Links in New Windows: Websites vs. Applications [en]

[fr] En 2011, c'est toujours pas bien de forcer les liens à ouvrir dans une nouvelle fenêtre. Sauf, on peut l'admettre, quand votre site est en fait une application web.

I was surprised recently to realize that in 2011, it wasn’t obvious to everybody that forcing links to open in a new window (with target="_blank") was not a good idea.

Actually, there are quite a few high-profile sites which force links to open in new windows, and I realized I actually don’t mind it that much: Twitter is one. I like that when I click on a link in Twitter, it opens in a new tab. I’ve learned that Twitter does this, and I now rely on it.

So, maybe links in new windows aren’t always a bad thing?

Here’s what I think: if your “site” is in fact an application, then it doesn’t matter much. People will learn to use the interface of your web application, and if links open in new windows, they will discover that and (hopefully) remember it. However, if your site is a real site, meaning it contains stuff that people are going to read, and that stuff might contain links to other stuff people might want to read or see, then it remains a Bad Thing.

Why is it a Bad Thing to force links to open in a new window (in your blog, for example)? Here’s the reasoning behind this.

  • Normal behavior is that links open in the same window.
  • When you force a link to open in a new window, you’re breaking that expectation — and there is no way to know, by looking at the link, that it’s going to spawn a new window.
  • Opening new windows is a user decision, not a website design decision. Windows are part of the browser, not the site.
  • A link can easily be opened in a new tab or window by holding down a modifier key before clicking on it (Cmd/Ctrl for example).
  • A link which is set to be opened in a new window cannot be opened in the same window if that’s what the reader would prefer.
  • Opening links in new windows may confuse the user (who might not notice the new window) and breaks the back button (to go back, you have to close the current window instead of hitting the back button — adding a different way to “go back”… more confusion).

Summary:

  • if your site is actually a web application, where links open is part of the application design, and forcing links to open in a new window can make sense in certain situations;
  • if your site is a “proper website” or a blog, don’t force links to open in a new window — where they open belongs to the way your reader chooses to use his browser, and not to the website design.

Similar Posts:

Blog, What Happened to You? [en]

When I’m asked what the difference between a blog and a website is, I usually make this drawing to explain it.

Difference between a blog and a non-blog website

It’s not perfect, but it helps. With a “traditional” topic-based website, you have a site structure which looks like a tree, with different pages on different topics. With a blog, you have a succession of posts organized chronologically (inverse chronologically, actually) on one page. Then each post has its page, and it’s archived forever in the back-office.

The two models tend to blend — more and more sites have characteristics of both.

There are two trends, however, which irritate the hell out of me. (If I know you and you’re doing this, please don’t take it personally — I don’t hate you for it. Really. But it annoys me.) They are:

  • the blogazine
  • systematic teasers or partial posts on the main blog page.

Prepare for the rant. I’m putting on my flame-proof underwear.

Blogazines

First of all, let me say that there is nothing wrong with making a magazine with a blog CMS. But Lord, why do blogs have to try to pretend they’re all magazines? It feels like bloggers are trying to make themselves look “high-profile”, because top “blogs” like TC, RWW, etc. are actually magazines. They might have started out as humble blogs, but they are not anymore.

“Media-blogs” are a special breed of blogs. Their content is there to generate revenue directly, through advertising and sponsorships. That has an impact on their content, and on the place they try to occupy, alongside old media. Why would everybody want to look like one? Dressing like a movie-star does not make you be one — and why would everybody want to be mistaken for one? If you’re a geek or a businessman or an entrepreneur, why don’t you just be that? There’s nothing wrong with being yourself and making you approachable.

There’s nothing wrong with having a blog that looks like a blog.

Coming to practicalities, there is a real concrete reason for me, as a user, to not like it when one of the blogs I read turns into a blogazine: very often, this transformation goes with the disappearance of the “main blog page”, the page which gave blogs the place they have in the publishing world of today, the unique stable page which you could go to at any time, confident that you would find the last 10 or so things the blogger you were reading had written.

