No Food for Thought

Food is something you should provide to your brain long before coming to this blog. You will find no food recipes here, only raw, serious, non-fake news for mature minds.

Modestly Moving Away from a Monstruously Mad Mozilla

admin Sunday April 22, 2018

In 2003, I was using Internet Explorer, Hotmail and Microsoft Windows when I discovered the Mozilla Suite. I could certainly count the number of open source applications I used on a single hand at that time. I installed Mozilla 1.4, a friend told me about Firebird, and then I switched to Firebird and never went back to MSIE as my main browser. Impressed by Firebird (now Firefox), and curious about free software, I migrated from Windows to GNU/Linux later that year. Since then, Firefox has always been the main browser on at least one of my operating systems. For most of these years, Firefox (or its close relative Iceweasel) was also my main browser.

Soon after discovering Mozilla's browser, I tried Thunderbird, which I started using instead of webmail. For most of these years too, Thunderbird (or its close relative Icedove) was my MUA. So it has been nearly 15 years since Mozilla became one of the projects with the most influence on the software I use.

For years, I was evangelizing friends and family about Mozilla software. I was also a developer of projects upstream of and downstream to Mozilla. When Firefox was released, I was helping early adopters on IRC channels. For more than a decade, I also sporadically contributed to Mozilla's wikis. I wrote and helped triage issue reports. For some time, I even edited Wikipedia's page about the Mozilla-Debian licensing/branding issue. When I discovered Firebug thanks to a friend, I immediately installed it. Over time, I would report many of its issues.

Despite the licensing issue and XUL's reliance on GTK+ instead of Qt (my favorite desktop environment's toolkit), I never hesitated before contributing to Mozilla. In 2010, I was thrilled to realize the milestone we had reached when Microsoft's Internet Explorer dropped below 50% in browser market share. Years later, Mozilla finally agreed to relicense its logos, and the old Mozilla-Debian conflict ended.

For years, we wrote history. And then, in 2016, I realized a major issue using Gmail via Thunderbird had been affecting me for years. That issue had been slowing me down and putting the security of my customers at risk. Following that initial realization, I spent numerous person-days dealing with the damage (cleaning up my mailbox, directly and indirectly). But in the course of that long process, I also realized more issues in Thunderbird/Gmail, as well as in the Mozilla project itself. Since these are numerous, I will not list them here, but instead ask those who volunteer in Mozilla to read the overwhelming report I sent to Mozilla's governance and bugmasters mailing lists.

This mail appears on the archives of Mozilla's governance mailing list, but careful readers will notice it does not show in the archives of the bugmasters mailing list. Indeed, although I did send the mail there (using Thunderbird!), the mail apparently never reached the list, for reasons unknown to me.

That additional problem is one of those I reported in a follow-up to "Issues, meta-issues and transparency" which I sent to the governance list on 2018-01-13. Weeks later, I hadn't received any response to my follow-up, and realized that mail had apparently not reached the governance list, so I resent it on 2018-03-10. Unfortunately, I can only link to a local version of this mail, since the second sending visibly also didn't reach the list.

I have to conclude these events did not just reveal issues in Gmail and Thunderbird, but huge issues in the Mozilla project - not just in its issue tracking, but also in its mailing lists (and yes, Mozilla also knows about these). And so, more than a year after I reported what I personally witnessed, the persistence of this situation shows not only that Mozilla is broken, but also that users are far from its priority (if it's not simply unwilling to fix itself).

Going forward

All of this is not to say that Mozilla products are worthless, nor that no part of Mozilla can be salvaged. Many large free software projects struggle with separation of duties, decision-making and prioritization. No matter how one looks at these problems, ultimately they "just" indicate a governance issue.

