Cuil.com – the next Google?

From the way Mike Arrington carries on about it (oh and a bit more here too), you’d think so.  My initial impressions are quite different (to be fair, he does state “…it doesn’t appear to have the depth of results that Google has, despite their claims. And the results are not nearly as relevant“).

I started with the old vanity surfing evaluation.  I initially searched for plain old “cederman” (I’m number 2 on Google behind darn Lars-Erik at the moment!).

Searching for \

Very confusing list of results without any sort of rhyme or reason to them.  None of the vaunted contextual search options delineating between the few Cedermans on the web (there are only three or four of us, with three of us being published authors).  There are also some incredibly irrelevant results there, and why is an extremely old copy of my twitter page listed there, but not cederman.com?

“cederman-haysom”, “tim cederman” and “tim cederman-haysom” didn’t fare much better either.  Ouch, it’s not hard when there are literally only three cederman-haysoms in the world!

Anyway, whatever.  The true test of a search engine is looking for stuff that you actually want.  While TechCrunch did several broad searches (such as for “dog”, “apple”, and “france”), these are fairly rare in real world searches.  At the moment I’m planning a trip to Belgium, so I tried one of the cities I’m looking at accommodation for and trying to plan a tour of.

No results for Ghent on cuil

No results!  Search for the alternative spelling “Gent” only showed pages in Dutch.  The most curious aspect to this is that contextual search worked here.  The options shown up the top are valid Ghent related things you’d search for, and it’s actually really nice having them identified and clickable in this manner.  So why have “no results” for the main page?  Doesn’t make any sense at all…

Finally, what’s with the ordering of results?  At first I thought I had to read sideways, in rows essentially, but the results don’t line up.  If I have to read by column, then I need to scroll to the bottom, and then all the way back to the top.  Without a sense of relevance, it’s very disconcerting, particularly if you’re searching for something you don’t know much about.  Sure it might help provide some users with a shotgun spray of results, but I think there’s a good reason why Google’s layout still works best.

As an example, imagine someone looking to buy a new Roomba:

Results for searching for a cheap Roomba on Cuil

While Arrington also states “I want to reemphasize that Cuil is only an hour old at this point, Google has had a decade to perfect their search engine.”  This is disingenuous to say the least.  Cuil is certainly not an hour old, and Google was VERY impressive when it first launched.

Still, I’ll have a close eye on what they do next.

Best analysis of a PhD I’ve ever read

Found over at Jamie Lawrence’s blog.

I found myself nodding the whole way through.

I remember hearing that, on average, 1.6 people will read your PhD thesis. I’m pretty sure that includes yourself, your spouse, your supervisor, your second supervisor and your examiner (yeah, that’s technically 5 people. If someone says they’ve read your thesis, they’re probably lying – they read page 9). You have to accept, that no one in the world will want to wade through this document. Ever.

Oh, and the glorious days of a constantly clean house:

If you’re like me, when it gets down to the nitty-gritty, boring detailed work of the PhD, you need to remove as many distractions as you can because, at this stage, just about anything is going to be preferable to your PhD. Computer games, good fiction and the Internet are all obvious distractions that can be minimised. Washing up was one of my favourite distractions, which I never found a way to avoid.

Required reading for any potential PhD student.

iRobot and lessons for design

It’s funny how something can be a great design, and yet with a few tweaks, it becomes completely awful.

Witness the Roomba.  I love my Roomba – I bought it in October 2006 when I moved in to my cottage and realised I needed a vacuum cleaner.  I bought the bottom of the line model, no accessories, and as of August 2008, with some minor maintenance it still runs great.  Even its battery, while definitely not as great as it once was, still has enough charge to clean the whole room.  It’s easy to use, cleaning it is very simple, and maintenance is also very user friendly.  And of course, it does a great job of cleaning the whole cottage (although I am lucky to have Roomba friendly floors and furniture).

With so much love for the Roomba, it was a natural choice to buy a Scooba as well.  I don’t mop and I could see grime building up.  Scooba was only $99 on Woot, so I bought one straight away.  I noticed immediately that it’s a lot bigger and heavier than Roomba, making it more difficult to move around and to manipulate while cleaning.  It doesn’t fit as nicely in our cupboard, and its increased height means the light sensor keeps getting caught under cupboard doors.  It leaks water when being moved from room to room, and will often insist to check the tank for no discernable reason.  There are a lot more parts to clean, and maintenance is a lot more complex.  Emptying it becomes a gross chore (instead of tapping a box into the bin, I usually manage to cop a bit of spray back when pouring out the dirty water).  When I lift it up, the tank often separates from the body.  I am struggling to get it to clean a single room at the moment with its myriad of problems, and it’s only 6 months old.  iRobot won’t support it because it’s refurbished.

Long story short, I’d never recommend the Scooba to anyone.  But I’d recommend Roomba in a heartbeat.

So how did they go so wrong?  There were several key areas:

  1. They identified an alternative niche and went for it at all costs.  I suspect their other products like the Looj will do better.  It seems like at no point someone wondered “will having <problem x> in addition to all the other problems mean people will just give up?”
  2. It’s overcomplicated to the point it can no longer complete its original purpose.  Mopping by definition is more complex than vacuuming because it is a 2 phase process.  However I can’t help but wonder if it would be more successful if they sacrificed some of the cleaning capabilities for simplicity.
  3. Too many choke points in the design.  The beauty of Roomba is it keeps working if some parts of it aren’t.  Scooba will fail if the tank connection gets clogged, if the hand-mixed formula is not done right, or if it detects a problem with its pipes.  One of these three things happens to me every time I try to deploy it.
  4. They forgot the Roomba design ideals.  It’s clunky, hard to use, makes a mess and has inherent design faults (such as a battery that fails after just a few months).  It was like they started from scratch without taking on-board any of the Roomba lessons.

I love domestic robots, but unless you’re really desperate, do not buy a Scooba.