.
   Also remarkable is that my Lenovo laptop, which had attempted to install various Windows 10 updates for over a year, and failing each time (I estimate over 40 attempts, as usually I let it run most times I turn that laptop on; as of April 18 it was at 31 attempts). That laptop was on near-factory settings, so the fact no Windows update would work on it was ridiculous. (I’ve even seen this at shops, where display laptops have Windows update errors.)
   Again, there’s plenty of advice out there, including the removal of Avira as the antivirus program. I tried that a few times over the first 31 attempts. It made no difference.
   I am happy to report that over the weekend, the spring Creators Update actually worked, using the Update tool, and the only alteration I made to Avira was the removal of its System Speedup program.
   And as of this morning, the same computer wound up with the newer fall update.
   There haven’t been BSODs there but to me it confirms that Microsoft’s earlier updates were incredibly buggy, and after two years they’ve managed to see to them.
   I can report that the advice on the Microsoft forums didn’t work and I never needed to result to using the ISO update methods. The cure seemed to be patience and allowing multiple attempts. Since Windows 10 behaves differently each time you boot it up anyway, one of those times might have been compatible with the update patches.
   Hopefully the above helps those who have been struggling with getting their Windows 10s to update. I’d advise against attempting some of the more extreme solutions, especially if your gut or your logic tells you that you shouldn’t need to go to those lengths just to update, when easier solutions worked perfectly fine when you were on Windows XP or Windows 7.

PS., December 12: After a day without crashes post-driver-update, they returned the following day. Investigations are ongoing … I’ve updated the Bleeping Computer link page.

P.PS.: Updated a remote-access program as well as Java (which hadn’t updated despite it having been set to automatic updates). During the former, I had another BSOD as it tried to shut down various network services. Wish I wrote down what they were. However, it does point at a networking issue. Also I saw some hackers in Latvia and the Netherlands try to get in to the system and blocked their IPs. Coincidentally, they had not attempted anything yesterday, which was the day I didn’t have BSODs.

P.P.PS.: Event Viewer revealed those hackers were really going for it. Hayden says it was a ‘port exhaustion hack’, which does, logically, affect TCP/IP. I’ve replaced the remote desktop program, though Java 8 wound back on the desktop because of another program I run. The PC has stayed on since the afternoon, so hopefully that is that. It does mean a day wasted on IT—and it does seem worrying that Windows 10 Creators fall has potentially more holes by default, or somehow falls over more easily when attacked. Those attacks had always come, but they never resulted in BSODs. It was, overall, more robust in updating but it may have some other problems, if the last few days are any indication.
   The external HD was also moved to another USB port. There could be a connection to USBs, as it crashed once after my partner unplugged her phone, and on another occasion I distinctly heard the external HD activate just before a BSOD.

P.P.P.PS.: The above never solved it, but one month on, this might have done the trick.

It didn’t do the trick. Here’s the next part.

Tags: , , , , , , , , , , , , ,
Posted in design, technology, USA | 2 Comments »


Fun for car anoraks—till you get to the factual errors

08.07.2017

I bought Steven Parissien’s The Life of the Automobile: a New History of the Motor Car, which started off as a good history. I’m 300-odd pages in now and the mistakes are really worrying. There’s also a shocking lack of editing (one part repeated, albeit in different language, and spelling and grammatical mistakes) in the parts I’ve got to now; it’s as though the editor got tired after the first 10 chapters and stopped caring. But the biggest errors are factual.
   I am astonished to learn, for instance, that Harley Earl was responsible for the concept of the Ford Thunderbird (p. 255), that Triumph TR7 production was transferred to Speke in 1982 (p. 293, though Parissien later contradicts himself with the correct fact), and that John Z. de Lorean was a protégé of Lee Iacocca (p. 309). I really have no idea how, but as far as I know, de Lorean was never at Ford, and he had a Chrysler stint long before Iacocca got there. I also never knew that ‘In 1968 the Toyota Corolla became the first Japanese car to be manufactured in the US’ (p. 314; that was the year it went on sale there, and from memory the Corolla didn’t get built there till the NUMMI deal in the 1980s) or that the Opel Ascona C was also sold as the Opel Vectra (p. 337). The Italian Job was released in 1969, not 1967 (p. 224).
   I am frustrated with this book—and now it makes me wonder if the stuff earlier on, which I know less about, was accurate.
   I can understand an editor not grasping the subject as well as the author but there is less excuse in professional publishing for the other problems. Maybe there are few professional proofreaders left, now that spellchecks have been around for a generation or more. I was prepared to recommend this book even a week ago and tolerated the spelling and grammar, but these factual mistakes are worse than what can be found in Wikipedia, and I often label parts of that site as fiction.