So in fact, if there were no issues other than those I reported, Mozilla - the open source project which has more resources than any other - could fix these quickly. Unfortunately, beyond the governance issue(s) exposed, there seems to be an extra issue at Mozilla. What this debacle and the aftermath shows is not just that Mozilla needs to review its governance. It also suggests that Mozilla does not want to solve the issue(s). My report was not met with silence; in fact, several people replied. Some contributors genuinely tried to help a bit, and I must thank at least Svetlana Tkachenko for offering a significant and credible contribution. But no one has offered thanks for the report, or even acknowledged the issues. In fact, I believe most replies were - most ironically - claims that the report was off-topic on the only forum which it managed to reach (the governance mailing list).

In my opinion, this reaction is evidence that Mozilla is denying its governance issues. Mozilla is in its 15th year and would not be the first once thriving organization trying to hide or deny its difficulties now that it is struggling. So unfortunately, even if all of this "just" comes down to governance problems which could be solved with reasonable resources relative to Mozilla's size, "just" solving these issues from within an organization which denies them seems like a challenge beyond what the old and busy man I have grown into can reasonably tackle.

In any case, what it certainly means is that this will remain my last contribution to Mozilla on a volunteer basis, and I will no longer endorse any Mozilla product. My departure is in no way because Mozilla's mission is complete; despite all the progress accomplished in the last decade, the web and its standards have evolved a lot too and much remains to be done. I hope that other contributors to Mozilla or other people interested in Mozilla's objectives can either fix it or get involved in other projects which share some of these objectives.

As for Mozilla, if it wants to remain a project where its goals can be accomplished, it will have to review its priorities and put users, quality and transparency first. If the current situation is just the result of a lack of resources, Mozilla should focus on its own issues. Proving contributors that Mozilla products and processes could be trusted will be a lot cheaper than "rewarding" contributors, and much more effective in winning the loyalty of remaining and potential contributors. Mozilla, just make our work effective, so we will become proud of our accomplishments and willing to pay for Mozilla clothes ourselves. Stop offering us to pay for traveling to your HQ, start valuing our work, and those who want to meet will be willing to pay their tickets themselves. Stop putting the resources you have into producing an Internet health report; people expect Internet health reports from the Internet Society, not from Mozilla; what people expect from Mozilla is reliable software. If you want to help the Internet, focus on your core mission and heal your own systems before worrying about the health of the rest of the Internet.

It's unfortunate to have to make this decision right after Firefox regained its relevance, with Firefox Quantum and the "integration" of my favorite extension (Firebug), both in the previous 12 months. I haven't decided yet what this resignation will mean for myself. I am redacting this resignation in Mozilla Firefox. I already mostly migrated to Google Chrome, even though the product is not strictly superior to Firefox, but even though this is recent, I already have doubts about the Chrome project's management. As for my MUA, I really don't see any free software alternative at this point (I concluded that KMail was way too buggy a decade ago, and I am under the impression this has not improved since).

Farewell, Mozilla


The year after my resignation, I installed a new instance of Thunderbird to use with Gmail. Even though Thunderbird facilitated the process, and even though I took care to set the same settings as my previous instance, I realized 5 months later that I had hit issue #651945. Once again. So I went back to the ticket to see what I had missed.
What did I find instead? That the ticket was closed last year by no other than... Mozillain Wayne Mery. The ticket was marked as resolved with not a word on why, and someone gently pointed out that was incorrect nearly a year ago, but the ticket remains... RESOLVED.
Ah, good ol' Mozilla!

Timex T128 alarm clock: Imperfect, but Smooth

admin Saturday April 21, 2018

Sometime in the second millennium, Realistic manufactured a Chronomatic radio alarm clock, which my parents bought for me. Over 2 decades of use later, my reliable alarm clock had only started dying. It had lost its radio capability, but it could still be set to beep loud enough to wake up an entire room, no matter its size. Unfortunately, if the volume control could possibly set the alarm's sound in the past, that feature had broken long ago, so when I realized the prospect of being awakened by such a heart-threatening sound had perhaps become scary enough to make me lose sleep, I decided my oldest electrical device had earned the right to its ultimate sleep, while I had earned the right to buy a new, smooth alarm clock.

