Aug 212019
 

I am reading an article in The Register about a major Internet outage that occurred last December, when a handful of rogue packets managed to clog up a backbone network for more than a day and a half, blocking even VoIP 911 calls.

There are two rather frightening aspects of this fiasco. Both rather horrifying, as a matter of fact.

First, that in this day and age, in late 2018, a backbone service provider can still be brought to its knees by something as simple as a malformed packet. What on Earth are you doing, people? Have you heard of penetration testing? Fault tree analysis? Auditing your equipment and system software? Or have these essential steps been dropped just so that you can report some cost savings to your shareholders?

But it really is the second point that I find particularly upsetting. To quote, “the nodes along the fiber network were so flooded, they could not be reached by their administrators”.

Say what? Are you telling me that you had no alternate means to access your nodes? Like, you know, something as crude and simple as a dial-up port with a command-line based management interface? I mean, this is something even my little home office network used to have, and when I dropped it last year, reacting to rising landline costs and the fact that I no longer used that data/Fax phone line at all, I did so because I have dual network connections. To learn that a major backbone provider doesn’t have the kind of redundancy that I take for granted for my own little network is disconcerting, to say the least.

I suppose I should stop rambling now, though. Truth to tell, I am ignorant as to how CenturyLink’s actual network is configured, and I certainly never managed a fiber optic backbone network. I am simply reacting to the main points of The Register‘s article even though I cannot independently confirm its veracity. In my defense, The Register‘s articles tend to be well written and accurate. Even so, criticizing from a position of ignorance is never a smart thing to do.

Nonetheless, if The Register is correct, this really is not how a transcontinental data network should be configured and managed. This also seems to be the FCC’s conclusion.

 Posted by at 5:04 pm
Jun 132019
 

The news this morning is that former PM Jean Chrétien suggested that Canada should stop the extradition proceedings against Huawei CFO Meng Wanzhou, as a means to win back the freedom of the two Canadian hostages in China, Michael Kovrig and Michael Spavor. (Yes, I called them hostages.)

The case against Huawei runs a lot deeper, however, than the financial fraud Ms. Meng is alleged by US authorities to have committed.

There is also the question of espionage, including the possibility that Huawei’s 5G equipment cannot be trusted because of firmware or hardware level backdoors.

I repeatedly encountered the suggestion that this issue can be trivially remedied by using end-to-end encryption. Unfortunately, end-to-end encryption, even if properly implemented (ignoring for the moment our own Western governments’ recurrent pleas to have built-in backdoors in any such encryption algorithms), solves only part of the problem.

It still allows Huawei to steal metadata, such as where calls are routed or the amount and nature of data traffic between specific endpoints. Worse yet, no encryption prevents Huawei from potentially sabotaging the network when called upon to do so by the Chinese government.

For this reason, I reluctantly came to the conclusion that the US ban against Huawei is justified and appropriate. It must, of course, be accompanied by a suitable increase in spending on researching 5G communications technologies, because otherwise, we risk shooting ourselves in the foot by banning the use of equipment that is technologically superior to the available alternatives. This is a new situation for the West: The last time the West faced a great power adversary that matched Western scientific and technological capabilities was in the 1930s, with Nazi Germany.

As for Ms. Meng, I think the suggestion to suspend the extradition process is wholly inappropriate. It would signal to the world that Canada is willing to suspend the rule of law for the sake of hostages. However strongly I feel about Messrs. Kovrig and Spavor, however strongly I desire to see them released, this is not a price Canada should be willing to pay.

 Posted by at 5:56 pm
Mar 302019
 

Content management software that I use, Joomla! and WordPress in particular, have been complaining for a while now that the PHP version that runs on my servers is outdated and potentially unsecure. Not exactly true, as PHP 5.4 remains part of the official Red Hat/CentOS release, but it would certainly be prudent for me to attempt an upgrade.

I tried to do just that last night, on a test server. And it was a miserable failure, a waste of many hours that I will never get back, to make no mention of the heightened risk of cardiovascular disease due to my elevated blood pressure caused by all that frustration.

The relatively easy part? PHP 7 complaining that its just-in-time compilation feature ran out of memory. Easy-peasy, I can disable JIT. Check.

But then: several of my Joomla! sites refused to run, with a cryptic and ultimately meaningless error message and nothing in the logs. And at least one Joomla! site just got itself into an infinite redirect loop. But why?

I tried many things. I kept looking for answers on Google. Nothing worked. Eventually I took two of my Joomla! sites that are very similar in nature, and began comparing their settings, side-by-side. One worked, the other didn’t. Why?

I then stumbled upon a custom Joomla! module, one that I wrote to support some ads that appear on my sites. This module was installed on the site that failed, but not used on the other. I disabled the module and, presto, the site was working with PHP 7. I re-enabled the module and the site was dead again. So… why?

