It just shortened “http://daytee.com/1f9” to “http://bit.ly/4h5OXK” without asking.

So it shortened a 14 character URL to a 13 character URL when there was plenty space left in the 140.

In shortening that URL readers lose a LOT.

Any meaning in the url itself is lost, as you can’t see where it’s going. You can no longer easily search for the URL in twitter to see who else tweeted it – a good way to gauge popularity of something.

The site owner also loses a lot.

Link love is diluted. Domain recognition cannot be built. And any site which tracks mentions of specific links to help rank content (like daytee) have to work a lot harder to do so.

Daytee

When I set up daytee.com I picked the domain specifically to be short, and wrote a url shortener for all content so that all shared links for the first year or two of the sites operation would be just 14 characters long. This was specifically to avoid the url shorteners.

I could have gone further, but wanted to keep a .com domain. “dayt.ee” is 3 characters shorter than “daytee.com” but costs more, looks less like a domain to most people, and so is less memorable. Also, it doesn’t feel like a hoop I should have to jump through just to have a link shared in a legible manner.

I’m sure there are excellent reasons for a twitter client to want to send traffic through a bit.ly link to gather stats and build a better understanding of usage patterns. Owning the link in this way should be a side issue, not the objective, of link shortening.

Turning a legible 14 character link into an illegible 13 character link isn’t convenient link shortening – it’s link obfuscation.

In order to give the service the quality feel I felt it deserved in the quickest time possible I needed a designer without a backlog. All the designers I usually use had backlogs… Oh oh.

Listening to twist 13 reminded me of 99designs. Without entering into the whole debate about creatives doing unpaid spec work – it seemed like a cost effective, quick, pain free way to get a new identity worked up. So I took 10 minutes to write a spec and fired up my first logo contest.

So how did it go? Well friending topics now has a new logo, so it went well.

The setup process was painless. Certainly quicker than briefing a human designer, involving less caffeine too.

I got over 90 entries to my contest over 7 days. Over half of these came in the last 2 days of the 7 day contest. There is an option to shorten your contest to just 3 days. Assuming these fall during Mon-Fri I’d guess the extra few $s is worth it.

It’s also worth pointing out that of the 90 easily half are duplicates with small changes – some designers drop a few designs in a batch all with a similar theme but subtley different execution. Coming from an agency background this feels a little like sneaking a look at the designers precious ’sketch’ folder when they are getting a coffee. I kinda like it.

Of the 90+ designs there were easily half a dozen really strong options.

A couple of tips if you are using 99designs which might help you to get the best out of the system.

It seems like you get a batch of new entries soon after doing a batch of star ratings. So rate things continually. You can always change your rating later – so if a 3 star grows on you it can be elevated. But failing to rate labels your contest as potentially problematic, deterring designers from spending time on it.

Also, give feedback to designers. I didn’t do enough of this due to a hectic week – and I suspect it would have allowed some of the designers to rework their design to fit my needs more specifically.

Also, be VERY clear in your comments. A lot of the designers don’t have English as a first language so if you are ambiguous you might get the opposite of what you expect. Avoid requests as statements: ‘it might look better in upper rather than mixed’.

While pondering this problem about 3 months ago I decided to build a web app to collect the newest designs from some of my favoured sites and present them on one page.

This got built pretty quickly, so I added a url shortener, some voting mechanisms, a popularity chart, a hookup to twitter, and some nicer styling. Daytee was born.

Why post now? – The site just covered its thousandth t-shirt, from around 100 different, mainly indie, t-shirt brands.

Friending topics gives you a more specific list of trending topics, based 100% on your network. So for me I see lots of coffee, code, twitter and apps. I don’t see #musicmonday, Jay-Z or #todaysweirdspammyhashtag.

Feedback on Twitter has been really great so far – and about 50% of visitors to the site are signing up. It’s also been listed in a few of the better twitter app sites.

If you’ve used the service do let me know what you’d like me to add next, and consider rating or reviewing friending topics over at the listings on oneforty.com and twitdom.

Interesting tale of homeopathic time travel with some disturbing lab mice based stuff. Nicely meta. Read it.

Rubbish ending though. Stop reading at page 490 and make something up instead – sorry Scarlett (oh, and your website seems to be a bit goosed – keeps pointing to www2. which doesn’t know what to do with the php).

Following the bye bye to the last book I’m now hunting for something else to read with a similar opening farewell in the title. Some hopes!

Note the sophisticated use of the carrot as foodstuff rather than nose, thereby subverting the language of the snowman. I also used bioethanol for the eyes instead of coal to reduce the carbon footprint of the shortlived snowbunny – but they don’t show up too well.

Went to bed last night – it was raining – woke up to today telling me snow Armageddon had paused for a bit. Looked outside and…