After a careful on-line search, I decided the Chronomatic's successor would be a Timex T128BC3, sold by Walmart for 25 CAD, part of Timex's T128 series. While this remains a fairly cheap device, after a few weeks of usage, I can say this purchase will have been worth it if the clock is still working as well in 20 years. The alarm can be set to be gradual, very smooth. The display is OK. The battery-based Sure-Alarm feature is well-designed, although it's disappointing to be unable to see the time when the power is out (there is no button to display temporarily). The nap feature is nice.

The weaknesses I found are:

  1. The DST control is a good idea, but it would save a lot more time to add a button for each digit to set times quickly, rather than simply time-consuming plus/minus buttons.
  2. Enabling and disabling the alarm cause a confirmation sound which is not incredibly loud, but still easily loud enough to wake up anyone else sleeping in the same room.
  3. The "24 Hour Set & Forget Alarm with auto repeat and auto shutoff" feature Timex brags about is worthless in practice. Unless you wake up at the same time every day (including weekend days), you'll still have to manually disable and re-enable alarms. If you wake up before the alarm time, there's no way to prevent just the next alarm. You have to disable the alarm completely (then re-enable it at night). Which makes weakness #2 considerably worst.
  4. Warranty of merely 90 days
  5. The feedback of buttons (ALM 1, "-", "+", ALM 2... all of them) is really bad. Very often, buttons do not actuate even when they have traveled down and are depressed. I did not investigate this deeply, but this seems to be related to button width. To ensure buttons actuate, you need to press in the center.
  6. The alarm seems to be buggy. Even though my clock is set to start smooth, twice its first 2 beeps were loud. Then it went back to quiet beeps and started its gradual increase.

Update 1: A few months after replacing my Chronomatic, I was shocked to notice that a woman my age was using the same alarm clock as I did. Hers had only started dying too. Wow, these Chronomatic clocks sure stand the test of time!

2018-10-13 Update: For the first time, I was awakened by a Sure Alarm during a power outage this week. I did not manage to stop the alarm without removing the battery, which is non-trivial (supposedly requires a tool, although I managed without). Either the button feedback issue combined with the lack of light and my lack of sleep that night explains my inability, or the clock failed to disable the alarm. I will try to update this next time Sure Alarm triggers.

Questrade's "Margin Account": A marginally considered nomenclature for a regular account

admin Saturday April 14, 2018

So you've joined Questrade and now want to open a main investment account to invest most of your money. It's not for an RRSP, or any other registered account. At the Account Selection stage, Questrade asks you which account type you want, and you can only see specialized account types. Questrade offers you a margin account, a managed account, but you fail to see a cash account. Could Questrade be stupid enough to not offer the simplest and most common account type?

Turns out the answer is negative. Questrade simply wanted to facilitate (and probably encourage) margin trading, so it made margin trading part of its regular accounts. So they named that account type "Margin account", only failing to specify, either in the name or in the account type's description, that using margin is completely optional, so that the majority which is just looking for a regular unregistered account should also use that type.

Unfortunately, that's not the only problem. Questrade also automatically allows you margin if you open a margin account. For example, although I just put the minimum 1000 CAD in my "margin account", I still have a 3330 CAD buying power. And that is not only the default setting; you cannot disable borrowing. So when you bid in your margin account, you'll have to consider "Cash" minus your open orders, not "Buying power".

Twitter's Bootstrap shall now bootstrap itself

admin Sunday April 1, 2018

Last Summer, as I was working on the Tiki project, which uses Twitter's Bootstrap framework a lot, I realized that Bootstrap is quite hard to discover organically and I decided to bite the bullet and go through Bootstrap's documentation.

