blog.crowdedweb - Home

Discontinued

2008-01-19T19:01:00Z

I am no longer blogging long-form, if you haven’t noticed.

Blogging just sucks, and it’s not worth my nor your time.

tumbling is where it’s at, so check out tumbl.es, my rocking tumblog.

I’ll be taking this site down imminently. In the meantime, subscribe to the tumbl.es RSS feed if you want to continue to follow me.

tumbl.es

2007-12-12T00:05:00Z

This is a public service announcement.

I just wanted to point out, if you’re already reading this, that I am living through tumbl.es. While this, my somewhat long-form blog, is updated here and there, tumbl.es is living, breathing, day-in day-out. The fact that it’s a tumblelog makes posting for me much easier than here and I feel I have a lot to say through what I read, like, and am willing to annotate.

So check out tumbl.es sometime.

[Previous Article: EarIdeas and Podcasting]

EarIdeas and Podcasting

2007-12-07T04:40:00Z

Hugh McGuire, who founded the Librivox project, and his partner Chris Goringe have launched EarIdeas earlier today. I must admit, I don’t currently listen to any podcasts. About 8 or so months ago I did follow a few, mostly Coverville and the RoR podcast. I eventually lost interest, and I don’t really know why.

I was at first reluctant to write about EarIdeas, not having much to say about it. But then I read this post by Hugh on the EarIdeas blog. It’s entitled “why we chose web 1.0 (so far)” and in it Hugh writes:

Earideas is a bit of a throwback to the old one-way web, but it’s like that for a reason. I’m a big fan of bottom-up organization (del.icio.us, digg), like-minded group organization (flickr), user-generated-relevance and all the rest of it. But it seems to me – especially after our experience with Collectik – that part of the problem with that approach is that MANY people don’t have time, or just don’t want to do all that work. There are so many demands on our attention these days, that many people just want something they know is good, or rather: they want someone to make sure things are consistent. When you turn on NPR or BBC, you have an idea of what you’re going to get. Not so with a podcast directory.

This is a great observation and I find it really applies when it comes to podcasts. Podcasts, I find, are incredibly difficult to “skim.” Even videos are easier thanks to thumbnails and the fact that most big web videos (e.g., stuff that makes it big on YouTube for example) are singletons, that is single big hits and not part of a series. Most people don’t follow a series on YouTube, for instance. The podcast format on the other hand implies “series.” A podcast “episode” is part of a bigger long-running show. Because of the boldness of this format (it’s a demanding format on the listener—when we say “Bob listens to podcast X,” what we mean is “Bob is subscribed to podcast X”), I feel I need an editor or curator to make it easy for me, to skim through it and tell me what’s worth my time. Other users similar to me can help here through social features, but I prefer a fixed category classification system such as the one on EarIdeas for podcasts rather than open-ended “tags” like del.icio.us has for bookmarks.

And if there’s one person who listens to a lot of podcasts and knows a thing or two about online community, it’s Hugh. So I trust him as a curator and I trust the currently-12-category taxonomy. Let’s see if I discover something good…

[Previous Article: Facebook, Beacon, Conversion Tracking, Oh My! ]

Facebook, Beacon, Conversion Tracking, Oh My!

2007-12-03T01:29:00Z

There’s been a lot of mostly whining surrounding Facebook’s recently-announced Beacon program, which is technically just a conversion-tracking implementation with a twist. Let’s recap.

First, what is Beacon? As greg yardley explains, it’s nothing new technologically speaking and we in the online advertising space have been doing it for some time. For instance, what do you think Doubleclick’s Boomerang product does?

Boomerang is a one-to-one targeting solution for advertisers that allows marketers to identify users who have come to their sites and then re-target those consumers on Doubleclick’s Network, which currently reaches 40 million unique users per month.

Boomerang, and other similar technology, allow publishers to tell Doubleclick to label users following some event in such a way that Doubleclick, and presumably those sites which are its clients, can then later re-target these users with hopefully more relevant advertising. This is all fine if it is done responsibly, i.e., anonymously. It’s actually more than fine if it’s done responsibly, it’s pretty good because it allows for making advertising more relevant, and that’s very very good for both advertisers and consumers.

If you don’t agree with me so far, then this argument can’t go on. It can’t go on because fundamentally, if you think targeting and using third-party cookies to anonymously track behavior so as to increase the relevancy of advertising is evil, then you think that anything involving the use of third-party cookies to make more intelligent decisions when it comes to ad-serving is bad (maybe you think advertising in general is evil and bad?), regardless of whether or not the use is responsible. Sometimes I get the impression that all this recent backlash at Facebook has more to do with people expressing their frustrations with years and years of irrelevant advertising than an actual attack on Beacon’s implementation itself.