Well, the module contains some PHP code. Which, after some preamble that allows it to connect to the internal data structures of Joomla!, begins the real work by accessing the MySQL database that contains the actual ads:

$conn = mysql_connect("localhost");
mysql_select_db("www");
$res = mysql_query("SELECT PAGEID,ADTEXT FROM ...

Oops.

You see, mysql_ calls have been deprecated and REMOVED from PHP starting with version 7.

And I have hundreds, if not thousands of lines of legacy code* (including, e.g., my calculator museum at rskey.org) that rely on this old library.

So I guess that PHP 7 upgrade will have to wait a while longer. Looks like I have no choice but to rewrite the affected pieces of code everywhere, as there is no other long-term solution. (Even if I find a third-party PHP plugin that re-enables mysql_ codes, how long will that continue to work? How reliable will it be?)

What a muckup. Grumble. And I do have other work to do.

 Posted by at 10:40 am
Mar 282019
 

Even as Facebook is battling white supremacism and fighting accusations of racial profiling, there is more nonsense going on.

In the past few days, I received several Facebook requests from accounts purportedly owned by young women, whose profiles contain sexually explicit, rather pornographic images and videos.

Here is one of the mildest ones (the majority of the images in this and other accounts from which I received friend requests were far, far more explicit in nature, including images depicting intercourse):

I do not wish to be a prude here; I am, after all, a middle aged male in relatively good health, and certainly not immune to, ahem, shall we say, visually stimulating images (though I admit I was never a fan of hard-core pornography. Not my cup of tea.)

But these Facebook accounts are obviously not accounts owned by bona fide young women trying to seduce older, happily married males like myself. They are probably overweight middle-aged male scam artists doing their shady business from their parents’ basements. Or worse yet, organized crime operating out of shady boiler rooms somewhere in Eastern Europe or Asia.

Thanks but no thanks. I have presently no desire to break my marital vow, but even if I did, there are better, safer ways.

As for these friend requests, I just block them and report the accounts to Facebook.

 Posted by at 2:08 pm
Mar 252019
 

The other day, I started listening to Google Music’s personalized music stream.

I am suitably impressed. The AI is… uncanny.

Sure, it picked songs that I expressed a preference for, such as songs from the golden age of radio that I happen to enjoy. But as I continue listening, it is presenting an increasingly eclectic, enjoyable selection. Some of it is quite new, from artists I never heard about, yet… it’s music I like. For some reason (maybe because I am in Canada? Or because it knows that I am trying to improve my French? Or was it a preference I once expressed for Édith Piaf?) it started presenting a whole bunch of French music, and again… some of it is quite likable. And now that I purposefully sought out a few classical composers, the AI realized that it can throw classical pieces at me as well, which is how I am suddenly listening to Schubert’s Ave Maria.

As a matter of fact, the eclectic choices made by Google’s AI remind me of two radio programs from the CBC’s past, long gone, long forgotten by most: Juergen Goth’s Disc Drive and Laurie Brown’s The Signal. Both these shows introduced me to music from excellent artists that I would otherwise never have heard about.

And now Google’s AI is doing the same thing.

I am also getting the sense that the more I listen, the bolder the AI becomes as it makes its choices. Instead of confining me to a bubble of musical genres of my own making, it is venturing farther and farther away from my presumed comfort zone.

Which is quite impressive. But also leaves me wondering how long before our machine overlords finally decide to take over.

 Posted by at 7:27 pm
Jan 092019
 

So I figured I’d try Office 2019. It came out a few months ago, and I have the MSDN license and everything, so why not give it a test drive. I downloaded the DVD installation image, and ran setup.

Immediately, I was presented with a cryptic error message:

What the devil does this mean? I am installing from a Microsoft-provided DVD image. Surely, it has all the required files?

The “Go online…” link was of no use. Just some generic stuff about installation failures.

A quick online search, however, revealed the culprit: Office 2019 won’t install unless Office 2016 is removed first.

Not something I am inclined to do at this time, not without thoroughly testing Office 2019 first to make sure that it behaves the way I like it (I am especially worried about Outlook and my encrypted imap connection, which can be a bitch to set up.)

In any case… in the software industry, 30 years ago already we had installers that gracefully recognized an existing installation of the same package, and offered either an upgrade or a side-by-side install. Or, worst case, they offered a meaningful error message, informing the hapless user that the prior version must be uninstalled first. Because, you know, chances are anyone installing Office 2019 might already have a copy of Office 2016 installed on their system?

I guess none of that is needed in 2019. After all, what kind of a dumb user am I if I don’t immediately understand Error code 30182-1 (2)?

 Posted by at 9:25 pm
Nov 072018
 

A little over two years ago, I made a decision to stick with Google-branded smartphones running unmodified versions of the Android operating system. I got fed up with brand-name phones that rarely, if ever, received even important security updates and were often hopelessly behind Google releases even when they were brand new on the market.

