When I got my hands on the DEWALT Ratcheting Screwdriver, it took me 5 minutes to tell whether I would use it as a multi-bit screwdriver. The weird mountable rack it came with made me wonder if it could store all of its bits inside the handle. After I decided to try it, I opened the package and put the bits in the internal bit storage. After putting 1 or 2 bits, I struggled to put the next in place. At first, I thought there must have been a manufacturing defect. But after trying more slots and bits, and considering the mountable rack, I started to smell a design error.

And indeed, the online reviews are poor. This review of a variant of mine says it all. Visibly, Dewalt didn't intend to have internal storage initially and rushed a broken solution. Or, it tried to patch its faulty internal storage with the mountable rack.

In any case, if you are looking for a multi-bit screwdriver, do the world a favor and do not encourage DEWALT by buying its ratcheting screwdriver.

SAP SQL Anywhere and its GUI Sybase Central (now SQL Central) have been part of my 5 most important tools for 4 years now. A few times, I noticed some oddities, but it wasn't until yesterday that I realized what was going on.

First, Interactive SQL starts transactions which are not automatically committed. By default, commits happen when closing or disconnecting (see the SQL Anywhere tab in Options). There is no warning to that effect. Second, the default SQL Anywhere isolation level for database transactions (at least in versions 11 to 17) is 0, "read uncommitted": http://dcx.sap.com/index.html#sqla170/en/html/3bf0ef526c5f1014b372935839b121fc.html
Which means that the changes from uncommitted transactions are all visible, from different sessions and even from different users. Which is why I did not realize how Interactive SQL behaved before now.

So far, this may be surprising, but perhaps nothing incorrect. Unfortunately, this comes with a ultimate surprise. When you execute a query in Interactive SQL in these default configurations, the transaction started is not committed. And after some time, it appears that the transaction is quietly cancelled, automatically. I reproduced this at the first attempt. If the abortion was indeed triggered by a timeout, I do not know how long the delay is, but it is inferior to 16 hours. The Interactive SQL window was still open, there was no popup, nor any indication that the transaction had expired. The Interactive SQL closed normally, without indicating that the transaction had timed out.

This happened with Sybase Central 16 and SQL Anywhere 11. Commit after every statement will workaround this, but will not cause automatic commits when modifying with the results table (either modifying cells or deleting rows).

Transactions are great, but when implemented properly. Thinking about all the manipulations which we must have lost due to this bug leaves a taste which is quite bitter, and certainly not ACID.

Based on David Carr's excessively simple technique to duplicate form sections using JavaScript, here is my enhanced but still minimalist take, adding support for:

1. select multiple
2. starting the form with several sections already
3. easy backend usage

This changes indices so they start at 0.
Duplicate Form Sections

//define template
var template = $('#sections .section:first').clone();

//define counter
var sectionsCount = jQuery('div.section:last').data('number');

function incrementIdentifier(id, sectionsCount) {
	if (id.substr(-1) == '0') {
		id = id.substr(0, id.length - 1);
	}
	return id + sectionsCount;
}

//add new section
$('body').on('click', '.addsection', function() {
	
	//increment
	sectionsCount++;
	
	//loop through each input
	var section = template.clone().find(':input').each(function(){
		//set id to store the updated section number
		newId = incrementIdentifier(this.id, sectionsCount);
		
		//update for label
		$(this).prev().attr('for', newId);
		
		//update id
		this.id = newId;
		
		var baseName = this.name;
		var array = false;
		if (baseName.substr(-2) == '[]') {
			array = true;
			baseName = baseName.substr(0, baseName.length - 2);
		}
		baseName = incrementIdentifier(baseName, sectionsCount);
		if (array) {
			baseName = baseName + '[]';
		}
		this.name = baseName;
	}).end()
	
	//inject new section
	.appendTo('#sections');
	
	jQuery('input[name=lastSection]').val(sectionsCount);
	return false;
});

//remove section
$('#sections').on('click', '.remove', function() {
	//fade out section
	$(this).parent().fadeOut(300, function(){
		//remove parent element (main section)
		$(this).parent().parent().empty();
		return false;
	});
	return false;
});

As a bonus, the PHP function I quickly crafted to ease processing the resulting $_REQUEST:

<?php
/**
 * Treat $_REQUEST when working with duplicate-form-sections.js
 * $_REQUEST['numberedParameters'] contains the base name of the parameters to be converted into arrays
 * $_REQUEST['lastSection'] is the number of the last section (the number of sections minus 1)
 * Alters neither $_GET nor $_POST
 * @author Philippe Cloutier
 * @license MIT
 * @return boolean false on failure
 */
function numberedParametersToArrays() {
	if (! isset($_REQUEST['lastSection'])) {
		return false;
	}
	$numberedParameters = $_REQUEST['numberedParameters'];
	$lastSection = $_REQUEST['lastSection'];
	foreach ($numberedParameters as $numberedParameter) {
		$_REQUEST[$numberedParameter] = array();
		for ($i = 0; $i <= $lastSection; $i++) {
			if (isset($_REQUEST[$numberedParameter . $i])) {
				$_REQUEST[$numberedParameter][] = $_REQUEST[$numberedParameter . $i];
			} else {
				$_REQUEST[$numberedParameter][] = array();
			}
			// We could unset the numbered parameters.
		}
	}
	return true;
}
?>

And here is an adjusted usage example:

<form>

<div id="sections">
  <div class="section" data-number="0">
    <fieldset>
        <legend>User</legend>

        <p>
            <label for="firstName">First Name:</label>
            <input name="firstName0" id="firstName0" value="" type="text" />
        </p>

        <p>
            <label for="lastName">Last Name:</label>
            <input name="lastName0" id="lastName0" value="" type="text" />
        </p>

        <p>
            Friends:
            <select name="friends1[]" id="friends1" size="2" multiple>
                <option>Alice</option>
                <option>Bob</option>
            </select>
        </p>

        <p><a href="#" class='remove'>Remove Section</a></p>

    </fieldset>
  </div>
</div>

<p><a href="#" class='addsection'>Add Section</a></p>
<input type="hidden" name="numberedParameters[]" value="firstName">
<input type="hidden" name="numberedParameters[]" value="lastName">
<input type="hidden" name="numberedParameters[]" value="friends">
<input type="hidden" name="lastSection" value="0">
</form>

Note that there are similar solutions elsewhere. I am not saying this one is better than these.

You know part of your life is behind you when Debian is dropping support for your first computer (as well as your second). I can't help but feeling nostalgic when thinking about my good old Pentium 120 MHz (80502) and my AMD 450 MHz (K6-2). Ah, the good old times when CPU-s were the stars. Today, I do not even remember the model or frequency of the computers I work on, even though I chose them alone.

That being said, I'm happy Debian is finally making that move.

Tangerine bank started promoting its new Money-Back Credit Card several months ago. Wanting to avoid spending time on anything "beta", I did not request a card until, last month, I fell under the impression that the card was ready.

The ordering process was not bad, except for asking you to make a ridiculous choice between purchase categories granting extra money-back rewards. The process lets you define your credit limit.

When the package arrived, I was surprised to find the contract was printed on 4 unbound sheets of paper. But even more surprised to see the card's design:

Out of 4 adults asked, a single one could tell the right direction in which the card must be inserted (which is chip first, of course).

The 2016 Debian Project Leader election has started. For a second time, its result is quite predictable, as there is a single candidature.

After a single term, Neil McGovern has not offered himself again. A few years ago, I wrote about governance issues in my description of my involvement in Debian. Although elections are infrequent and the number of candidates has never been very high, the following table seems to show a downwards trend:

While I believe Debian should use liquid governance, I suggested that distributing the DPL role (using for example a Second in Charge) would help meanwhile. Some may argue that a second single-candidature DPL election this decade implies that candidatures for the 2IC role would be rare, since so few contributors are interested in even the DPL role.

There may be several factors behind the diminution of candidatures, including a diminution of activity. But it should still be observed that Software in the Public Interest, which is a lot less active than Debian, manages to find 8 directors. So I have an alternative interpretation for the low number of candidatures in past years.

In my description of my journey in Debian, I showed that half of Debian Project Leaders do not propose themselves again after their first term. Veteran developers know that becoming DPL often means burnout and the end of one's Debian involvement. So I propose that the decline in DPL candidatures is not due to the role's importance having diminished, but rather a realization of that importance being excessive (in other words, excessively demanding for the (null) compensation).

I am not convinced this year's candidate has as much maturity as the DPL role would call for. But even him seems aware of the DPL's importance, as he shows at the very end of his platform:

I am disappointed to see his platform does not propose any major improvement to governance. Which does not mean things cannot improve a bit:

Hopefully, by distributing the role of spokesperson, the DPL can get a bit more time for decision-making.

Updates

2016-04-16

The leader in office at the time of publication included the following in his electoral platform:

Neil's Final bits from the (outgoing) DPL was only his third "bits" mail. And it contains:

While the above is very revealing, I encourage leaders who will leave the office to provide an actual estimate of the amount of time they dedicated to the office, as well as an evaluation of how well they think they did.

2016-08-21

After Neil dedicated most of his A year in the life of a DPL talk about his DPL term to a single issue (ZFS and its Common Development and Distribution License), he was asked by his predecessor about the time spent by DPL-s on licensing matters (at 33:05). Neil's answer eventually strayed from the question to discuss the DPL role and concluded with:

2017-04-02

The 2017 election is quite comparable to 2016's. There are twice as many candidates this year - 2 - but no more focus on governance. Mehdi Dogguy's platform is unchanged regarding governance. Mehdi also doesn't estimate how much time he committed to the role in his first mandate. As for Chris Lamb, his platform does not have 1 word on governance.

2018-04-10

This year again, a single candidate is walking for leadership. The candidate, the current leader, presents a new platform which still has no word on improving governance, but which now discusses governance by describing his first term:

While I congratulate Chris for describing his activity probably more than any other DPL, and while I know very well how painful time-tracking is, I deplore that once again no estimation of the time required/invested during the current term is published.

Designing Web forms can still be a pain today. I thought HTML version 5 and its tel type input field could replace the complicated triple text fields used by our web tools. But due to lacking global phone number standardization, tel is quite useless. So I ended up integrating the jQuery Mask Plugin, which adds characters in a plain text field so "1234567890" is turned into "(123) 456-7890", or the format you want ("(000) 000-0000" is called a mask).

Everything seemed fine until I realized the plugin was pushing the masked values to the server. If the server had pushed a customer's current phone number as "1234567890", which was then masked by the plugin, it would come back to the server as "(123) 456-7890". In August 2015, raphaelm22 requested his solution to be adopted. Igor Escobar, the jQuery Mask Plugin's author, commented that he appreciated the "suggestion but [did not] belieave that [raphaelm22's] decision matches with what [he hoped] in terms of future of jQuery Mask." And without further explanation, he closed the ticket.

I took that occasion to properly report the problem raphaelm22 was trying to solve in ticket #371. And 2 days ago, Igor Escobar closed that ticket, without any explanation.

Dear Igor, if you read this, I am thankful to you for writing the plugin, but for authoring such a bug and for pretending it does not exist… I am hereby enshrining you into the infamous FHoS.

As always, admission into the FHoS is best considered as an incitation to leave it. In this case, you will have yourself expelled from the Hall by simply solving this issue, either by a code fix or by properly documenting it. I do hope your solution will be better than that proposed by raphaelm22.

Meanwhile, we are using the following ugly workaround at work, similar to raphaelm22's suggestion:

jQuery("form").submit(function() {
      jQuery('input.telephone').each(function() {
            jQuery(this).val(jQuery(this).cleanVal());
            });
      });

If you know of a fork without this issue or an equivalent plugin, please leave a comment to let me know.

TODO: Evaluate https://github.com/digitalBush/jquery.maskedinput
See also http://wiki.jqueryui.com/w/page/12137996/Mask

In January, I announced my "resignation" from Debian's public relations. My last contribution to the team was a few fixes and other improvements to its wiki page. All of these changes were integrally reverted by an editor who used the helpful edit summary "remove changes from chealer".

Those who have worked with me know how I deal with troublemakers. I am as respectful to them as they are respectful to their colleagues. I reverted the editor with the equally helpful edit summary "uh... I guess "remove changes from x" shall be a good enough explanation". The editor then again integrally reverted my revision with an even worse edit summary: "hi chealer, please do not make edits in this page.".

At this point, you may think this is just another episode from an open wiki. And you would be right, except for the fact that the editor in question is no other than Ana Beatriz Guerrero Lopez, who has been a Debian member for a decade, and who is one of the 4 team members recently "appointed" by project leader Neil McGovern. I was already contributing to Debian KDE when Ana joined the team, so she knows very well who I am.

All of that explains why I went as far as creating a discussion page for the team's page before going further. But that still wasn't enough to get Ana to try justifying what she had done. She would still "[revert] unwelcome changes". Not only that, but by the same reversion, she removed the link to that discussion page. Appropriately in a sense, since she also deleted the discussion page.

Thankfully, that latest reversion was not integral. Apparently, she finally accepted most changes. Those she still rejected may be an indication of what really bothered her.

To be fair, I am not the only former member Ana removed. Her very next edit removed the current members and emeritus members sections, commenting that "there is not[sic] such thing as "emeritus" members and everybody belongs to the publicity team". I certainly agree with the former statement. Congratulations, Ana, for being the first to have the guts to make that statement. Unless you were simply the first one to need such an edit, to justify your previous edit. As for everybody belonging to the publicity team, if the delegation was not a coup, and if "The listing below includes regular contributors also.", it is strange that this listing only includes the so-called "DPL-delegated members". And if your co-delegate found it important to have all members listed on the page, is it consistent that all of those who added themselves at her request (Thomas Vincent and yours truly) have been removed by your edits, just a few months later?

In the end, the answer to those questions does not matter. That we have to speculate is enough to diagnose a communication problem. And since this is not the first time you have serious issues communicating, I have the dishonor of hereby enshrining you into the infamous FHoS.

Your last edit to the wiki page is surely just the tip of your contribution to Debian. I have not followed your activity closely, but I am convinced you are a useful contributor. But this is no excuse to undo other people's work and to be incapable of civil discussion after a decade of participation in a huge project like Debian, at the risk of demotivating your colleagues.

Do not despair. The FHoS is one very special hall, which you can (and are encouraged to) leave. This is no more a condemnation to permanent humiliation than a challenge for the continuation of your contribution. Demonstrate that you have understood the importance of communication and I will have the true honor of removing you from the FHoS. Have the Debian wiki become a real discussion platform, so that no one can do something quite as bad again. Show me you got it to run a discussion system a la Flow, and I will replace this post with one singing your praises.

I was in first grade the first time I had something stolen from me. In fact, the stealer took Micro Machines I had temporarily exchanged with my electronic helicopter, so the owner of the stolen Micro Machines did not give me back my helicopter, which produced sounds when its buttons were pushed. I remember the poor child I was regretting that loss years later.

Since then, the only things which were stolen from me are 2 winter bikes (i.e. cheap bikes), a seat-post and a wheel. And only one of these thefts involved breaking a security device. Eventually, I realized I have always been lucky, in particular on that front. In recent years, I realized how lucky in particular I was for having never lost a wallet.

On January 23rd, I went to the gym. I put all my stuff in a locker to go warm up swimming. After a shower, I arrived at the pool and noticed it was unavailable. I took a couple minutes to memorize the schedule, went back to the locker room, dried my hair and went back to my locker about 15 minutes after leaving it. I noticed half of my stuff was missing, and realized it was not my lucky day.

I wondered if the stealer would bring back some of the stolen stuff he could not make use of, but the middle-aged security agent who received my deposition said he had never heard of such luck. While I was waiting for footwear and clothes needed to go back home, I realized how much time (and money) it would take to recover and buy again the more than 20 items I had been stolen. I asked 2 or 3 times whether some of my stuff had been recovered, without success.

At least initially, I even felt that some of my trust in others had been stolen. I was not done dealing with that mess when, 23 days after the theft, I received a call from the gym asking me to come get my stuff, which they had picked from a locker in a different locker room than mine 2 weeks earlier. I was so shocked that I failed to realize that they did not realize what had happened. They had cut the lock because users are not allowed to permanently let their stuff in a locker, and they had been waiting for 2 weeks for the owner to come claim his belongings. They did not realize it was not the owner of those personal items who had let them in that locker, which is why they did not feel compelled to contact me earlier.

When I claimed my stuff, I confirmed the attendant was right. Everything I had been stolen was there, except for my lock. Either that lock was cut by the stealer, or—much more likely—I had forgotten to lock it and the stealer used it to lock my effects in the locker where he left them. Everything else was there, including payment cards, gift cards, 170 CAD (which is more than I estimated I had been stolen), 2 controlled keys, and various other personal items worth a few hundred dollars.

I wish I could confirm I had forgotten to lock my locker, and I wish I could understand why. But I am a lot more curious about the stealer's behavior. Did his relatives ask him to bring back my possessions? Did he realize who he had stolen and decide the victim did not deserve this? Or did he simply change his mind fearing the consequences of getting caught?

Recovering my effects is far from having compensated the trouble of canceling and obtaining new cards, buying replacements, dealing with the risk of having a stealer with the keys of several buildings and possibly my address, temporarily bringing down my website, and more, but a lot of that could be blamed on my gym. The stealer involuntarily taught me a few important things about security, as well as how one's vision of his fellow citizens can easily be influenced by anecdotal events of one's history. For these lessons, for my goods of course, but also for your restitution's large restoration of my trust in mankind and my desire to work for it, I thank you, mysterious stealer.

I should use this opportunity to thank Vincent Amyot, who caught me trying to steal a pog when I was in sixth grade, and rightly warned other kids at school to stay away from me because I stole pogs (one by one!). And thanks to Guillaume Talbot, who caught me stealing his Internet account (in the good old time of dialup) and The Palace account (although he then remained my best friend for years). And thanks to the RFID tag I could not see which prevented me from stealing a magazine about Microsoft Windows from my college's library, an attempted theft which the free software zealot I was at that time considered ethical. You all made me realize I was not good enough to be a robber, and made me a bit less uncomfortable for having the privilege of being the victim of such a good stealer.

C'est bien connu que notre vie dépend de nos organes.
C'est aussi bien connu que nos organes sont moins nécessaires après la mort.

Pourtant, à chaque fois que je reçois une carte d'assurance maladie, on me dit de déclarer mon consentement au don d'organes. Cet effort serait si nécessaire qu'on tente de me persuader à coup de marketing. « Signez don », dit le pamphlet. Je suis bien prêt à faire un effort pour le don de mes organes, mais pourquoi devrais-je soit…

• Apposer un autocollant sur chaque nouvelle carte et le signer, tout en avisant mes proches de mon consentement
• ou envoyer, par la poste, un formulaire de consentement

Si les beaux chiffres (comme les 8 vies potentiellement sauvées) que le gouvernement avance pour inciter au consentement sont vrais, la présomption de consentement n'aurait-elle pas autant plus d'intérêt?

Ce que je serais prêt à faire, c'est de demander la présomption de consentement. Mais la pétition 411 est fermée depuis des années. Le gouvernement y a même répondu, rejetant en prétextant que le consentement présumé dans d'autres pays a donné «un résultat mitigé». La réponse mentionne « la possibilité d'effets pervers, comme une perte de confiance du public dans le système de prélèvements », prétendant même que le consentement présumé « pourrait, à la limite, aggraver la pénurie d'organes ». Mais quelle confiance du public dans le système de prélèvements? Et comment le consentement présumé pourrait-il diminuer une telle confiance?

Je suis désolé si mon boycott vous condamne à la mort, mais en réalité, ce meurtre aura été commis par le gouvernement du Québec. Certainement pas par moi, car je serai mort.

Vous pourrez au moins vous conforter en pensant que vous ne serez privé que d'une vie dans une société encourageant l'égoïsme, et avec plus de considération pour ses cadavres que pour ses membres. L'altruisme est généralement plus coûteux que l'égoïsme. Mais sommes-nous condamnés à ce que l'altruisme soit toujours plus coûteux?

Je ne ferai pas de déclaration officielle de consentement tant qu'il me faudra le faire plusieurs fois ou par la poste, mais j'autorise par le présent paragraphe que mes organes soient donnés à ceux qui boycottent le don d'organes à refus présumé, à ceux qui ont consenti au don d'organes et aux mineurs. Sauf pour mon cerveau (que je réserve à notre gouvernement brain-dead, qui en a tant besoin d'un nouveau, même si le mien peut sembler sans cœur).

En espérant que ma foie dans ce boycott ne sera pas vaine et qu'il me permettra de voir le consentement présumé (tel qu'il existe en France) au Québec de mon vivant. Et pourquoi pas avant la baisse anticipée des opportunités de dons d'organes?

Mise à jour 2017-05-14 : Ma carte expire à nouveau. Aucun changement de situation, et toujours aucune action de ma part pour devenir complice de cet ignoble massacre. Nouvelle insulte : le formulaire qu'il faudrait remplir réfère au site de Transplant Québec avec une mauvaise adresse. Gros progrès!