PS. (September 17): How much worse can it get, as I continued through? A lot.
   On p. 320, we get an admission that Parissien was wrong on p. 314: the Honda Accord was the first Japanese-branded car to be made Stateside. At least an earlier error was corrected. But they begin again on p. 321: Parissien claims the V30 Toyota Camry dominated the US mid-size car market (it was never sold outside Japan; he’s thinking of the XV10, or the Japanese-market Scepter, which was badged Camry). Correcting his error on p. 322, the Camry was not specifically targeted at the US; it was Toyota’s attempt to create an efficient car from the ground up, and it was not done in 1980, but 1982 (the 1980 Celica Camry was not sold outside Japan). The Paykan deal was cemented long before George Turnbull got to Iran (p. 324), though local content rose in the 1970s for it to be truly Iranian-made and Parissien might mean the shifting of the engine tooling there, if I’m being generous. There is only one world, not multiple ones (also p. 324), unless Parissien knows something about parallel universes that the rest of us don’t. Surely Chrysler managed to launch its T-115 minivan (Dodge Caravan, Plymouth Voyager) before Renault launched the Espace (p. 330), and the Triumph Acclaim was never called the ‘Ronda’ (p. 334), though its successor was.
   On p. 360, in a single paragraph, Parissien makes several errors. The MG 6 launched as a five-door car, not his claimed four. There is no such thing as the Roewe 150, in China or elsewhere. The MG 3 has five doors and is not a three-door counterpart to the 6. The five-door MG 6 launched not in 2011, but in 2010, and the Magnette name was only used in the UK for the four-door. The founder of Chrysler was Walter Chrysler, not Walther (p. 364). The Lancia Delta only became a Chrysler in the UK and Éire, as far as I know (p. 365), and remained a Lancia in most countries. The Belgian designer is Dirk van Braeckel, not van Braeckl (p. 368); Mercedes-Benz never bought an 18·53 per cent stake in Volkswagen (p. 369); and Citroën’s BX was not the last car in that range to have ‘pneumatic suspension’ (p. 372). The Malibu was not a Saturn, but a Chevrolet (p. 375), and Buick was never sold off (p. 376). The Ford Mondeo did not replace the Telstar in all Asian markets (p. 377), and it shared far more than the ‘windscreen, front doors and rear’ with the Contour and Mystique (in fact, the rear was not shared, though there were common engines, platform, and plenty more). It’s not entirely certain that the US market judged the Contour to be too small (p. 378), but there was a lack of marketing (which would have made an even better story than the one Parissien writes about). Ford subsequently filled the Contour’s niche with the smaller Focus Stateside. The CD338 Fusion was never sold in Australia (also p. 378). Ford never resurrected the Taunus in Germany under Alan Mulally (p. 381)—this invention is incredible. VAZ did follow up the 2101 with something similar after it ended production in 1983 (not 1984, p. 382), viz. the 2105, which was about as similar as one could get to the 2101.
   On p. 384, Parissien claims Acura’s Legend sales were ‘disappointing’, after saying they were ‘beginning to sell rather well’ 50 pp. before. The Hindustan Ambassador was not based on the 1954 Morris Oxford (p. 389): that car was actually the Hindusthan Landmaster. The Ambassador was based on the 1957 Morris Oxford III, and was in production from 1959, not 1958. The Red Flag (or Hongqi) marque was not reborn on an Audi A6 (p. 391), but the marque had been used on a version of the Audi 100 C3 from 1988, and no Hongqi bore an Audi–Chrysler–Hongqi brand name. The Chinese company is Dongfeng, not Dongfen (p. 391), and Parissien’s claim that the Everus was sold in the west (p. 392) is news to me, as I am sure it is to its own management. I’ll stop there for now.