The OK

Doing so, I hit quite a few issues, some of which I reported:

From these 4 initial reports, 2 were reportedly solved and closed. #23754 was a wide report which could be considered as reporting several problems. Several of these were fixed, and when the ticket was closed, I opened more specific/clear reports, #24705 Grid system documentation refers to undefined columns and #24706 Inappropriate paragraph in grid system documentation's "How it works" section ("Sounds good?"). As I could not verify the fix for #23769, I filed #24288 Indicate getbootstrap.com website freshness (documentation version). The only worrying case was #23772, which was closed without solution nor explanation, and which remains closed as of ½ year later. Since I couldn't reopen it myself, and since no one saw my request to reopen it, I filed a new ticket, #25225 Blockquotes example and others use cryptic spacing classes before their introduction.

The Bad

At that point, I had 4 new issue reports : #24705, #24706, #24288 and #25225. All of these were closed. Worryingly though, only #24706 was actually solved. The 3 other issues still persist, to my knowledge. Even though I asked to reopen 2 of them months ago, the 3 remain closed.

That's when I figured it was time to ask the mailing list whether Bootstrap contributors can expect serious treatment. And that's when things got ugly.

The Ugly

The next problem I hit was I could not find any official mailing list about Bootstrap development. When I realized Bootstrap's documentation did not refer to any discussion forum, I had to give up, and instead reported #25824 Documentation does not refer to discussion forums.

I suggest anyone who would consider contributing to or using Bootstrap to read that last report. The report was closed in less than a day, without any result being mentioned­. Not only did Bootstrap not get any official forum, but the documentation still fails to acknowledge that there is none.

Given the level of interest the Bootstrap project seems to have in getting forward, I am afraid the most helpful thing I can do is to stop wasting my time on it. This resignation shall be my last contribution to Bootstrap.

HP LaserJet Pro M227fdn (Debian GNU/Linux, Microsoft Windows)

admin Saturday March 31, 2018

After giving up on my Epson WF-3620 all-in-one, I had gone through a sixth consecutive defective inkjet printer. 6 printers of 3 different models, from 2 manufacturers, in 2½ years. Fearing loss of even more of my time and mental sanity by getting a seventh printer which would also fail near-instantly, I reviewed our needs and choices. Although I had always avoided laser because we print little (most of our usage is non-business), I decided to drop inkjet printers altogether and decided to replace with a laser all-in-one. After realizing that color laser printers were too massive, I decided to go with a monochrome HP all-in-one. After I gave up trying to add a M227fdw to my cart, I called HP and was told by an agent that was because HP had no such device in stock...

I then decided to go with a slightly less flexible but immediately available and considerably cheaper M227fdn (no touchscreen, no front USB port, no wireless). I thought I was lucky, since Debian stretch offers the very first hplip version which supports the M227 series. Setting up the printer was indeed easy after I remembered I needed to install the hplip package. I was impressed by printing speed.

Things got considerably worst with the scanner, which I connected via USB. I tried launching Skanlite, which refused to open with an error message. When I tried launching from the command line, it whined:

skanlite wrote:
sane_open(" "hpaio:/usb/HP_LaserJet_MFP_M227-M231?serial=VNB3J02222" ", &handle) failed! status = Error during device I/O

With such a helpful error message, I decided to try Simple Scan, which opened, but failed when asked to scan, complaining it was "unable to connect to scanner". After searching the web significantly, I was lucky enough to find the solution, given by another poor LaserJet user. The solution is to install hplip-gui, launch HP Toolbox, and install a proprietary plugin for that scanner (UPDATE: I can't get figure out how to do this in Ubuntu 19.04. I ran $ hp-plugin -i; instead.). While that solution does fix the problem, if the M227fdn does require a proprietary plugin to scan, it's easy to consider HP Linux Imaging and Printing's table of supported printer models, which lists the M227fdn as having "Full" support and as having "Scan to PC", as fraudulently misleading.

The LaserJet has a relatively powerful web interface, but reports cannot be viewed, only printed. Inexcusably, the 10 reports offered, including the usage report, can only be directly printed on the LaserJet, even with firmware 20180510. If you want to see how many pages you actually printed before buying a new cartridge, you can thankfully use Supplies Status in the Home tab. That will also give you the total number of pages printed (in fact, the number printed with your drum, which should support about 18400).