Our Nexus 6P phones served us well, but recently, first in mine then in my wife’s phone, the battery started to lose charge much too rapidly. This, apparently, is a known and common problem with these Huawei-made phones. Given that Google’s support for the Nexus 6P was coming to an end this month anyway, I decided to look for new phones. I was hoping to get another zero-dollar deal from Rogers, which is how we got those two 6P’s in the first place.

What a disappointment. Sure, the Rogers Web site does show certain phones available at a deep discount, even at $0. But they are available only with certain plans. And the cheapest such plan that I could find would increase our combined phone bill by a whopping $35 (plus tax) per month. No matter how I look at it it seems like a ripoff. Thanks but no thanks. I might as well just go out and buy a pair of unlocked phones, I figured, instead of signing up for these ridiculously expensive plans.

So I started looking, and soon enough my attention was focused on phones produced under Google’s Android One program. This program makes it possible for manufacturers to produce phones that run unmodified (or minimally modified) versions of Android, with the same monthly security updates and same system update schedule that Google-branded phones enjoy.

And that’s when I stumbled upon the Nokia 6.1, also known as Nokia 6 (2018), a supposedly entry-level phone at the ridiculously low price of 320 Canadian dollars.

Nokia, you ask? Indeed, the Nokia brand is still alive, or perhaps coming back to life is a better way to describe it. Microsoft purchased Nokia’s devices business years ago, but in 2016, it sold the Nokia-branded feature phone business back to a newly formed Finnish company founded by former Nokia folks. It is this company that has since created a range of beautiful, low-cost, entry-level smartphones.

Well, the phone may be entry level, but this particular model (TA-1068) still has an octacore processor, 4 GB of RAM, 64 GB of storage, not to mention that it’s an unlocked, dual-SIM phone (or alternatively, I can have an SD card in the second SIM slot.) And it still has features like a fingerprint reader and NFC, not to mention an FM radio. Solidly built, with an elegant design, it does not feel like a cheap phone at all, quite the contrary.

And so far, its battery life proved spectacular. Here is what it showed moments ago:

OK, it was a few hours less than 3 days, and spent mostly at home. But I use this phone a lot!

My only worry was that after a reliable series of monthly updates (which the phone dutifully downloaded after initial setup, necessitating several reboots) there was no October security update and October came to an end. I was guessing that it was because we were waiting for the Android 9 Pie update instead. My suspicion proved correct; the phone is downloading Android 9 right now.

This phone reminds me of my current travel laptop, which I purchased over three years ago I think. It was a kind of emergency purchase; my old laptop was “on the fritz”, and I did not want to spend a lot of money, so I ended up buying this lightweight ASUS laptop for just over 400 dollars. Best purchase ever! My previous laptops, always in the $2000+ price range, consistently proved disappointing. Not this one. Even three and a half years later, it remains snappy, eminently useful, capable of running even decent games on its dual core (four virtual cores) i3 CPU. It also has a touchscreen, handy at times, and something else that’s increasingly rare: plain VGA video output, which makes it much easier to use when I give a talk using older projection equipment. Once I realized how much better this laptop was compared to my expectations, I swapped out its 500 GB hard drive in favor of an SSD and that, of course, also significantly improved its performance. All in all, it does not feel “old” at all, quite the contrary, it is a decent, capable machine that is still a joy to use when I travel.

Of course in the smartphone era, we rely on our laptops less and less. With the ability to check my e-mail and social media accounts on the phone, I found that sometimes I would not even turn on my laptop for days while I am traveling, especially if I am not traveling on business.

 Posted by at 3:47 pm
Aug 282018
 

If you are still running the desktop version of Skype on a Windows 10 computer, be careful before you click “Install now” when the Skype Update dialog appears on your screen.

You see, “the latest version of Skype” that is ready to install is the crappy, worthless, dysfunctional “Metro” version of Skype, which now Microsoft is keen on pushing out to unsuspecting customers, who may naively think that this update, like so many updates before it, is merely another security and functionality fix.

Why am I so hostile to Metro Skype (aka. Skype 8)? I tried using it in the last few weeks. I really did. Here are just a few of my issues with it:

  1. It is not possible to change incoming call notifications, which currently interrupt whatever the hell you might have been doing, stealing focus.
  2. It is not possible to tell Skype 8 to use your desktop speaker for incoming call rings, when you are using an earpiece for the actual call.
  3. The “Metro” UI is buggy. By way of a rather blatant example, when I call a landline contact in Hungary (+36 country code, 1 for Budapest), once the call connects, it shows up in the UI as a call to +1 361 … Needless to say, any attempt to redial fails.
  4. Other “by design” annoyances in the UI, for instance, when there is an incoming message or call that I do not answer, I get a notification in the Chat tab, but the contact is not brought to the top of the list; I have to scroll down manually, hoping that I do not accidentally skip over the contact that appears in boldface.
  5. Though I do not use it, many do: apparently Skype 8 is not integrated with callto: links on Web pages.