P.PS. (September 18): Some bedtime reading, or should I say error-finding, last night. On p. 394, Toyota and Aston Martin did not jointly develop the Cygnet: Toyota developed the IQ in 2008, and Aston Martin converted that car to become its Cygnet, and ‘hot hatch’ is a very optimistic description for a city car. Toyota did not launch the Cygnet in 2008 as Parissien claims, nor did it have a say in what customers were expected to purchase the Cygnet: it was aimed specifically at existing Aston Martin owners, not ‘Toyota and Aston expected initial demand to be limited to those who already owned an Aston Martin sports car.’ It was certainly no ‘eccentric experiment’ of Toyota, but of Aston Martin. Volvo never made a model called the A40 (p. 395), and I bet Nissan is surprised to find that the original Qashqai was designed ‘at the firm’s Milan design centre’ (p. 397) when it was designed in London. Maserati never launched a Jeep-based SUV called the Kubang (p. 397), but it did have a concept of that name, and the Levante appeared in 2016 after the book was published. There is no such car as the Porsche Cajun, and if Parissien refers to the smaller Porsche crossover, then that is called the Macan, and it has five doors, not the claimed three (p. 397). The Volkswagen New Beetle was not on a Polo platform (p. 399), but a Golf one, as was its successor (though a newer Golf); and Ford would dispute that its Mustang is a sedan (p. 401). If J Mays’s first name is J (as footnoted), then there is no need to refer to him as ‘J. Mays’ (p. 401). The Ford Ka’s name is not derived from StreetKa (p. 402): that was a model spun off from the Ka in 2002; and some would regard the Mk II model was being superior to the Fiat 500 on which it is based (especially as Fiat adopted some of the changes for its own model). I have yet to see a Smart with a Mercedes-Benz three-pointed star or marque anywhere, unlike Parissien who seems to think they are badged Mercedes (p. 403), and a Smart SUV does not exist unless Parissien is reporting again from his parallel universe (p. 403). There is also no such car as the Kia Exclusive (p. 410).

Tags: , , , , , , , , , , , , , , , , , , , ,
Posted in cars, design, interests, publishing, UK, USA | 1 Comment »


Facebook lets me have full access on someone’s public page—but I’m not an admin

02.06.2017

I have long maintained that Facebook’s databases are dying (hence their need to force people to download malware) and tonight’s discovery is a case of ‘What more proof do you need?’
   Tonight, I can edit a verified (blue-ticked) Facebook page with a fan base in the high five figures that is not mine. I can view all the messages, remove admins, receive notifications, and comment and like as that page. The one thing I cannot do is notify the real owner of that page via Facebook messaging.
   This is not unlike in 2013, when someone found themselves a fan of my public page—but they never liked it. Fortunately for me, they believed us when we said we knew nothing of it.
   And fortunately for this person, I am (a) not dodgy and (b) I know her in real life, though I have not spoken to her in over three years. She hasn’t made me an admin. I’ve looked on the list of pages I really administer and hers isn’t there. I’ve gone into her page’s settings and the page roles, and I’m not listed as an admin. Yet I can do everything an admin can. There’s a box right there for me to add other people as admins to her page. I could kick her off.
   I tried contacting this person’s private profile via Facebook messaging as myself. Impossible. I can’t attach screen shots to show her what I discovered, and clicking ‘Send’ does nothing. I will, of course, email her.
   How did I find out? Someone shared an article from the Lucire Facebook page. I clicked through to see if the sharer had written anything. I wanted to ‘like’ the share as Lucire rather than myself, and discovered that I could only be me and this other person. In fact, I could do nothing in the name of the pages I actually run. The sharer does not have either me or this person as Facebook friends.