The blogazine goes with excessive categorization and silofication of blog content. And I think that’s a real shame for most bloggers who take that route. Hey, even if all your last posts are on a big mixed-up main blog page, you can still point people to individual categories if you like. That’s what category pages are for, right?

Partial posts

People put forward all sorts of good reasons to display only partial posts on their main blog page (or archive pages) — roughly the following:

  • improved SEO
  • more page views
  • increased scannability

Until somebody shows me convincing data for either of these three claims, I am going to simply say “bullshit!” (and I’m remaining polite). I’m going to put the culprits on the stage one by one and tell you why I think my reaction is justified. I don’t have any research to back me up (am planning to do some though, so if you want to lend a hand, get in touch) but I do have some reasoning which I believe holds together.

Improved SEO

I have to admit I’m biased against SEO. For me, most SEO aside from “markup your stuff properly (be search-engine friendly) and have great content” is a pile of rubbish. I mean, there are some very obvious things one needs to do for SEO, but they are “common sense” more than “secret tricks”.

If a search engine is doing its job correctly, it will pull out the page that is most relevant for the human being who typed the keywords it based the search on. Make it good for humans, roughly, and it’ll be good for search engines.

When SEO gets in the way of the human experience, I have a big problem with it. And partial posts on the blog page does get in the way of a good reader experience. Why do I know that? Because of what I call the “closed door” phenomenon. A link to click, like a folder to open, is a closed door. You don’t know what’s behind it. You don’t know if it’s worth your while. Chances are you won’t click. Chances are you won’t read the rest of the post.

Even if you know the post is going to be worth it, to read the ten posts on the home page of such a blog, you’re going to have to click on each title (all ten of them), and open them in different tabs, or go back and forth, and maybe get lost in the process.

The original blog format puts all the articles neatly one beneath the other. You start reading at the top, scroll down as needed, and before you know it you’ve read the ten articles.

So, if it really does improve SEO to display only partial articles, I would say that the problem is with the way the search engines work. We should never be creating bad user experiences for the sake of SEO.

(I’m aware that what I claim about the “bad user experience” of partial articles on the main blog page needs to be demonstrated. Working on it. Get in touch if you want to help — or if you can save us the work by showing somebody has already done it.)

How exactly are the partial articles supposed to improve SEO? Well, as you can tell, I’m no expert, but based on what I’ve heard it has to do with duplicate content. Yeah, Google is supposed to penalize duplicate content. And of course, if you publish whole posts on your main blog page, and in your archives, then you’re duplicating the content from the post page — the one you want people to land on directly when they put the magic words into the search engine.

Only… I remember very clearly, in 2007, when Matt Cutts was asked about duplicate content on blogs. (And Matt, if I’m misremembering because it feeds my theory, please set me straight.) He didn’t seem to be saying that it was really a problem. And for what it’s worth, make a note that he’s providing complete posts on his main blog page — not excerpts.

The way I understand it, the duplicate content penalty is a weapon in the war against spammers and link-farms and splogs etc. Having 2-3 copies of the same post lying around do not make your blog sploggy.

Enough for the SEO.

More page views

What can I say about this? First, the reason people obsess about page views is because of advertising. If you’re rewarded for each ad impression, the more pages are viewed, the more money you get.

Sure.

But this begs the question: how much are you willing to sacrifice of the user experience (see above) for a few dollars? Most advertising revenue on blogs is miniscule.

People imagine that “more page views = more articles read”. Nope. I can read ten articles on your home page for only one page view if you publish whole articles. So of course, if you switch to excerpts only, you’ll see an increase in page views. But it doesn’t mean you’re being read more. Don’t be fooled. (This would need to be proved, of course — but the so-called proof that the excerpt method increases page views is worthless in my book, because it’s measuring something that isn’t really meaningful, unless your purpose in life is to sell ads on your blog rather than be read, which is your right, but in which case maybe I’m not going to be that interested in reading you anymore.)