In short, the new UI offers a much poorer user experience, it has serious functional deficiencies, and it is buggy.

What I would like to know is what on Earth is going on in the heads of those at Microsoft who are pushing this “upgrade”. And who is going to give me back the hours that I wasted on this boneheaded nonsense.

At least for now, it is still possible to reinstall Desktop Skype by following the direct download link. Just be sure not to “upgrade” by accident afterwards.

 Posted by at 1:31 pm
Aug 212018
 

I often advise my clients that although automation is great, it must be accompanied by well-tested fallback procedures and training, to ensure that they can continue operations even when systems fail. This is especially important in the case of life-critical applications or, for that matter, in applications related to major centers transportation and infrastructure centers.

Today, there was a perfect example that took place at Gatwick Airport in London. Due to a problem with an underground cable, their flight information system failed. That meant no more functioning displays showing departure times and gate numbers. Yet the airport was able to continue operations with no cancellations or delays, and with very few passengers missing connecting flights.

How? Why, they used whiteboards.

Yes. Whiteboards, maintained by officials carrying walkie-talkies.

Now here is the thing. Although the scenes were described as “chaotic” by some, reality was the exact opposite. To quote a spokesperson: “The airport’s manual contingency plan – which included displaying information manually in the terminals and having extra staff on hand to help direct passengers – worked well and tens of thousands of passengers departed on time and no flights were cancelled.”

This is exactly what “business continuity” means. I hope that the person at Gatwick who insisted that such plans must be in place and ensured that not just whiteboards, walkie-talkies and markers but also adequately trained personnel were available when needed, will get not just a commendation but a nice raise.

 Posted by at 6:57 pm
Aug 012018
 

Over the years, I ended up with several Microsoft accounts, and it is a mess. Here is how it happened.

I had a Microsoft account since time immemorial, associated with my personal e-mail. I had my MSDN subscription under this account.

I also had a Hotmail account since time immemorial.

I had a Skype account since time immemorial, too, associated with my personal e-mail. I used my standard, preferred username as my Skype name.

The Hotmail account became an Outlook account once Microsoft acquired Hotmail and created outlook.com. Thus, this became a separate Microsoft account. My standard, preferred username became a Microsoft Live ID.

So here is what I wanted to do at one point: I wanted to use my personal e-mail as my Office 365 and SharePoint online login. But for reasons I no longer remember, there were obstacles along the way. To resolve this problem, I first moved my MSDN subscription to my Outlook account. I then changed my old Microsoft account to be identified not by my personal e-mail but by my Gmail address. This freed up my personal e-mail address to be used as an Office 365/SharePoint online user account.

But then, one day when I was trying to use Windows 10 Quick Assist to offer assistance to someone, the software told me that I need not only to log on to Microsoft, but associate my account with my Skype account. OK, I’ll bite the bullet, I said… and associated my original Microsoft account (now under my Gmail address) with Skype.

And now I am having a problem. Skype tells me that my account has two aliases: The Gmail address and my standard username. But if anyone sends a contact request to my standard username, I get nothing. Today, I figured out why: these requests go to my other Microsoft account! (The one I never used with Skype.) Presumably it’s because my standard username also happens to be my Microsoft Live ID.

Curiously, if I actually log on to Skype using my standard username, I get connected to my Gmail-associated Microsoft account (which is what I want.)

Needless to say, there is no option to merge two Microsoft accounts. There is no option to unmerge a Microsoft account and a Skype account either. I cannot even add my old personal e-mail address as an alias to either of my Microsoft account; presumable because it is now set up as my Office 365/SharePoint online user account, I receive an error message indicating that a “work or school” e-mail cannot be an alias.

As far as I am concerned, this is an unholy mess. Just writing down what happened made my head spin a little. And I really wish I didn’t have to waste a good two hours of my life earlier this morning simply to get to the bottom of it all. (It all began when I made an unsuccessful Skype-to-landline call to Hungary and tried to call again using Skype on my mobile; the software, having updated itself, asked me to log in, and inadvertently, I logged in with the wrong ID. Bad idea, which I eventually remedied by shutting down Skype, deleting all Skype data on the phone, and then restarting Skype and starting all over again.)


As a reminder to myself, here is an excellent page that explains the difference between Microsoft and Office 365 accounts.

 Posted by at 7:28 pm
Jun 162018
 

When I was a teenager, the classic novel, The Count of Monte Cristo by Alexandre Dumas, was one of my favorites.

And one of my favorite chapters in that book was a chapter with an uncanny (not to mention unusually long) title: “How a Gardener May Get Rid of the Dormice that Eat His Peaches”. In it, Dumas describes a classic hack: exploiting the human in the system. By bribing an operator of France’s early optical telegraph network, the book’s protagonist is able to plant a false message, which ultimately contributes to the downfall of one of his mortal enemies. In short: a targeted cyberattack on a telecommunications network.

What I did not know, however, is that this chapter may have been inspired by real life events. About ten years before Dumas published his novel, the brothers François and Louis* Blanc managed to hack the telegraph network in a manner even more sophisticated than the hack described in Dumas’s book. Yes, the real-life hack relied on bribing operators, too, but it also involved a case of steganography: inserting a coded message that would piggyback on the original telegraph transmission. Not only did the scheme succeed, like any good hack it remained in place and undetected for two years. And when it was finally detected, the Blanc brothers were charged but never convicted; there were, after all, no laws on the books back in the 1830s against misuse of data networks.


*Well, that’s what Wikipedia tells me. It appears that the twins are misidentified as Francois and Joseph in several English-language publications. Francois was later known as The Magician of Monte Carlo, a casino that he owned and where he first introduced the single-0 style roulette wheel.
 
 Posted by at 7:52 pm
Apr 142018
 

Yesterday, we said goodbye to our old car, a very nice Honda Accord that served us faithfully for four years.

The lease expired, so we opted to lease a new one. Another Honda Accord. (Incidentally, 2018 marks the 30th year that I’ve been purchasing Hondas, from this very same dealership.)

The old car was nice. The new car… Well, it’s amazing what even four years can mean these days when it comes to vehicle automation.

The level of automation in this vehicle is amazing. It can start itself, it can steer itself. It has full situational awareness, with radar all around. Apparently, it even monitors the driver for alertness (I’ll have to read up on exactly how it accomplishes that.) During the short drive home, it once applied the brakes when its adaptive cruise control was on and someone moved into the lane ahead of us. It was braking a little harder than I’d have preferred, though. And at one point, as the lane markings were a little ambiguous, it gently resisted my attempt to depart from what it thought was the correct lane.

In principle, it appears, this car has all the components for it to be fully autonomous, except that perhaps its array of sensors is not sufficient for it to be fully safe. But really, the only thing missing is the software. And even the way it is, it is beginning to feel more like a partner in driving than a dumb machine; a partner that also has a well-developed instinct for self-preservation.

Welcome to the future, I guess.

 Posted by at 9:54 pm
Nov 102017
 

I’ve seen several news reports commenting on the fact that Donald Trump was using Twitter while visiting China. That despite the fact that Twitter is one of those Western services that are blocked by China’s “Great Firewall”. Some even speculated that Trump was using a military communications network or some other exotic technology to circumvent Chinese restrictions. (As if the US military was foolish enough to let this idiot of a president’s unsecure smartphone access their network.)

But reality is much more mundane, as I know quite well from personal experience in China.

When you are traveling with a phone registered to a foreign service provider, your Internet connection initiates from that provider’s network. So insofar as the Internet is concerned, you are not even in China. Your connection initiates from your home country. In my case, whenever I used my phone in China for Internet access, I accessed the Internet from an IP address registered with my Canadian cellular service provider, Rogers. I had unrestricted access to Google, Facebook, CNN and other news sites, with no Chinese restrictions.

Trump probably did exactly what I did, except that he probably worried about international data roaming charges and data caps a little less than I. He grabbed his phone, turned it on, and used it without a second thought. (OK, that’s not exactly like me. Trump was probably not surprised to see Twitter work on his phone in China, because he probably knows very little about the Great Firewall. I was mildly surprised myself, especially as I went there prepared for the worst, with multiple overt and covert VPN options prepared just in case I needed them. Which I did… but only when I was using the hotel Wi-Fi instead of the cellular network.)

 Posted by at 9:21 am
Sep 262017
 

Lately, I’ve been spending my free time playing post-apocalyptic computer games. Most recently, Fallout New Vegas, from which this screen shot of a full moon rising is taken.

I’ve now played a couple of Fallout games, the two Metro games and the S.T.A.L.K.E.R. game trilogy. Sure there are some common traits but these games are nonetheless quite different. Yet they all have their poetic moments.

The Metro games left the deepest impression on me, to be sure. The characters in these games were perhaps the most realistic, their despair as they clung to life in the tunnels of the Moscow metro under a dead city, almost tangible. S.T.A.L.K.E.R. was something else… for starters, these were games set not in a dead world, only a dead zone in a world that was otherwise alive and well, presumably. And then, Fallout 3 and Fallout NV. When I began playing Fallout 3, I thought that the game lacked soul. Soon, I realized how wrong I was. While the message of the Metro games was that often, it was more satisfying to be merciful (indeed, you can pretty much play through both games without ever killing a human), the message of Fallout is that often there are no good options, only a choice between bad and worse. Killing bad guys or monsters is easy. Killing good guys because they must die for the greater good… Not the choice I’d care to make in real life.

