I clicked dial, and this is what filled my screen:

An absurdly long phone number. Bemused, I went to inspect the link I'd clicked. This is what it showed:

The tel: URl scheme is brilliant. You can write something like:

<a href="tel:07700 900123">Call Me!</a>

And when you click on Call Me! your phone dialler will pick up the phone number and offer to place the call. Nifty!

In this case, the crappy marketing system is adding Urchin Tracking Module parameters to every link. Including the phone numbers.

tel:02083178830?utm_medium=ecrm
&utm_source=email
&utm_campaign=TRAN
&utm_content=UK_TRAN_01_0_ORDER-CONFIRMATION_E_V01
&campaign=TRAN
&adgroup=UK_TRAN_01_0_ORDER-CONFIRMATION_E_V01
&utm_term=restaurant_phone_3_3

To be clear - this is useless. The user clicks on the phone number, the device passes the URl directly to the phone dialler. An HTTP request is never made and those parameters are never sent to a server.

Now, in fairness, perhaps my Android dialler should probably be smart enough to recognise the cruft at the end of a phone number and discard it. This is Postel's Law in action.

Except... My reading of the RFC says that the dialler is handling things correctly.

If the reserved characters "+", ";", "=", and "?" are used as delimiters between components of the "tel" URI, they MUST NOT be percent encoded. These characters MUST be percent encoded if they appear in tel URI parameter values.

For example, if I wanted to dial +447... I should use tel:%2B447...

The dialler sees the unencoded ? and treats it as a delimiter. It then sees utm and assumes the letters are part of the phone number. Just like you can dial 1-800-FLOWERS, you can write tel:1-800-FLOWERS and have it go through to 1-800-356-9377.

utm medium on a telephone keypad is 886633486 - which is exactly what appeared on my phone screen.

So, if you're writing link tracking software, please make sure only to add parameter to URls where it makes sense.

In the time it took me to write this post, my meal got delivered and it was delicious!

Just Eat UK

@JustEatUK

Replying to @edent@edent Thanks for getting in touch and sharing this feedback with us Terence, we appreciate it and we'll be sure to bring this to our tech team's attention. ^EM

---

A few weeks ago, the .tel registry released their domain restrictions - users can now host their site anywhere. I'd already experimented with about.me/edent as a contact site - but even that seemed too weighty for my needs.

Here's how I used CSS to craft a beautiful homepage, HTML5 to give it a sprinkling of syntactic sugar, and SVG to make it as tiny as possible.

The entire site is less that 16 KB when gzipped for transport - and only 36KB uncompressed.

All CSS and Images are inline - that reduces the number of requests which need to be made to the server.

Depending on your hardware, the entire page should load and render in under 0.3 seconds.

Oh, and there's no JavaScript to slow things down either!

SVG

For the touch screen icons, I used SVG. I've long been fascinated with the magic of vector graphics - I love how they obliterate file sizes while looking fantastic when shrunk or enlarged.

As an example, the regular Twitter logo when optimised as a PNG is around 16KB. The SVG is only 436 Bytes. Yes, that's right, just 3% of the size!

In total, those 25 SVGs added up to less than 20KB. They are so small, web browsers spend longer connecting to the server than the time taken to download them. So I placed them inline to the HTML - that saved a heap of downloading and rendering time.

I didn't skimp on accessibility either! Léonie Watson's brilliant blog post on accessible SVGs gave me everything I needed to add alt-text to the images. I was even able to mung the HTML enough that the site works on the text-only Lynx browser!

Browser support for SVG is thorough. Because of their small size, mobile browsers have always had good support for them - and that support has filtered up the chain to desktop. It was only when going back as far as IE8 that I experienced any real problems.

Ultimately, there are too few people using those old browsers for me to bother with fall-back PNGs.

CSS

In keeping with the .tel theme, I decided to make the site look like a modern touch screen phone.

I found the Marvel Pure CSS Phones - a collection of fairly tight and responsive CSS classes which simulate the look of a phone. I trimmed down the files and made the "screen" portion resize depending on the number of elements inside it.

As with all moderately complex CSS projects - there's a tension between relative and absolute sizing. I managed to bash enough % signs into the existing code to make the phone scale with the size of the screen. On absolutely tiny screens, the font does get rather cramped - I should probably use some media queries to sort that out!

HTML5 - the semantic web

The site has three main pieces of semantic markup.

1. Schema.org Person - a way of structuring data on the web. This allows me to define facts about myself in a predictable manner, including how the linked sites are related to me.
2. hCard microformat - to me this seems identical to the above, but my even nerdier friends complained that I'd left it out.
3. OpenGraph Tags - when the page gets shared on services like Facebook and Twitter, the sites are able to pull in metadata about the page.

Hopefully, that will have a positive effect on SEO as well.

I also considered WebFinger, but that seems like dead standard - no-one is using it.

All the chat app links should open up your messaging app - if you have one. I've got a long rant about URI Schemes coming up.

Security

Finally, I used Let's Encrypt to generate a security certificate. I considered making it mandatory to use https when connecting - but there are still too many browsers out there which can't cope with modern security requirements. For now, I'm giving people the option of accessing it in the clear.

Further Reading

The site is entirely Open Source and shared on GitHub.

I've released all the logos as Super Tiny Social Icons on GitHub. I used svgo to optimise them.

Take a look at Marvel Pure CSS Phones.

Finally, if you'd like to get in touch - you know where to find me!

What is .tel

Your address book is probably a mausoleum - stuffed with the rotting corpses of long dead phone numbers. Perhaps you took my business card back in 2002, duly entered it on your Palm Pilot, and never spoke to me again. That address book entry has a phone number I've not used for a decade, an email address provided by a defunct start-up, and a postal address for a country I no longer live in.

Isn't there a better way?

That's what .tel was supposed to be.

• I register a .tel domain - http://edent.tel
• I fill it with my contact details.
• You store my .tel in your address book.
• When I change my phone number, I update my .tel and your phonebook receives the changes.

The magic of .tel is that everything is stored in the DNS. It shouldn't matter if the website goes down - or even if you've got low connectivity. All you need to do to get my details is:

dig @8.8.8.8 edent.tel naptr

Or, to get everything in the DNS records:

dig +nocmd edent.tel any +multiline +noall +answer

The tragedy of .tel is that there was almost no UI customisation available. Every site looked close to identical, corporate colour schemes couldn't be easily integrated, and the design was limited.

This is what it looked like back in 2009:

Which, thankfully, had improved by 2013:

No further improvements were made.

So, how well did it work in practice?

Lack of critical mass

Back in 2012, there were 256k .tel domains. In 2016, it's a mere 105k domains. Those numbers need to be in the multi-millions in order to get the traction needed for success. In their original proposal, they were expecting 20 million registrations five years after launch.

As registrations fell, so did income. Senior staff left the .tel organisation, and the infrastructure was left to rot. There were no updates, and it looked like .tel might collapse - an unprecedented event in DNS history.

I worked in the mobile industry for a decade. I don't think I ever met anyone else with a .tel. I got mine when they first launched in 2009 - and have been lonely ever since.

As far as I can tell, no mobile phones were ever released which had .tel capable address books.

Relaunch!

Late last year, .tel owners were sent emails describing the upcoming relaunch and reinvigoration of the service.

• ✓ Lifting of usage restrictions.
• ✓ A new Telhosting platform.
• ✓ Android and iPhone apps.
• ✘ No porting of data!
• ✘ No sub-pages.
• ✘ No search.
• ✘ No advertising.
• ✘ Limited foreign language support.

It's a bit of a mixed bag. But, hopefully, there's enough to sustain numbers - if not increase them.

The most important is the lifting of hosting and design restrictions. Users will be able to point their .tel at any site they like. The idea of it just being an address book is disappearing.

For those people who do want to keep it as their virtual contact card - a new platform is being launched with an improved interface and fewer design restrictions. It will still be free of charge for domain owners.

As far as I can tell, this also means that sites can be secured with https - something which was unavailable on the old system.

Apps will be available for editing your site - but it would be a lot more useful to integrate with native address books.

It is downright odd that they're not automatically porting over peoples' data. There's going to be a one month grace period before launch in mid-March, but that isn't a huge amount of time.

The lack of sub-pages and search probably reflects how little those features were used. Removing Ad-Sense seems weird - but people can always add their own advertising.

They're also dropping support for "Arabic, Czech, Japanese, Korean, Portuguese and Russian" - I have no idea what those languages have in common! I assume they just mean that their hosting platform won't contain translations for those languages.

Is it enough?

I doubt it. Sorry to be so pessimistic - there are now literally hundreds of available top level domains. Including .mobi, .mobile, .phone, .call, .me - all of which could serve the same purpose.

If .tel had built on their early momentum - and perhaps done some deals with mobile networks or manufacturers - then perhaps .tel would be in a better position.

It is pretty neat that they can store data like this in the DNS, and it is more discoverable than .hcard or other microformats - but I fear that the idea of placing one's address details in DNS is doomed to failure.

Because they aren't porting existing data to the new system, I expect that a lot of existing .tel sites are going to be empty.

Bonus Retro Video

This is how .tel launched itself back in the day.

Bless!

Since then, the site has got easier to use and now looks a lot better - especially on mobile.

When http://edent.tel/ came up for renewal last year, I decided to add Google Analytics to it so I could see if it was useful.

So, at the moment it bumbles along with 2 to 5 hits a day - more when my blog is popular. I get a couple of genuine emails per month from people visiting it - sometimes leading to work. But mostly it generates spam.

The total cost to renew for another year is €18. That's not a huge amount in the grand scheme of things - and I did originally get the domain free for a couple of years. I hate the idea of linkrot - but it turns out hardly anyone links to it.

It's handy to have a dedicated microsite which contains all my contact details - but as .tel has never quite got the momentum it deserves, it's hard to justify keeping around.

I'd love it if everyone had a .tel and my phone's address book just pointed to their entries. I'd never have to worry about updating people's details when they changed job or phone number. Sadly, that doesn't look likely to happen. The TelNic registry put out its last press release in June 2011 and they haven't seen any significant press since September 2011. Similarly, their blog has gone over a year without an update.

I know some of the guys behind .tel - they're smart, passionate, and committed. They created some great developer resources. But, still, for some reason, the service has never quite broken into the mainstream.

So, should I renew it or just let it lapse? I've got until the 22nd to decide. Leave a comment in the box

"Visit aitch-tee-tee-pee colon slash-slash edent dot tell... No... Tell. It's spelled TEA-EE-EL. Yes. Just one EL. No, I don't know why. Here, let me write it down for you on a little cardboard oblong..."

So, as you've guessed, I'm not the world's biggest fan. But I was "lucky" enough to win a free domain courtesy of TelReg. As I'm a freebie customer, I don't think it's fair of me to critique them in any way - but I'm sure going to critique the whole .tel idea.

First of all, why not take a look at the site. edent.tel. Go ahead, I'll wait.

Yeuch! What a dull looking site. There is, at the moment, no way to customise the look and feel, I can't even add my picture (not that it would make the site look much better...). I can, however, add some very limited information.

Now, I'm not a professional usability expert - but even I can point out the dull text, the huge dead space, the reliance on scrolling. Yeuch.

To make matters worse, there is a mobile version of this page. Seeing as most of the people to whom I'd give a business card will have a mobile phone, this makes sense. Let's take a look at it in two popular phones.

iPhone 3G.

BlackBerry Bold (9000).

Yeuch! Again, no panache, no style. Just dull dull dull text. You'd think that the .VCARD link would be the top link, wouldn't you? That would be a basic usability feature. But no.

As an aside, the iPhone can't even download the .VCARD. This is the fault of Apple.

But the .tel people shouldn't even offer a link to a phone they know can't support it.

Anyway. As I've said, I've got this free for a year. A few hours in and I can't see myself renewing it. If .tel can improve some of their basic flaws and improve their customer perception, I might reconsider.

I sort of like the idea. But I prefer SyncML sites. Take ZYB - (owned by Vodafone, who own me. Yes, I am biased) - if I give you my ZYB address, and we connect, whenever I change my address or telephone number it is automatically updated on your phone. Smart. And also pretty. Two things that .tel is not.