The first clue. How come I can comment as this person?


I can only comment as myself as this one other page that I have no current connection to.



Sure enough, I have full access to the site settings and messages.


I’m not an admin, though I seem to have all the admin privileges.



Full access to mess around with her posts, and further proof I can comment as her.

   This blog post is a warning to anyone with a Facebook page. Just know that at any time, access to your page can be granted to someone else.
   If pages are no longer secure, then I have to ask: what is the point of Facebook?
   This isn’t good news for us at all because one of the businesses I am involved in relies on Facebook.
   But it’s certainly a risky platform to be on, and I am willing to bet this bug will become more widespread.

Tags: , , , , , , ,
Posted in business, internet, technology, USA | 2 Comments »


My first day with Windows 10

17.11.2015

I never expected that the Windows 10 download would ever begin. I had registered for it, but the Windows 10 notification window kept coming up with various excuses, talking about my drivers being out of date, then claiming that because of the automatic log-in, the download would not start. Clicking ‘Tell me more’ never did a thing: that took you to Microsoft’s home page. This went on for months.
   But on a day when I upgraded a new office Mac to El Capitán (and migrated the old data on to it), and reinstalled Lubuntu on another PC which threw a wobbly after being asked to adopt the Cinnamon desktop (which took many hours), Windows 10 decided it was ready after all on my main Windows 7 machine, via Windows Update.
   First signs were promising, with the 2½ Gbyte download coming in five minutes, although the computer stayed on ‘Preparing for the upgrade’ overnight. I had heard that the upgrade process would take an hour or so, not overnight, so something was wrong. Typically I had this trouble with one of the Macs on OS upgrades—one took 18 hours while another took 30 minutes once—but this was new territory for Windows.
   One thing I will say for Windows is that when things go wrong, there is help. One piece of advice, which proved right, was to crash out of the process (using the process manager), and to start it all over again.
   Windows just need a second stab at it, and recommenced the download. This time the ‘Preparing’ window flashed up and was gone, and the hard yards then began. It did wind up taking just over an hour. Getting it on a second attempt isn’t bad, considering I’ve had Mac OS upgrades fail far more times than that.
   First impressions are pretty good though most Windows 7 initiés will tell you that things are a bit harder to find. Don’t believe a soul when they tell you it’s faster to boot up: it isn’t. I’m sure it takes an extra minute compared to 7. Doing some basic things in the File Manager takes more movements of the mouse, to open menus and to click, and the menus aren’t as streamlined once you open the panel to find the functions you used to see at a glance. Little annoying things included Windows 10 forgetting that I had set Cyberfox as my default browser—it really loves Edge, and admittedly, it is a nice, fast program—and the time zone changing without you noticing (I prefer GMT, but Windows kept altering it to NZDT). You have to dig a bit deeper into the menus to make these things stick, such as going through the default programs’ dialogue box, and turning off Windows’ ability to check the time. Having opted for UK English, Cortana refused to work—curiously, it claimed that the installed US English pack was an unsupported language, until I downloaded the same for UK English.


Cortana gives completely the wrong address for me. I wonder if the resident of 39A Aparima Avenue is getting identified as the home of a lot of Windows 10 users.

   There’s not an awful lot that Cortana can tell you. Most enquiries wind up on Bing, and she’s only really good for the weather and exchange rates (as I have discovered so far). There are a few fun questions you can throw at her, asking if she’s better than Siri, or whether if she’s met Bill Gates, but generally, but we’re far from Knight Rider or replicant technology here. A New Zealand accent presents no problems. One thing she gets very wrong is my location, which is allegedly 39A Aparima Avenue in Miramar. I’m not sure how she arrived at that address, as I don’t live there and I don’t believe I know the person who does.
   It’s not too unpleasant to look at although the mobile-specific features can get a bit annoying. The menus feel too large overall, because it’s all designed from a mobile-first standpoint, while the biggest gripe from me comes with the typography.
   Microsoft has ruined ClearType here in its attempt to make something for mobile first, and most type looks very poor on screen. Fortunately, a Japanese website still hosts the MacType plug-in, which brings the font display closer to what we experience on Mac OS X. It even goes beyond what we were used to in Windows 7, which had been Microsoft’s best use of its ClearType technology to date.