 Posted by at 5:50 pm
Sep 252017
 

Today is September 25. In one of the coldest capital cities in the world. Yet this is the temperature according to the weather monitor gadget on my desktop (but also according to the thermometer on our balcony):

Yes, 3233 C. Or 9091 F for my American friends. The record for this day? A little under 30 C.

No, it does not feel like autumn at all.

On an unrelated note, yes, I do like to use desktop gadgets on Windows 10.

 Posted by at 3:38 pm
Sep 122017
 

I have an old (11 years, to be precise) Konica-Minolta magicolor 2450 laser printer, with the duplexer option.

The quality of this printer’s output is superb, but mechanically, it was never perfect. Just a few weeks after I purchased it, it stopped printing. Faced with the prospect of having to return a 20+ kilo monster, I figured I’d take my chances and look for the cause; after partially disassembling the printer and re-seating some internal mechanical parts, it started printing again.

It worked for many years, but it was becoming rather unreliable. Sometimes, the output was shifted down from the top margin. Paper jams became frequent. It started to fail in mysterious ways, such as complaining that a toner cartridge was missing. Eventually, it stopped printing altogether; every attempt resulted in a paper jam, as the last set of rollers, responsible for pulling the paper out of the printer, no longer worked.

I bought another printer in the meantime, so I retired the mc2450. But I loathed the idea of turning it into e-waste or sending it to a landfill. Today, I decided to take one final look at this fine machine, to figure out what could possible be wrong with its mechanism.

Eventually, I stumbled upon a set of YouTube videos, which offered advice about cleaning some solenoids. Say again?

Solenoids are used in some internal actuators that turn on and off specific parts of the printer’s mechanism. These events are timed with precision. And as it turns out, little plastic pads that are used, I presume, to eliminate clicking sounds and perhaps reduce mechanical wear in the metal parts that are actuated by the solenoids became sticky over time. Just a teeny bit sticky. But that teeny bit is enough for the actuator to become a little lazy. Move a little too slowly. Not much… a few ten milliseconds. But when the paper moves through the printer at, say, 20 cm/s, 50 milliseconds amounts to a centimeter… more than enough for timings to be off and for the mechanism to fail.

Still, it sounded like a stretch. After all, the stickiness was just barely noticeable. Nonetheless… I followed the video’s advice (except that instead of removing/replacing the plastic pad in question, I covered them with kaptonthread seal* tape.) After I reassembled and fired up the printer (and fixed a paper weight adjustment that I managed to set incorrectly), presto: it was printing test pages flawlessly!

Yippie. My old printer was working again. I put it fully back together, and decided to give it another test, this time with its duplexer installed. A huge disappointment: as the paper was feeding through the duplexer, it acquired a nasty fold, very consistently, each and every page. What could possibly cause this?

By this time, I downloaded the service manual for this printer, and studied the diagram of the duplexer a little. It looks deceptively simple, just like an extra back cover for the printer, but it hides complex machinery inside. And guess what… a solenoid actuator, too. And when I disassembled the duplexer and looked at the bit in question, sure enough, its plastic pad was sticky. Ever so slightly sticky, but the stickiness was (just barely) noticeable.

Another few square millimeters of thread seal tape later, after reassembly, my old printer is now printing double-sided documents again flawlessly.

This exercise was not just satisfying but also very educational. That such a tiny flaw can cause all these symptoms. And symptoms that I attributed to (possibly) bad sensors, misaligned or failing mechanical bits, or aging plastic were all caused by actuators that were slowed down, by no more than a few ten milliseconds, tops, by a bit of sticky plastic. Amazing.


*Someone told me it was kapton tape. No, it’s really teflon. My mistake.

 Posted by at 11:40 pm
Sep 082017
 

Jerry Pournelle, the noted science-fiction writer, political pundit and early computer enthusiast, is dead at the age of 84.

Pournelle was a long-time collaborator of science-fiction giant Larry Niven, with whom they co-wrote some amazing science-fiction novels, like The Mote in God’s Eye or Oath of Fealty, not to mention their take on Dante’s Divine Comedy, Inferno, and its sequel, Escape from Hell. Novels he published under his own name included the memorable Janissaries or West of Honor.

Pournelle was well known to readers of the once legendary BYTE magazine. His Chaos Manor column, in which he reviewed software, hardware, new technologies, was very popular.

Pournelle was a political conservative, one of the intellectuals behind Reagan’s Strategic Defense Initiative (aka. “Star Wars”) space defense program. He was a thinking conservative, not blinded by ideology; his views were based on facts and reason.

I corresponded with Pournelle a few times, going back to the late 1980s, when I exchanged e-mails with him on BYTE’s long-defunct dial-up bulletin board, BIX (the Byte Information Exchange). Later, I was an on-and-off subscriber to his Web site and blog. I wasn’t a regular reader, and certainly didn’t always agree with him, but I liked to read his views.

Pournelle suffered a stroke in 2014 and it certainly slowed him down. Even so, he never stopped writing. His passing is not exactly a surprise, but it still came a little too soon. May he rest in peace.

 Posted by at 9:43 pm
Sep 032017
 

In the last few days, I’ve been spelunking in our basement and crawlspace.

I was looking… for many things. Old computer hardware. Boxes to be thrown out. Boxes to be kept, preferably original retail boxes, for packaging things in them that I no longer need.

And finding the unexpected.

For instance… I’ve had an unused old MSI motherboard that has been lying around in my study for ages. Now I don’t recall ever using MSI motherboards. For quite some time, my manufacturer of choice was Gigabyte (no, not married to them, it’s just that whenever I was searching for motherboards, their offerings came closest to what I was looking for.)

So then, during my spelunking, I found the cardboard shipping box of a computer case, and inside it, several parts boxes. Including the retail box for the aforementioned MSI board.

But wait. I had another, identical computer case shipping box, also filled with parts boxes. Including a second box for an MSI motherboard.

So perhaps I did use MSI motherboards after all? Maybe in my server and backup server, around 12 years ago? But if that’s the case… where is the second motherboard, which goes with the second box?

Hmmm… maybe it’s in this test machine? No, the test machine has a Gigabyte board. But let’s double check… Gigabyte branded internal cabling alright… but the board is the second MSI board!

Mystery solved. Except that I still do not remember ever purchasing a pair of MSI motherboards or build computers from them.

But my truly prized finding was something else altogether. (This, I did know about.) Here it is, in its fully functioning glory:

Yes, ladies and gentlemen, that is my first ever server for the vttoth.com domain, decommissioned approximately 22 years ago, in 1995.

The machine has a 386SX motherboard with a whopping 4 megabytes (yes, mega) of RAM. It also has two MFM hard drives: a MAXTOR XT-1085 with about 68 megabytes (yes, mega again) of storage space, and a Magnetic Peripherals 98205-051, with 43 megabytes. Together, about 110 megabytes of storage space.

This machine began its life as my first ever PC-compatible computer that I owned, purchased from a small local company (MICS Computers, no longer in business as far as I can tell) in late 1989 or early 1990 I think. About three years later, I bought another system from them: A powerful monster indeed, with a 486 processor, maybe 16 MB of RAM, but most importantly, a gigantic 500 MB SCSI hard drive, a 525 MB tape backup drive, and, yes, a SCSI CD-ROM, complete with CD caddies. Double speed, too, which means it could read an entire data CD in a mere 30 minutes! My old 386SX system was thus retired.

But it didn’t stay retired for long. Later in 1993, I was asked to serve as the sysop of the UNIX forum of the short-lived National Videotex Network, a service provider that tried to compete with the likes of CompuServe just as the Internet put an end to that business model. I took over from someone who already began creating content, including a brand new upload of the Softlanding Linux distribution, complete with version 0.98pl12 of the Linux kernel. I figured that as a brand new sysop, I ought to know what I was going to be in charge of, so I downloaded the SLS distribution myself and set it up on my old system. It ran beautifully. It was, for all intents and purposes, the same real UNIX that I loved and enjoyed. I was hooked.

Just a few months later, I signed a contract with UUNet Canada, my first commercial Internet service provider. From that point onward, I had a dial-up connection for e-mail, Usenet news, and on-demand Internet. More importantly, UUNet arranged for me a so-called Class C block of 256 portable IP addresses, a block that is assigned to me directly, and which I still use. As the shortage of IP addresses loomed, the powers that be stopped issuing such individually assigned IP address blocks just a few months later.

But when I signed up, the Internet was still mostly non-commercial. So much so that I had to sign the NSFNet Acceptable Use Policy, promising never to use the NSFNet backbone for a commercial purpose! Fortunately, this nonsensical, unenforceable policy was discontinued not long thereafter, but for me, it remains a reminder of just how different the Internet was back then.

Anyhow, this server ran flawlessly for several years, although its limited power and storage capacity were both rather constraining. So it was a relief when I was able to retire it finally in 1995. When we moved to our current home, the machine came with us, only to settle down in the basement for good, where it mostly remained, though I recall powering it up once about a decade or so ago.

So tonight, I dug it out, cleaned it, hauled it upstairs, and powered it up. It came on just fine, along with the monitor, but then an unexpected snag happened: Its BIOS backup battery long dead, the machine asked for the hard drive parameters. You see, ladies and gentlemen, back then there was no plug and play. You needed to know things like the number of cylinders, heads, sectors per cylinder, and precompensation cylinder for your drive. I had to look them up, but fortunately, the Internet knows (almost) everything. Soon, I was booting Linux. Then, another snag: I could not for the life of me remember either the root password or the password to me personal account on this system. Finally, I reminded myself that back in those innocent days, I used much simpler passwords than today… and I was in.

Not much to see, mind you. There isn’t room for much in a mere 110 MB of disk space. But I did see some old e-mails from 1995.

This machine is a keeper. It has history. I just need to find a nice place for it in the house. Oh, and I might want to vacuum its interior, as I noticed a few spiderwebs in there.

Before shutting the machine down, I noticed its performance rating: 2.57 of Linux’s infamous BogoMIPS. In contrast, here is what my current server, built early last year, reports:

Calibrating delay loop (skipped), value calculated using timer frequency.. 4199.71 BogoMIPS
smpboot: Total of 16 processors activated (67195.42 BogoMIPS)

Yup… a machine built about 26 years later, roughly 26,000 times faster. How about that.

 Posted by at 12:22 am
Aug 122017
 

Machine translation still leaves a lot to be desired.

I was watching a cute YouTube video this morning, about a tiny kitten kept warm by a chicken.

The title of the video was in Spanish. My Spanish being nonexistent (in fact, at first I thought the title was in Italian) I used Google Translate. When I used Google Translate most recently, I was tranlating something into Hungarian, so that was the default target language. And Google dutifully translated the sentence, “gallina cuida gatito del frió”, into “Sült csirke cica-ellátás”.

Which means, literally, “Fried chicken kitty-supply”.

Not sure how Google managed to produce this gem of a translation. It offers a reasonably decent English translation: “Hen cares cold kitty”. But the French (“soins chaton de poulet frit” – “kitten care by fried chicken”) and German (“gebratenes Huhn Kätzchen Pflege” – “fried chicken kitten care”) versions are just as atrocious. And the Russian version? “Fried Chicken уход за котенком”… Google didn’t even deign to translate the “Fried Chicken” part (but where did it come from in the first place, when I am translating from Spanish to Russian?) although the rest of the translation (“care for the kitten”) is acceptable.

As I said… machine translation still leaves a lot to be desired.

 Posted by at 9:47 am
Jul 282017
 

I have two “live” uninterruptible power supplies (UPSs) in my study. One powers my servers and network equipment, the other, my workstation. They are identical models: APC Back-UPS Pro 1500. They were both purchased recently, just a few months ago, replacing units that were more than ten years old, and for which replacement batteries were no longer available from a reliable source at a reasonable cost.

So here is the thing: for the second time in a month, the two UPSs decided to do a self-test at the exact same moment.

Now let me make it clear: apart from being identical models and located in the same room, the two units are not in any way connected. They are not synchronized. There is no data connection between the two. When two UPSs simultaneously go offline and start running on battery power, this usually means a power event; but there was no power event, utility power was steady at 124 volts, with no brownout or interruption.

Looking at various logs, I think I solved the mystery. There was a power event recorded on my server on June 30, a brief brownout with line voltage dropping all the way to 74 V:

19:14:33 122.0 13 27.0 100 2910 OL
19:14:34 74.0 11 26.0 100 2910 OB
19:14:35 74.0 11 26.0 100 2910 OB
19:14:36 74.0 11 26.0 100 2910 OL DISCHRG
19:14:37 74.0 11 26.0 100 2910 OL DISCHRG
19:14:38 121.0 14 26.5 100 2910 OL

And lo and behold, my workstation’s UPS sensed the same event at the same time:

<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
  <System>
    <Provider Name="APC Data Service" /> 
    <EventID Qualifiers="0">0</EventID> 
    <Level>4</Level> 
    <Task>0</Task> 
    <Keywords>0x80000000000000</Keywords> 
    <TimeCreated SystemTime="2017-06-30T23:14:33.228802500Z" /> 
    <EventRecordID>6000</EventRecordID> 
    <Channel>Application</Channel> 
    <Computer>VIKTOR</Computer> 
    <Security /> 
  </System>
  <EventData>
    <Data>PowerEvent handled successfully by the service.</Data> 
  </EventData>
</Event>

When I look at the dates of the two UPS self-test events, the first happened just 44 minutes less than exactly two weeks after this brownout; the second, just 44 minutes less than two weeks after the first.

The UPS manual says that this UPS performs a self-test every two weeks. So what’s with the 44 minutes? It can’t be uncalibrated clocks, since the two UPSs remain in sync to the second. Could they be synchronized to utility power? Perhaps, but that does not explain the 44 minutes. The frequency stability of the East Coast grid is much better than 0.2%. In fact, it is purposefully maintained to have an accurate average frequency that deviates from the standard by less than a few parts per million over the course of a day.

So my best guess is that the UPSs are factory configured to perform a self-test every 13 days, 23 hours and 16 minutes. They do this in sync because they are synchronized to the same time source, namely the utility power frequency.

 Posted by at 9:19 pm