Google is so crude
by Ross Lasley
The pace of technology often makes it difficult to see the big picture – and lately this seems to be especially true of Internet search.
Google is perhaps the best current example – they are doing some amazing neato things over there. Using google will improve your world – they offer an excellent search toolbar for your browser, they are working on email services (Gmail), they offer a shopping engine (Froogle) and they even have a way for people to write applications that use the technology (like Google Alert or TouchGraph Google Browser) It is a phonebook, a calculator, and will clearly be much more – they want you to get lots of data through Google.
This is impressive stuff, no doubt about it – and it presents the cutting edge of where text based search is going.
One thing that I haven't heard many people talk about is that text based search – as a technology – basically sucks.
See – fundamentally – Google isn't any different today than when it launched. They have a system that crawls the web looking for content, makes a copy of it, and then analyzes that content. They keep improving the analysis machine – algorithms are much better than they were, and all of this work does in fact make it more likely that when you type something into their engine you'll find what you want.
It is a very impressive steam engine, no doubt about that. It is much better than the other steam engines like AOL Search or Yahoo – but at it's core don't forget what it is – a technology that is doomed to failure and will need to be replaced outright.
There are many fundamental problems with how text based search works, and the best and brightest have concluded many of them simply can not be solved.
One big issue is the sheer size of the data – it depends on how you count but the web is big... really, really big. It makes the social security database look like a drop in a bucket – there has never been a pile of data anywhere near this large, ever. Google says they are searching 4,285,199,774 web pages (yeah, that's billion with a "B") – and they acknowledge this is a fraction of what is out there – no engine claims to index more than ¼ of the web. On a practical basis they index a much smaller amount. It is reasonable to think that 10 years from now there will be a hundred web pages for every living man, woman, and child on planet Earth.
And here is why you don't have a steam powered car - “syntactic search barrier”
This is basically a super geeky way of saying these machines can't read or really learn. Text based search will never be able to read (understand the context of the content), it just isn't possible. When reading is really possible it won't be reasonable to call it text based search anymore. If you're interested in a good textbook on this sort of arcane topic, click here.
What the businessperson needs to know and recognize is that we are beginning to come to the end of text based search - the next half decade or so must include fundamentally new technologies. The current tool will get polished, the tool will get faster, better, etc – but eventually we'll need a genuinely new tool.
I'll keep you posted.
|
KISS News |
|
KISS is pleased to announce we have
formed a relationship with MX Logic to offer our customers
Anti-Spam and Anti-Virus Services.
We did lots of research and this is the best solution we
know of to this growing problem. This sophisticated technology
utilizes off site resources – there is nothing for you
to install on your computer and no settings for you to change.
The way it works is that your email flows through this system
before it ever even gets to you – and you just get reports
of what has happened that allow you full control of how it
all works, e.g allow all mail from a sender, block all with
a particular subject, report frequency, etc.
KISS-anti-Spam is just $4 a month ($48/year) per mailbox
KISS-anti-Virus is $5 a month ($60/year) per mailbox
Please note: you must host your site with KISS and
have your domain name registered with KISS at this time. In
late July we'll be making these services available to anyone.
Interested? Drop
us an email. |
|
 Tasty Tip |
|
by Michael
McGrath
Everyone likes chicken. It's low in fat (without the skin),
high in protein, and oh so very versatile. It's also the lowest
common denominator menu item. Think about it, when's the last
time you saw a restaurant menu that didn't have at least one
chicken dish on it?
KISS's “Everything Chicken” cookbook, to be
released late in 2004 or early in 2005, presently stands at
100 recipes, but continues to grow. Here's a sneak peak at
one tasty finger food item from it:
The Last Chicken Wings Recipe You”ll Ever Need
3# chicken wings
1 large bottle Teriyaki Sauce
3 tb minced garlic
2 tb corn starch
¾ cup dark brown sugar
Preheat oven to 400 degrees. Cut the tips off the wings,
and split them at the joint. Mix the Teriyaki Sauce, corn
starch and brown sugar in a bowl. Line a caserole baking dish
with the minced garlic and add the split wings in a single
layer.
Pour the Teriyaki Sauce mixture over the wings. It should
come just to the top of the wings, but if it is a little deeper,
that is okay.
Bake in the preheated oven for 90 minutes. Allow to cool
for 30 minutes before eating. They don't fit well in a formal
setting, and plenty of napkins will be needed. They do eat
pretty well, though.
Enjoy! |
|