I thought setting up the scanner on Windows would be easy, but I was overestimating HP once again. After installing HP's software, I was somewhat able to scan using Microsoft's built-in utility, but I could not see a user-friendly application from HP. I tried the menu entry "HP TWAIN Administration". It is not clear whether that application is purely administrative or if it's supposed to allow scanning directly. The home screen says to click on "Numériser" ("Scan" in French), but there is no such button. The name suggests it's just an administrative tool, but I could not find where the presets controlled there show up.

So I went to the manual and found that after installing the software on Windows, there should be an "HP Scan" shortcut in the Start menu. Unfortunately, even though I installed 2 different versions of the software on 2 Windows 10 installs, none had such a shortcut.

Thankfully, I then noticed that in one session, HP did install an application called "HP Smart", which does allow easy scanning (among other things), and which works. Yet, it wasn't installed on the PC on which I used the latest installer, and it apparently only installs for the current session (you have to repeat for each user!).
A big applause for HP for such a failure... how Smart is that? It's HP Smart.

But, after all this time wasted, it seems to do the job. So here's hoping I didn't get a seventh consecutive defective HP all-in-one, and that this will be my last post about printers for a very long time...

Update 2019-01-02: After trying to copy a document, I must say this printer's interface is crap. I knew it would be less intuitive than a touchscreen, but having to use 4 different buttons to launch a copy, only 1 of which is labeled, caused 2 adults together to fail to copy a document until we gave up and consulted the user guide.

Update 2019-03-02: We printed 1979 pages on the original cartridge, but the output was starting to get gray rather than black. The web interface didn't show any fill percentage anymore (although a bar suggested 1% left). It had been several weeks since the printer started whining about low toner though, so I just adjusted the low toner threshold from the factory default 7% to 4% in the web interface's System tab (Supply Settings), since it's apparently quite annoying on Microsoft Windows. Changing the cartridge is not as easy as it was with previous printers, but it's straightforward enough once you know you just need to follow the pictograms, letting the printer figure out automatically it needs to initialize the cartridge.

Epson Blocked Printer Head Cleaning Kit review

admin Wednesday March 14, 2018

After my Epson WorkForce WF-3620 became essentially unusable, I looked for explanations, since this would be the third printer I bought which got badly defective in 2½ years. One plausible explanation was that the printer head was clogged, even though I had tried automatic printer head cleaning. The prospect of searching for a new printer model, setting up a new printer and scrapping the old printer with its partially filled cartridges once more making me sick, I decided to try manual printer head cleaning.

The only way to do that which looked reasonably easy was to buy a cleaning kit. And unfortunately, the only way to get one for my printer model seemed to be to buy 118Ink.com's Printer Refresh Epson Blocked Printer Head Cleaning Kit, for £10, even though I am in Canada, which means paying a total £20 (36 CAD) to cover shipping. Even though 118ink features a big logo which pretends that delivery is free. My sickness convinced me to buy the kit anyway.

2-3 weeks later, the package arrived and I tried it. The first deception was that the kit shipped is not the one advertised. The syringe is different, the bottle is different (in fact, you get 2 small bottles), the gloves are different, even the instruction manual is different. The gloves are not large - it took some time to put them on without breaking them, even if my hands are not large. And unexpectedly, most importantly, the blotting paper is different - it's not pages of paper, but 2 or 4 strips. As 118Ink.com recommends 2 strips per cleaning, that means barely enough to clean 1 or 2 printer heads once, even though 118Ink.com also recommends cleaning all printer heads when you clean one. So the kit doesn't even provide enough blotting paper to clean 1 printer once!

So did it work?

Even though I only had problems with the black cartridge, I followed their recommendation and cleaned all cartridges. I did not expect the journey would be so complicated, so I did not fully document what I did from the start, and I can't remember everything I did, but I must have cleaned each head 2 or 3 times.

Initially, only the black cartridge was problematic.
At a later point, magenta remained perfect. Yellow was good, others were worthless.
Later that day, black and cyan were close to perfection, but magenta was worthless, and there was no yellow left at all.
The day after, magenta and yellow improved a bit.
After 3 automatic cleaning cycles, black was very close to perfection still, but all colors had become worthless.
After 2 more cleaning cycles, yellow and magenta got perfect, cyan was as close to perfection as possible, but black had gotten worthless.
2 days later, black was back to near-perfection, but all colors had become basically worthless.
3 days later, after the black, cyan and magenta cartridges had been replaced, cyan was good, magenta was bad, and yellow and black had gotten worthless.
After 1 automated cleaning of all cartridges, the situation remained unchanged.
After 1 more automatic cleaning of the black cartridge, the situation was mostly unchanged, but black got half-OK.
After 1 more automatic cleaning of the color cartridges, the situation was mostly unchanged, but magenta got good.
After 1 more automatic cleaning of the black cartridge, the situation was mostly unchanged, but black got nearly perfect.
After replacing the yellow cartridge, the situation was still black near-perfect, cyan almost perfect, magenta good, no yellow at all, so we seemed pretty close to the target.
I then manual cleaned the yellow cartridge. In the next test, yellow was still worthless, but magenta had gotten closer to cyan than magenta (!), and black had degraded a bit.
After 1 automatic cleaning of the color cartridges, the situation was unchanged, but magenta has degraded.

So, after all of this (and much more which I did not document), my black cartridge was printing OK, the yellow cartridge was not printing at all, the magenta cartridge was not printing magenta and the cyan cartridge was printing OK. And then, when we tried to actually print monochrome, there was no ink left after a few lines. And the printer started failing with the helpful error code 0x97 (actually, it started doing that a few days before, while I was trying to clean). The screen said to look at the manual or the website for more details. The manual obviously doesn't say a word about any error code, and Epson's website apparently doesn't even mention that error neither. Bravo, Epson. I found one source which claimed 0x97 was a symptom with many possible causes, one being clogged printer nozzles. But oddly enough, even though black seemed to be completely clogged initially, printing didn't trigger that error.

After I ran out of blotting paper, I tried using regular sheets of paper. That doesn't work well at all. You should rather follow 118Ink's suggestion to use kitchen roll.

During all of these cleaning attempts, we had to change all cartridges. Because as soon as 1 cartridge is considered empty, the printer simply refuses to print. It doesn't matter if what you print uses the empty cartridge or not, you just have to replace the cartridge. And automatic cleaning uses lots of ink! So we started with some levels of ink, and despite all these replacements, we ended with approximately the same levels of ink.

At that point, I was forced to conclude I had failed. I had started with a printer which "only" had problems printing monochrome, and after paying 30 CAD for a kit which I waited for weeks, replacing all cartridges with original Epson cartridges, manually cleaning each printhead several times, letting the fluid in some heads for hours, and at least 10 automatic cleaning cycles, I was left with an even less usable printer.


The steps I tried clearly had an effect. Just not the effect I was looking for. Overall, the situation seems to "evolve" approximately randomly. Either:

  1. The kit is hard to use, or the instructions are incomplete.
  2. The kit is not designed for the WF-3620, even though there is no indication to that effect.
  3. My printer's problem was not clogged printer heads, which I doubt very much considering the effect cleaning has had.

Ignore 118Ink's marketing. 118Ink does charge for shipping. And despite the "90 DAY MONEY BACK GUARANTEE", 118Ink requires notification within 7 days of receiving items.

If your WF-3620 is in a situation similar to mine, and if you're not looking forward to an investment which will require more work than replacing your printer, over a longer period of time than replacing your printer, which will cost you more than replacing your printer, and end up with a printer which prints even less, having to scrap not only the printer but a cleaning kit somewhat specific to that printer, then, rather than buying 118Ink.com's Epson Blocked Printer Head Cleaning Kit, just replace your printer. That advice is probably true for other Epson printers too.