After installing MacType, ITC Legacy Serif looks far more like it does in print.

   You can alter the fonts through the Registry Editor, and I set about getting rid of Arial as always. Windows 10 doesn’t like you removing a system font, so the trick is to replace it with something else called Arial, then remove the original from HKLM\Software\Microsoft\Windows NT\CurrentVersion\Fonts.
   Windows 10 removes your ability to change the icon and menu fonts, and they now have to be changed in the registry, too, at HKCU\Control Panel\Desktop\WindowMetrics, and very carefully.
   After tinkering with those, the display began looking like what I was familiar with, otherwise there was a bit too much Segoe on screen.
   There have so far been no program incompatibilities. As upgrades go, it hasn’t been too bad, and I haven’t been stuck here forever downloading updates. Apple still gets higher marks for its OS upgrade processes (when they work) but given how much data I have on my main Windows machine, and how different each PC is, Microsoft has done a good job. I’m glad the system waited till now, and delivered me a relatively bug-free transition. Software upgrading is one area where I don’t mind not being first.

Tags: , , , , , , , , , , , ,
Posted in design, technology, typography, USA, Wellington | No Comments »


Facebook’s still deleting drag accounts and keeping bots

21.06.2015

Some interesting bugs out there on Facebook that my friends are telling me about. One has been removed from all her groups, including one that I run (we never touched her account), another cannot comment any more (an increasingly common bug now), while Felicity Frockaccino, well known on the drag scene locally and in Sydney, saw her account deleted. Unlike LaQuisha St Redfern’s earlier this year, Felicity’s has been out for weeks, and it’s affected her livelihood since her bookings were in there. Facebook has done nothing so far, yet I’ve since uncovered another bot net which they have decided to leave (have a look at this hacked account and the bots that have been added; a lot of dormant accounts in Japan and Korea have suffered this fate, and Facebook has deleted most), despite its members being very obviously fake. Delete the humans, keep the bots.
   Felicity didn’t ask but I decided to write to these people again, to see if it would help. There was a missing word, unfortunately, but it doesn’t change the sentiment:

Guys, last year you apologized to drag kings and queens for deleting their accounts. But this year, you have been deleting their accounts. This is the second one that I know of, and I don’t know that many drag queens, which suggests to me that you [still] have it in for the drag community.

https://www.facebook.com/DoubleFFs/

Felicity Frockaccino is an international drag performer, and you’ve affected her livelihood as her bookings were all in that account. This is the second time you deleted her, despite your public apology and a private one that you sent her directly. What is going on, Facebook? You retain bots and bot nets that I report, but you go around deleting genuine human users who rely on you to make their living. Unlike LaQuisha Redfern’s account, which you restored within days, this has been weeks now.

   That’s right, she even received a personal apology after her account was deleted the first time. I had hoped that Facebook would have seen sense, since Felicity has plenty of fans. The first-world lesson is the same here as it is for Blogger: do not ever rely on Facebook for anything, and know that at any moment (either due to the intentional deletion on their end or the increasing number of database-write issues), your account can vanish.

Meanwhile, my 2012 academic piece, now titled ‘The impact of digital and social media on branding’, is in vol. 3, no. 1, the latest issue of the Journal of Digital and Social Media Marketing. This is available via Ingenta Connect (subscription only). JDSMM is relatively new, but all works are double-blind, peer-reviewed, and it’s from the same publisher as The Journal of Brand Management, to which I have contributed before. It was more cutting-edge in 2012 when I wrote it, and in 2013 when it was accepted for publication and JDSMM promoted its inclusion in vol. 1, no. 1, but I believe it continues to have a lot of merit for practitioners today. An unfortunate, unintentional administrative error saw to its omission, but when they were alerted to it, the publisher and editors went above and beyond to remedy things while I was in the UK and it’s out now.