I don’t care about my page views. I just want people to read my articles.

Increased scannability

This one is easy to deal with. Of course, it makes it easier to scan the articles on the first page, if it’s kept short by trimming the articles. Personally, I’m all for a display option that will allow you to see just a list of post names, or a list of post names plus excerpts. Feedly allows this kind of thing.

But do you want to be read, or scanned? Do you want people to read the first two paragraphs of your articles, or the whole articles? Do you prefer to have them scan more headlines, but click less to access the whole articles?

Again, the choice is a non-choice as far as I’m concerned.

The blog is not dead

For the last years, we’ve seen the “blog is dead” meme pop up regularly. I was recently interviewed on this topic by the Swiss National TV — just to show you it’s still around. Aside from the rise of Twitter and Facebook, the rise of the blogzine is often cited as proof of the death of blogs.

Bullshit. The bloggers are still there. We’re still there. We’re not going anywhere. (I need to write more about the so-called death of blogs.)

Now, please go and get rid of those partial articles on your blog pages.

Similar Posts:

Being My Own Travel Agent With Kayak [en]

[fr] En mars, je vais en Irlande, puis à Austin (Texas), puis à San Francisco. Ça fait pas mal de vols à organiser. L'agence de voyage que j'ai contactée me propose un circuit à CHF 2800. En utilisant Kayak, j'arrive (non sans mal, sueur, et heures investies) à faire le tour pour CHF 1650.

Cet article est le récit de la façon dont j'ai procédé.

I have some **serious travel** planned for March.

First, I go to Cork, Ireland, for [Blogtalk](http://2008.blogtalk.net/) and the preceding [WebCamp on Social Network Portability](http://webcamp.org/SocialNetworkPortability), from 2nd to 4th.

Then, I head for Austin, Texas for [SXSW Interactive](http://2008.sxsw.com/interactive/), from 7th-11th.

I’ll be **speaking** in both places.

As I’m in the States, I’ll then head out to spend two weeks or so in San Francisco. Here are what my travel dates and destinations look like:

– 1st: GVA-ORK (ORK is Cork, yes, funny)
– 6th: ORK-AUS
– 12th: AUS-SFO
– 25th: SFO-GVA

I chose the 25th to go back because it seems to be the cheapest day around there. The other dates are fixed by hotel or event constraints.

After fooling around with [Kayak.com](http://kayak.com) for a fair number of hours, and finding it a little confusing (I’ll detail below in what way), I caved in and **called a travel agent** in Lausanne to ask them to sort it out for them.

They got back to me, speedily and kindly, but with a surprising price tag: **2800 CHF** for the whole thing. That’s $2400 for those of you who like dollars.

Now, even though I wasn’t very happy with what I came up on Kayak, I had figured out that this trip would cost me around about 1200$. Not the double.

So, **back to Kayak**. In the process, I’m starting to get the hang of how to do searches for long, nasty, complicated journeys, so I thought I’d share it with you.

A side issue before I start, though: flights to and from the USA have a **much more generous luggage allowance** than flights elsewhere (20kg + cabin luggage). If the first leg of a journey to the USA is inside Europe, though, you still get the “US” luggage allowance for that flight. I was hoping I could make things work out to have the more generous luggage allowance for the GVA-ORK part of my trip too, as I tend to have trouble travelling light (particularly for 3 weeks). But it seems that won’t happen.

As I understand it from the kind explanations a few people have given me, the GVA-ORK part of my journey is considered a completely separate one from ORK-AUS, AUS-SFO, and then SFO-GVA. In short, I’m dealing with **four separate flights**.