If you try using the kit anyway, you'll notice that the liquid bottle will float in the hot water container. I suggest you attach something heavy to it to stabilize it.

Winter sport helmets for commute bicycling (Marker Consort)

admin Wednesday February 14, 2018

The other weekend I went to a local sports shop and asked a nice merchant who daily commutes to work on bike even during winter about their helmets for winter sports. The guy had a Marker helmet and was satisfied with it. He offered me a ¼ price discount and convinced me to buy a Marker Consort 2.0 helmet for 105 CAD, to replace my usual bike helmet during winters.

While I'm happy I satisfied my curiosity and must say the helmet has interesting features, I returned it after trying it for a few days. The dealer highlighted these advantages:

  1. Quick clipping/unclipping thanks to magnets, with no risk of skin squeezing
  2. Better protection
  3. Temperature
  4. Visibility - attachment for goggles


The main disadvantage I saw was the lack of built-in position lights, particularly useful during winter, and the incompatibility with Mountain Equipment Coop's Turtle LED lights (which I already owned).

Unfortunately, after buying, I confirmed there is really no way to install Turtle lights on the Consort. Moreover, at least in my use case, which is often using my bike to reach a station where a bus brings me to the office, there are quite a few more small issues:

  1. While carrying the helmet in the bus, it is heavier and bigger. It is also a lot warmer, so it's not necessarily comfortable to keep it on your head in the bus.
  2. At destination, the risk of having the helmet stolen increases. It's more costly to replace, and it's more attractive for thieves.

Finally, the cyclist who sold me the helmet said the only disadvantage he saw was reduced hearing. While I can confirm that the helmet obviously makes it harder to hear traffic when ear isolation is maximal than in the summer, I did not use it enough to compare to the effect of the other headwear one would use with a regular helmet.

Not-so-great advantages

Additionally, even if I don't need to wear a tuque while I use the helmet, I still need to carry one for cases when I go out on foot at destination. And the attachment for goggles is imperfect; there is one attachment at the rear. That ensures the goggles stay attached to the helmet, but is insufficient to ensure the goggles stay on the helmet. While there are small dents intended to help hold the goggles, if you're used to putting your helmet upside down, the goggles won't stay on for long, unless - perhaps - you have compatible expensive Marker goggles. So unless you put them back often, the goggles will flap.

Another claimed advantage which is not that useful is the cold protection, which only helps with the top and back of the head, while the coldest part is the front.

As for advantage #1, clipping and unclipping were still not instant for me. I suppose you get quicker after a while, but clipping would still be complicated by a moving band for chin comfort which can easily get between the 2 parts.

This leaves better protection as the most interesting advantage. When I returned the helmet, my logic was that if a regular bike helmet is safe enough during summer, it should be good enough for winter too. After more thought though, this logic is significantly wrong because:

  1. The tuque or other headwear change the helmet's position, so sides are less protected. Unfortunately, I did not see any evidence supporting that (although I did not research).
  2. The risk of falling during winter is way higher.


Overall, Marker's Consort 2.0 is not a bad winter sport helmet. I wasn't convinced enough to use it instead of a regular helmet in my use case (which does not include any activity other than cycling), but it will surely be worth its price for others. If it had integrated lights, I would buy it.

Scenario for Advertising Proportional Representation

admin Sunday January 28, 2018

I was just reflecting on how I could promote proportional representation in governments using FPTP. My provincial reform movement unfortunately does not sell any gear. I was thinking a T-shirt picturing dis-proportionality could easily catch attention. It would be great to have a picture of my face with proportions changed so one size is 50% bigger than the other side on the front, and on the back, "Proportions matter - Visit democratienouvelle.ca".

I was then thinking it would be terrific to produce a short promotional video showing 2 voters inserting their ballot in a booth. A third voter arrives saying the path was icy, but the travel was worth it. The old lady is welcomed by the same 2 tellers, with the same authentic smiles, who invite her to "insert" her ballot in a trash bin. "Thank you madam, we value your opinion.", one teller says on an optimistic tone. "Be safe on your way back home.", says the other. The advertisement could conclude "If you are part of the numerous people who did not care to vote in the previous election, this time, consider voting for a change."