Tags: , , , , , , , , , , , , , , , ,
Posted in branding, business, internet, marketing, technology, UK, USA | 2 Comments »


Remember when you could post, comment and like on Facebook? Those were the days

29.05.2015



Let’s see: Facebook doesn’t work on Wednesdays and Fridays. Check. Thursdays are OK though.
   It’s another one of those days where the Facebook bug that began on Wednesday (though, really, it’s been going on for years—including the famous outage of 2013 where what I am experiencing happened worldwide to a large number of users) has decided to resurface and spread. Not only can I no longer like, comment, post or share without repeated attempts, I cannot delete (Facebook makes me repeat those attempts even when a post has been successful, but doesn’t show me those till an hour later) or upload photos to messaging without repeated attempts.
   The deletion is the hardest: while commenting will work after three to twelve repeats, deletion does not work at all. The dialogue box emerges, and you can click ‘Delete’. The button goes light for a while, then it’s back to the usual blue.
   And this happens regardless of platform: Mac, Windows, Firefox, Opera, Android, inside a virtual machine, you name it. Java’s been updated as have the browsers on my most used machines; but it seems the configurations make no difference.
   I am reminded how a year ago I had even less on Facebook. Quite a number of users were blocked for days (Facebook isn’t open on weekends, it seems), but eventually the message got through and things started working again.
   My theory, and I’d be interested to learn if it holds any water, is that older or more active accounts are problematic. I mean, if spammers and spambots have more rights than legitimate users, then something is wonky; and the only thing I can see that those T&C-violating accounts have over ours is novelty. Facebook hasn’t got to them yet, or it tacitly endorses them.
   As one of the beta users on Vox.com many years ago, I eventually found myself unable to compose a new blog post. It’s an old story which I have told many times on this blog. Even Six Apart staff couldn’t do it when using my username and password from their own HQ. But, they never fixed it. It was a “shrug your shoulders” moment, because Vox was on its way out anyway at the company. (The domain is now owned by another firm, and is a very good news website.) Unlike Facebook, they did have theories, and tried to communicate with you to fix the issue. One woman working there wondered if I had too many keywords, and I had reached the limit. I deleted a whole lot, but nothing ever worked. It suggested that these websites did have limits.
   Computer experts tell me that it’s highly unlikely I’ve reached any sort of limit on Facebook, because of how their architecture is structured, but I’m seeing more and more of these bugs. But we are talking about a website that’s a decade old. My account dates back to 2007. Data will have been moved about and reconstituted, because the way they were handled in 2007 is different to how they are handled now. There have been articles written about this stuff.
   What if, in all these changes over the last eight years (and beyond), Facebook screwed up data transfers, corrupting certain accounts? It’s entirely conceivable for a firm that makes plenty of mistakes and doesn’t even know what time zones are. Or deletes a complainant’s account instead of the pirate’s one that she complained about. (This has been remedied, incidentally, the day after my blog post, and a strongly worded note to Facebook on behalf of my friend.)
   The usual theory I hear from those in the know is that certain accounts are on certain servers, and when those are upgraded, some folks will experience difficulties. That seems fair, but I would be interested to know just what groups us together.
   Last time I downloaded all my data off Facebook, and this was several years ago, I had 3 Gbyte. It wouldn’t surprise me in the slightest that that was now 6 Gbyte. That’s a lot to handle, and when you multiply that by millions, some will result in buggy accounts. Ever had a hard drive with dodgy fragments? Or a large transfer go wrong? Facebook might have better gear than us, but it’s not perfect.
   I don’t believe for a second that certain people are targeted—a theory I see on forums such as Get Satisfaction, with Republicans blaming Democrats and Democrats blaming Republicans—but I do believe that something binds us together, and it is buried within the code. But, like Vox, it may be so specific that there’s nothing their boffins can do about it. You simply have to accept that some days, Facebook does not let you post, comment, like, share, delete or message. The concern is that this, like random deletions, can happen to anyone, because these bugs never seem to go away. Looking at my own record on Get Satisfaction, they are increasing by the year.