So, let’s do the obvious thing first, and **ask Kayak.com to do all the work**. My dates are fixed, but I’m open to the idea of using nearby airports. This is what I gave Kayak.com:

Kayak search: GVA-ORK-AUS-SFO-GVA

And here is what I got:

Kayak.com GVA-ORK-AUS-SFO-GVA

Oops. It seems Geneva dropped off the map. If I select the “neighbouring” airport LYS (Lyon), I get this. Slightly more encouraging, but…

Kayak.com: GVA-ORK-AUS-SFO-GVA

…slightly expensive. Roughly what my travel agent told me, actually. Gosh, I wonder which part of the journey is costing so much? **Let’s try and break things down.**

**First, GVA-ORK:**

Kayak.com GVA - ORK

Wow, is that their best price? $384 and 9 hours of travel to go from Switzerland to Ireland? I should be able to find something better. So, I hunted around a bit on my own. I know I can get to London for around $100 or less with [easyJet](http://easyjet.com), so what about the other low-costs? From the Cork airport site, I got a [list of airlines flying there](http://www.corkairport.com/flight_info/airlines.html). Then I went to individual airline sites — I’ll pass you the details, save to say that [RyanAir](http://ryanair.com) has got some “virtually free” flights (1 penny + taxes) but as they only allow 15kg of check-in luggage (I can make sacrifices and try to stick to 20, but 15 is really low), flight + excess luggage fee actually comes down to not-that-cheap.

Oh, wait a sec! Let’s enlist Kayak’s help for this. Here are GVA-LON flights, according to Kayak:

Kayak.com GVA - LON

That’s helpful, actually. I wouldn’t have thought to check [BA](http://ba.com). The flight is way too early, though. And Kayak.com now gives results with European low-cost airlines — I don’t recall it did this early December when I first tried.

What about LON-ORK?

Kayak.com LON - ORK

I removed RyanAir from the results (they were the cheapest, around $48 — plus extra luggage tax!), and the winner is… [Aer Lingus](http://aerlingus.com)!

So, if I manage to get the timings right, and accept that I’ll have to pick up my luggage and check in again in London, I should be able to get a better deal than the $384 Kayak suggested “out of the box”.

Oh, another idea. Let’s tell Kayak I’m flying through London, and see what happens. Here are the results for GVA-LON-ORK:

Kayak.com GVA - LON - ORK

Still no luck. The first flight is the same as the one I got when I asked for GVA-ORK. Clearly, Kayak introduces constraints (like… airlines must be working together) when asked for a trip. That probably explains why my total trip seems so horrendously expensive.

Right, now we’ve dealt (more or less — at least there seems to be hope) with the first part of the journey, let’s look at the rest.

**ORK-AUS-SFO-GVA:**

ORK-AUS: $509

Kayak.com ORK - AUS

AUS-SFO: $125

Kayak.com AUS - SFO

SFO-GVA: $530

Adding all that up, we’re quite far from the $2400 my travel agent or Kayak suggest for the whole flight.

Now, let’s dig in a little further. How about I ask Kayak for ORK-AUS-SFO-GVA? I’ve already identified that the GVA-ORK part was problematic, so maybe… maybe:

Kayak.com ORK - AUS - SFO - GVA

$1029! And all with American Airlines! That sounds nice. Add to that a bit less than $200 for the GVA-ORK bit, and I should manage to do all this flying for roughly $1200. Much more reasonable (though still a big hole in my bank account credit card, given the sad state of my finances these days).

So, ready for the details? Because, no, in case you were wondering, the fun doesn’t stop here. Sick around, there’s still work to do.

**First, GVA-LON-ORK.**

London has a problem: it has too many airports. Aer Lingus fly out of LHR to Cork, so ideally, I should plan to arrive there. I don’t think I want to go through the fun of commuting from one airport to another if I can avoid it.

That unfortunately rules out easyJet, who don’t fly to LHR. They fly to LGW, Luton, Stansted, but not LHR. So, let’s check out BA, who were actually cheaper (though at an ungodly hour, and for LGW).