Meanwhile, the French-language "Rose nanane" video produced by Quebec's Mouvement pour une démocratie nouvelle will remain one of my favorite advertisements.

Marin Fairfax SC1 performance hybrid

admin Monday January 8, 2018

Last year I started working for an organization more than 10 km away from home. After asking my local dealer about the hybrid bike I had bought second-hand a decade earlier and being told it wasn't worth 100 CAD, I decided it was time for me to buy a new bike.

After much research, I settled on the first proposal from the dealer, a Marin Fairfax SC1. With a 489 USD MSRP, the model is inexpensive, but Wirecutter's review of performance hybrid bicycles convinced me it was a great choice.

The SC1 is a clear upgrade over my previous hybrid. For commuting, the worst problem was the quick release wheels and seatpost. I managed to get these replaced with attachments needing a hexagonal key for 15 CAD (not quite secure, but getting true protection was not affordable). I got Evo mudguards which cover plenty from the local dealer for 32 CAD and a new kickstand since this bike won't fit the usual model attached on the wheel's axis.

After a few months of usage, I was about to write I couldn't find any flaw in the SC1 when the rear brakes degraded. The same local dealer who had sold me the bike told me the brake pads needed replacing. I didn't get new pads yet, but the pads do look used, and I'm a little annoyed to be told I should replace these after less than 5 months of non-continuous usage, with well under a megameter traveled. In the end, the 1 weak point mentioned by the Wirecutter, brakes, which I thought wouldn't really matter, seems to constitute a significant con indeed.

But I remain very satisfied, and I must thank Wirecutter for its incredibly comprehensive and high quality guide.

Don't Quit Your Day Job... but Don't Trust DQYDJ

admin Sunday January 7, 2018

In 2016, Don't Quit Your Day Job... asked Start with us: how do you know you can trust DQYDJ?

In October 2017, I searched for historical data on ROI to convince my mother to be less conservative in her investment. The most interesting tool I found was DQYDJ's S&P 500 Historical Return Calculator.

When I presented the data to my mother, I realized the data did not make sense. I apologized, stopped, then looked at the data again with much focus. Eventually, I understood what the results mean. In fact, the statistics are even more useful than my initial understanding. But unfortunately, nothing in the tables nor in the accompanying text invalidates my initial interpretation.

I then showed the calculator to a brilliant friend who knows personal finance at least as well as I do, and asked him to explain his understanding of the results, without directing his answer. His interpretation was the same as the one I made initially. Undoubtedly, the calculator is misleading. And badly misleading. If the -6% 10-year minimum return is understood as the return of the worst year from 2008 to 2018, it is very far from the reality, which is -38% (2008). Making financial decisions based on this bad interpretation could easily lead to bankruptcy.

So I scrolled through the comments at the bottom of the page and found none reporting this issue. I was about to write one when I realized that the page did not allow new comments. In fact, it seems impossible to comment most DQYDJ pages.

So I went to the Contact Us page, which offered a form to send a message. I informed DQYDJ of their issue with comments through that form. Months later, the calculator remains misleading, and DQYDJ did not even reply to my message. And guess what - DQYDJ has been posting numerous new articles meanwhile.

I am sure that DQYDJ is actually trying to help its readers. This issue is clearly unintended. Unfortunately, finance is not a simple topic, and trying to help someone with finances can easily have the opposite effect in the end. DQYDJ's calculator is not wrong. If DQYDJ was targeted at economists, I would not be denouncing. But DQYDJ claims its audience includes beginners, and its calculator has the "Personal Finance" tag.

Investing its resources in creating new material while failing to give minimal care for existing material is an inexcusable management issue, and for that, my answer to DQYDJ's question is unfortunately that DQYDJ cannot be trusted.