But assuming you do agree that cookie-based targeting can be done responsibly and is ultimately a good thing, it’s still worth wondering whether Facebook’s Beacon program is implemented responsibly. And this is what I think the argument should be about – not about who Facebook is setting cookies to, but rather, what they are doing with the data afterwords. And my understanding is that part of what they’re doing is associating some of that data with your FB user profile, going so far as to publish some of your transaction history via your newsfeed to all your contacts and friends without requiring an explicit and clear opt-in, and that, if it is true, is bad and irresponsible because it’s taking what is otherwise presumed to be anonymous data living in the private realm and making it personal data living in the public realm. That would be my beef with Beacon, not that it’s setting cookies on my machine.

Personally, I had hoped and speculated that Facebook’s “SocialAds” announcement, as it was called at the time, would be more interesting than Beacon turned out to be, both for publishers and advertisers, but I digress now…

[Previous Article: Web UI Polisher]

Web UI Polisher

2007-11-22T02:31:00Z

So I’m looking for a Web UI Polisher. What the heck is a Web UI Polisher? I just invented the title, but the description is this:

You know a little bit about designing something that looks good and is usable, and so consequently you at least know what does look good and is usable. You are flexible and can translate descriptions into a functional UI and implement, with polish. You know little tricks that make it not just look good, but look great, and you care about making things look great. You don’t just draw, though, you do. You know HTML, you know CSS, you know Javascript (or you’re on your way). You’re imaginative and pragmatic.

Are you such a person, or do you know just the person? Want to help me out? lemme know.

[Previous Article: Itojun]

Itojun

2007-11-09T22:41:00Z

A couple of weeks ago, on October 29 2007, Jun-ichiro “itojun” Itoh Hagino passed away at the age of 37.

To those in the BSD communities he was simply Itojun, best known in his role as IPv6 KAME project core researcher. Itojun did the vast majority of the work to get IPv6 into the BSD network stacks. He was also instrumental in moving IPv6 forward in all aspects through his participation in IETF protocol design meetings. Itojun was helpful to everyone around him, and dedicated to his work. He believed and worked toward making technology available to everyone. He will be missed, and always remembered.

I had the pleasure of meeting itojun at the first BSDCan conference in 2004, where both he and I were speakers. He was a really great guy and needless to say played a huge role in advancing IPv6 work, in particular making the KAME project possible.

Really sad news.

[Previous Article: Facebook and SocialAds Speculation]

Facebook and SocialAds Speculation

2007-11-06T15:05:00Z

Today Facebook is to announce SocialAds. Adweek Magazine says the following:

Full details of Facebook’s SocialAds are not yet known. The company declined to comment in advance of the meeting. According to sources, however, the aim of SocialAds is to find ways to embed advertising more naturally into the site and develop an ad network for showing placements on other sites. (As part of the deal it inked with Microsoft that included a $240 million investment, Microsoft sells the display ads on the site.) Facebook is expected to use the data available in user’s profiles to more carefully tailor offers served through its News Feed tracking feature. Those ads have already performed well, with some gaining click-through rates over 10 percent.

With Microsoft’s recent investment in Facebook, I’m predicting that SocialAds will be a competitor to AdSense. Facebook will launch an advertising network allowing publishers to embed Facebook SocialAds tags. These would deliver ads for advertisers which have the potential to be targeted based on cookies set by Facebook’s own ad servers and which will contain data allowing advertisers to target based on a combination of behavior and Facebook user profile information, or information from Facebook’s social network.

I really hope that’s what it turns out to be. If it works, advertisers will win and so will publishers. Google AdSense on the other hand… that remains to be seen.

UPDATE: So it turns out I was wrong with my prediction. At first glance I’m pretty underwhelmed by FB’s announcements. Adding to a person’s Facebook newsfeed from creative is interesting, but not that much; and I worry about the increasing irrelevancy of the FB newsfeed, especially as it becomes populated with more and more stuff.

[Previous Article: When prototype.js is Inappropriate]

When prototype.js is Inappropriate

2007-10-06T16:26:00Z

[Fjord is hiring! Check out Web Enthusiasts Need Apply and Web Enthusiasts Need Apply, part 2 for details.]

I like the prototype.js javascript library. It’s a great and convenient library for web applications, particularly when used in conjunction with Ruby on Rails which in fact ships with it as well as script.aculo.us, a visual effects library built on top of it. I also liked Gary Haran’s talk at Montreal on Rails the other day, Gary’s a great technical speaker and has a knack for explaining things simply and concisely.

With respect to the use of prototype.js however, there are some caveats. In particular, it shouldn’t be used if you’re writing widgets or any javascript that you would like others to embed on another website. For instance, web applications meant for mashing up. There are 2 reasons:

prototype.js is opinionated and doesn’t try to nor encourage namespacing, so co-existence with other frameworks/libraries. If you’re building something for inclusion in another site where you cannot control the choice of javascript conventions used, you are potentially breaking the rest of the site.
prototype.js adds methods to the prototype of Array (javascript’s associative array), and there is code out there (like e.g., the site where your code might be included) that relies on being able to iterate over elements of an associative array by just doing “for (var element in my_hash) { ... }”. If you really want to see the potential impact of this and you’re feeling particularly masochistic, open up script.aculo.us in Firefox, turn on Firebug, and type and run the following snippet in the console (note that the script.aculo.us site loads prototype.js):


var foo = new Array();
foo['blah'] = 'blahaaaa';
for (var element in foo) {
    alert(element);
}

There is also a third but in my opinion less important reason: prototype.js defines $ (the dollar sign) and $$ (the double dollar sign), but according to the ECMA-262 on page 14, “the dollar sign is intended for use only in mechanically generated code.” Personally, though, I think that ignoring the standard here is fine (it’s not a very clearly written standard to begin with).

Also, since prototype.js has hit the mainstream, people seem to increasingly believe that it gives you classes and inheritance. Only half of this widespread belief is actually true. First, while it does give you the notion of a “Class,” all you’re really getting is the ability to define the constructor code inside an “initialize” method defined in the class’ prototype. For instance, here’s how I’d define a prototype.js style Class without prototype.js in pure javascript:


var MyClass = function() {
    this.abc = 1;
    this.def = 2;
}
MyClass.prototype = {
    methodone: function() {
        this.abc = 5;
    },
    methodtwo: function() {
        this.def = 6;
    }
}

With prototype.js, the above code looks like this:


var MyClass = Class.create();
MyClass.prototype = {
    initialize: function() {
        this.abc = 1;
        this.def = 2;
    },
    methodone: function() {
        this.abc = 5;
    },
    methodtwo: function() {
        this.def = 6;
    }
}

Second, the Object.extend() that prototype.js defines is not really an implementation of inheritance. Rather, it is an implementation of object augmentation. Object augmentation is when you have an object, like for instance MyClass.prototype from the example above, and you’re copying key value pairs from another object into it, so that MyClass.prototype is then augmented by that other object and, consequently, objects constructed with “new MyClass()” get access to their newly augmented prototype. What this doesn’t give you, though, is the ability to chain method calls across the inheritance chain. In an actual inheritance model, you’ll really have two separate objects with references from one to the other so that you can optionally invoke methods of the parent from the child (this is what “super” does in most OO languages).

Now is the lack of real inheritance a big deal? Perhaps not. Perhaps you can get away with just object augmentation in the vast majority of cases. But the distinction between inheritance and object augmentation is more than academic. For example, when you’re using augmentation you must make sure to not augment an object without being aware of every other piece of code that is augmenting that same object, otherwise you risk overwriting someone else’s freshly added methods. In a true inheritance model, you can relatively safely inherit from the same object from many different places without having to worry about overwriting methods defined by another inheritor (is that even a word, inheritor? :-))

It is worth noting that Yahoo’s excellent BSD-licensed (free!) and extremely well documented YUI library not only cares about namespacing (and so co-existence with other libraries), but also gives you both YAHOO.lang.augment() and YAHOO.lang.extend() which do object augmentation and prototypical inheritance. Since we’re on the subject, I’d also highly recommend the following Douglas Crockford talks from the YUI Theater: “The JavaScript Programming Language,” “An Inconvenient API: The Theory of the DOM,” and “Advanced JavaScript.” All of these, and more, are available for free at the YUI Theater.

[P.S.: Fjord is hiring! Check out Web Enthusiasts Need Apply and Web Enthusiasts Need Apply, part 2 for details.]

Web Enthusiasts Need Apply, part 2

2007-10-03T03:35:00Z

As I wrote previously, Fjord Interactive is looking to hire an ambitious web developer. If you know someone who fits the bill, please let me know.

In addition to the above, we’re also looking for a talented UI implementor who knows how to make things happen and who’s got a good eye for visuals. Knowledge of Javascript is an asset but not an absolute requirement (we’re willing and can afford to teach). This person would join an existing small two-person team of web UI integrators and would work on both Rails and .NET projects alike, from websites to applications.

To get an idea of what we’re about, check out my previous job posting.

Once again, if you are or know someone who might be interested, get in touch with me.

Web Enthusiasts Need Apply

2007-09-18T02:54:00Z

Fjord Interactive is the interactive advertising division of Cossette, Canada’s largest independent advertising agency. Fjord is looking for a talented and ambitious mid-level software developer particularly interested in the web and web application development.

As a web developer at Fjord, you will be working with a small core team of developers in Montreal on applications and platforms both large and small. Fjord develops and maintains the AdCentric adserver software (which runs Fjord-produced campaigns) as well as a slew of other delicious bits and bytes. We’re avid supporters of and part-time contributors to open-source.

The happy looking people in the picture above are not only software developers – Fjord is also art direction & design, account management, strategy, and business solutions.

We’re also happily a part of the larger Cossette community. Feel free to get in touch with us on Facebook.

Things you would ideally be familiar with (to various extents): ruby, rails, unix, gentoo, perl, C, open-source projects of sorts, del.icio.us, the Atom publishing protocol, HTTP, REST

If you’re interested or know of someone who is, please let me know.

/clear

2007-09-17T02:35:00Z

Ah, refreshing.