BA: GVA-LHR

Right, so for 144 CHF, I get to fly out around 10am, which is actually quite nice. I land around 11am. Let’s look at Aer Lingus flights to ORK, then:

Aer Lingus: LHR-ORK

I’m very tempted to take the 14:05 flight instead of the 18:05 one, **but**. That would leave me with only 3 hours in LHR to get my luggage, go from terminal 1 to terminal 4, and check in again. The London crew on Twitter tells me it’s a little tight, though others seem to think it’s OK.

So, well, that would be it for the first part of the journey.

Now for the rest.

**Then, ORK-AUS-SFO-GVA.**

Here are the details I get from Kayak for this multi-city journey:

Kayak.com ORK-AUS-SFO-GVA 1029$

As you can see, American Airlines seem to like Chicago airport, ORD. [Dennis Howlett](http://twitter.com/dahowlett) warns me against going through that airport, but it seems the other options are going to cost me an extra $1000.

But that’s not all. What exactly are the “layovers” here? I’d assume they are plane changes. But 55 minutes in Chicago and 1h35 in Brussels on my way back don’t really seem to allow time for that. Chances are I’d miss the connection — but then why would Kayak.com (and AA!) suggest this kind of combination?

It’s not the end of the world if I get home a day late, so I guess that for $1000, I’ll take my chances.

Let’s not stop there, though, shall we? I decided to dig a bit deeper into all this. See, for example, I tried asking Kayak.com about:

AUS-SFO-GVA: $1669

Kayak.com AUS - SFO - GVA

Why isn’t Kayak coming up with one of the (obviously cheaper) combinations for the SFO-GVA leg? Why is BA suddenly the cheapest option? I don’t get it.

See, for example, this flight option for SFO-GVA, $550, is much more exciting than the AA one via ORD and Brussels:

Kayak.com: SFO-GVA

Just one change in Newark. And it’s a shorter overall flight, too.

That means I need to get the ORK-AUS-SFO part separate. Let’s look at it now:

Kayak.com ORK-AUS-SFO

The cheapest deal is $624 with AA and Frontier, which is an immediate (and logical! what a surprise!) combination of the two cheapest deals for ORK-AUS and AUS-SFO taken separately. I don’t seem to gain anything (financially) by booking them together.

Now, the problem here is that the flight times are really long (20h). I’m quite tempted to force my journey through some European city other than London and see what happens.

A quick trip to the Austin airport site seems to say there are [no direct flights there outside the US](http://www.ci.austin.tx.us/austinairport/nonstops.htm). I can’t find that kind of information for DFW, unfortunately. I’m keeping an eye on [DFW](http://www.dfwairport.com/) because I could land there and take a road trip to Austin with a friend. It’s 3.5 hours on the road, though, so I need a flight that lands early enough.

For example, let’s take Dublin, as I’m already in Ireland.

Here are Kayak flights from DUB to AUS: most interesting deal $484 with Delta for a 19h flight:

Kayak.com: DUB-AUS

Come to think of it, you know what I’d like? I’d like to be able to place all the flights on a chart, with for example “price” on the x-axis and “total flight duration” on the y-axis. I’d be willing to pay $50 extra or so to cut of a certain number of hours of travel, but as of now there is no way to visualise this kind of thing easily. The “Matrix” tab in Kayak has a promising name, but all it does is give best price and number of stops per airline. Not very exciting.

What about ORK-DUB? Well, the fine folks at Blogtalk recommend [Aer Arann](http://2008.blogtalk.net/travelling) (they have a great “travelling” page, btw, I’ll have to take example on them for [Going Solo](http://climbtothestars.org/archives/2007/12/14/announcing-going-solo/):

Aer Arann: ORK-DUB

Cheap flight, $36. What would Kayak say?

Kayak.com: ORK-DUB

Well, RyanAir is cheaper but I don’t want them, and the Aer Arann flights are there, but a bit more expensive than what I found. Hidden costs, maybe? Or maybe just an update glitch — I’m aware it’s difficult to keep everything perfectly in sync.

Gah. This is turning into another nasty headache.