Tags: , , , , , , , , , ,
Posted in internet, technology, USA | 6 Comments »


Switching to Cyberfox, after Waterfox and Firefox stopped displaying text

23.12.2014

Since the Firefox for Windows updates in November, I’ve had a big problem with the Mozilla browser, and the Waterfox 64-bit version based on it: they won’t display text. I had to downgrade to Waterfox 32.0.3 for the last month or so, but it’s begun crashing more and more regularly (from once a day to thrice today—I visit largely the same sites, so why does software “decay” like this?).

   On the latest incarnations of Firefox and Waterfox, linked fonts work, but the majority of system fonts vanished from the browser. And, for once, I’m not alone, if Bugzilla is any indication. It is probably related to a bug I filed in 2011.
   I’ve had some very helpful people attend to the bug report—it’s great when you get into Bugzilla where the programming experts reside—but sadly, a lot of the fixes require words. And, unfortunately, those are the things that no longer displayed in Firefox, not even in safe mode.
   As many of you know, there’s no way I’d switch to Chrome (a.k.a. the ‘Aw, snap!’ browser) due to its frequent crashes on my set-up, and its memory hogging. There’s also that Google thing.
   After some searching tonight, I came across Cyberfox. It’s not a Firefox alternative that comes up very often. Pale Moon is the one that a lot of people recommend, but I have become accustomed to Firefox’s Chrome-like minimalism, and wanted something that had a Firefox open-source back end to accompany it. Cyberfox, which lets you choose your UI, has the familiar Firefox Australis built in.
   I made the switch. And all is well. Cyberfox forces you to make a new profile, something that Waterfox does not, but there isn’t much of an issue importing bookmarks (you have to surf to the directory where they are stored, and import the JSON file), and, of course, you have to get all your plug-ins and do all your opt-outs again. It also took me a while to program in my cookie blocks. But the important thing is: it displays text.
   You’d think that was a pretty fundamental feature for a web browser.
   The text rendering is different, and probably better. I’ve always preferred the way text is rendered on a Macintosh, so for Cyberfox to get a bit nearer that for some fonts is very positive. It took me by surprise, and my initial instinct was that the display was worse; on review, Firefox displayed EB Garamond, for example, in a slightly bitmapped fashion; Cyberfox’s antialiasing and subpixel rendering are better.

Firefox and Waterfox on Windows 7
Firefox_Screenshot_2014-12-23T14-14-06.693Z

Cyberfox on Windows 7
Firefox_Screenshot_2014-12-23T14-13-23.809Z

Here’s where the above text is from.
   Gone is the support for the old PostScript Type 1 fonts (yes, I still have some installed) but that’s not a big deal when almost everything is TrueType and OpenType these days.
   The fact Cyberfox works means one of two things: (a) Cyberfox handles typography differently; or (b) as Cyberfox forces us to have a new profile, then there is something in the old profiles that caused Firefox to display no text. That’s beyond my knowledge as a user, but, for now, my problems seem to be solved—at least until someone breaks another feature in the future!

PS.: That lasted all of a few hours. On rebooting, Cyberfox does exactly the same thing. All my text has vanished, and the rendering of the type has changed to what Firefox and Waterfox do. No changes to the settings were made while the computer was turned off, since, well, that would be impossible. Whomever said computers were logical devices?
   Of yesterday’s options, (a) is actually correct—but how do we get these browsers behaving the way they did in that situation? In addition, the PostScript Type 1 fonts that the browser was trying to access have since been replaced.

Tags: , , , , , , , , , , , ,
Posted in internet, technology, typography | 3 Comments »