Let’s go back to letting Kayak take care of ORK-AUS-SFO. I had a look at flights from [Shannon](http://www.shannonairport.com/), but the price difference is not worth the couple of hours by bus to get there. I also considered SAT (San Antonio) but it’s really out of Austin, so not interesting. I’m willing to fly in another airport than SFO though.

Sidenote: this is where I discover I can “favorite” flights in Kayak. I should have started doing that hours ago. So, here’s the flight I’m favoriting for the ORK-AUS segment. I don’t want to land at 12:15am in Austin, so the choice is easy to make. Will have to get up early in Cork, though. Ugh.

Kayak.com: ORK-AUS favorite

You know what would be really cool? If I search for ORK-AUS-SFO, I’d like Kayak to let me know which flight combinations contain that flight I’ve favorited. I wonder if it does that. Let’s see! But before that, I’ll go and favorite the flight I want for heading over to San Francisco. So, here is what Kayak gave me for that segment, remember?

Kayak.com AUS - SFO

The cheapest flight is $125, but if you have a close look, you’ll see that all these are either dreadfully early, or quite late. I’d rather leave sometime later in the morning. Luckily, Kayak provides a “filter” that allows me to select that. (Remember, earlier on, I was wondering why Kayak was suggesting routes with 55min stopovers? Well, there’s a “stopover length” filter too that I could have used to avoid that.) Here’s what happens if I decide to leave between 8 and 10am:

Kayak.com: AUS-SFO Flight Time filter

For roughly $200, I get to sleep a bit more. This is another case where the price/something-or-other graph would come in handy: it would help me visualise how much I have to pay to leave later. (I’m learning to factor in cab fares and stuff like that when making flight decisions.)

So, back to our combined ORK-AUS-SFO trip:

Kayak.com: ORK-AUS-SFO best choice

By playing with the time sliders for flights 1 and 2, I managed to filter out the flights that didn’t contain my two favourites (at no surprise, Kayak doesn’t tell me that this “multiple flight” actually contains a single flight that I favourited… too bad). Result: $695 and decent flying times.

**So, let’s recap.** (I’m going to be doing the actual booking tomorrow, it’s getting late and I’m tired, which is usually a recipe for mistakes. Also, the prices the airlines and Kayak give could be slightly different, so this is an approximation.)

GVA-LHR: BA, $125
LHR-ORK: Aer Lingus, $60

That’s $185 for me to go to Cork.

ORK-AUS-SFO: AA and Frontier, $695

SFO-GVA: United and Qatar, $550

Total: $1430 = 1650CHF

That’s a bit more than what it seemed I’d get away with at first, but there are less stopovers and the flying times are nicer than the cheapest deal. That’s worth a couple hundred $.

So, thanks Kayak. That’s more than 1000CHF less than my travel agent came up with. But God, did I have to work hard for it. There is definitely room for improvement in the business of helping people sort out their travels.

While I was writing this post and [twittering about my trials](http://twitter.com/stephtara), [Bill O’Donnel](http://egopoly.com/) (find him [on Twitter](http://twitter.com/agentbillo), he’s the Chief Architect at Kayak!) sent me a message saying he [wanted to read my post](http://twitter.com/agentbillo/statuses/524594472) when I was done. He also added that he was [forwarding my twitters to the UI team](http://twitter.com/agentbillo/statuses/524596032). So, guys, hope you enjoy the free [experiential marketing](http://climbtothestars.org/focus/experiential-marketing/)! In a way, only — it’s not really an experiential marketing campaign because nobody asked me to do anything, but this is typically the kind of stuff I *would* write up in such a campaign, and an example of *authentic user behaviour* that experiential marketing “re-creates”.

So anyway, hope you enjoy this tale of user experience. And I also hope my fellow travellers will find useful input here to help them sort out their travels.

Thanks to everybody who answered or simply put up with my numerous questions and tweets during the process of sorting out this trip.

Similar Posts: