Microsoft To Launch Homegrown Search Engine

Mr. Christmas Lights writes "While Google is currently the king-of-the-hill in search engines, Microsoft continues to lag in market share and uses Yahoo's technology/results. But Cnet reports that they'll launch on Thursday their own homegrown search engine , although it appears this is mostly a face-lift (despite a year of development and $100 million investment). According to Bill Gates, they 'will introduce a homegrown web crawler and algorithmic search engine ... later this year,' which is almost certainly their tech preview (you can look at this now) -- but will that be ready for prime-time in less than two months?"

Comments

About time

[lightdarkness]

I've seen their cralwer (their new one I presume) around for at least a month, without any indication on where the results were being shown. At least I have another spider to add to my list of robots that steal my bandwidth.

Re:About time

[b374]

Is this their crawler?

Re:About time

[FireFury03]

The msnbot has been around for many months. I have seen many complaints about the amount of bandwidth it uses and I know many web masters (me included) have blocked it's access because of this so I dunno how useful the search results will be. I've seen reports of it sucking gigabytes off a site in a day, and then doing exactly the same again the next day, which is really quite serious for those people who have a reasonably small bandwidth limit on their web space.

For me it was sucking several gig a month off my site, and was obviously very badly coded since it was refetching the same pages over and over (cachable pages, non-cachable pages and 404's). So in the end I gave up and outright blocked the damned thing - yet another bit of shoddy MS code out to break the internet.. :(

Re:About time

[FireFury03]

It pays attention to robots.txt directives (finally, a small amount of standards compliance!)

Re:About time

[AndroidCat]

It looks like it checks for meta tags too. (Useful when /robots.txt isn't convienent.) MSNBot page and other info. Also note that it only checks /robots.txt once a day, so policy changes might not take effect right away.

Re:About time

[FireFury03]

It payed attention for me with:

User-agent: msnbot
Disallow: /

iptables -A INPUT -p all -s 65.54.0.0/16 -j DROP

Or even better, if you have the TARPIT module:
iptables -A INPUT -p tcp -s 65.54.0.0/16 -j TARPIT :)

So

[jbartone]

Who's actually going to use this instead of Google?

Re:So

[MoonFog]

Since it'll probably end up being default start-up page in IE, lots.

Re:So

[Draveed]

You would be amazed. This week I discovered someone in my office who knew nothing about google.

Re:So

[zakezuke]

Since it'll probably end up being default start-up page in IE, lots.

Fortunately IE has enough in the way of exploits so the default start-up page gets hijacked often enough.

Re:So

[Technician]

Since it'll probably end up being default start-up page in IE, lots.

You mean the same people who use the default favorites? I looked at the default list once, then deleted it. It looked like a paid list from the yellow pages of the travel and media sections in the phone book.

Re:So

[Titusdot Groan]

Did your company fire them?

I've got a running gag with my team that if I can find the answer to their problem in 3 google searches I get their pay for that week. The number of dumb question I get is WAY down.

Search for Windows.....

[alistair]

No Results Found

Needs some fine tuning before it's ready for the prime time, me thinks.

Microsoft's problem

[ArbiterOne]

Microsoft is branching out too much. Without ripping off Google, I don't really see how they can pull this off. In order to reverse the current trend in market share, they'd have to have a better algorithm than Google, a massive ad campaign, and the popular opinion on their side. Oh, and start giving things away for free (Google: Blogger, Picasa, etc.)

Re:Microsoft's problem

[irc.goatse.cx troll]

Yeah! And whats with AOL using AIM as the default messaging service? And aol's mail for the default email? Its so monopolistic! They should license YIM and embed it into aol to be fair to their competition.

Seriously, its their browser, why shouldnt they make the homepage their search? Moz's default is a moz branded google, how is this different?

Prediction for most popular queries on day 1...

[seanellis]

"mozilla firefox" download
google

Algorithmic search engine? WTF?

[jagripino]

Don't *all* search engines have to have, hmm, some kind of algorithm in them?

Marketing speak confuses me! Please stop!

this article is oooooold

[evil_one666]

This article is from june 30th

Comment removed

[account_deleted]

Comment removed based on user account deletion

Doubtful this will take any ground.

[Ambient_Developer]

I say it time and time again.. Microsoft is not a company of innovation (besides user interfaces), they are a company that aquires other companies. It is doubtful that a home-grown engine will beat the likes of google.. Especially being so late in the game, not only that what good will a face lift do? Google is already one of the easiest things out there, how can Microsoft make search even easier? THAT is the 100 million dollar question!

Tech Preview?

[beders]

No results for my name or a random word. More worryingly no result for "porn". Got a long way to go chaps, $100 million seems a little steep for a input box linked to an error page...

Dutch?

[tompercival]

Fantastic... dutch search results - just what I was after too.

Best search engine

[Libor Vanek]

Just try to search "best search engine" and enjoy what comes out:
http://www.search.msn.com/results.aspx?FORM=SRCHWB &q=best%20search%20engine

Re:Best search engine

[jimicus]

That looks remarkably like a Google results page, in terms of structure rather than content.

Example:

http://www.search.msn.com/results.aspx?FORM=SRCHWB &q=best%20search%20engine
http://www.google.com/search?hl=en&ie=ISO-8859-1&q =best+search+engine&btnG=Google+Search

Surely Microsoft haven't chosen to rip-off their design based on the market leader?

lack of trust

[vinsci]

Microsofts search engine lacks the most important feature: trusted results.

In the past, it has been shown that Microsoft blocks search results that are contrary to its own business interests.

wtf?

[Mr_Silver]

although it appears this is mostly a face-lift (despite a year of development and $100 million investment)

I thought it was only marketing that didn't understand that just because it looks the same, doesn't necessarily mean you've done nothing under the hood.

THE bot?

[knipknap]

I wonder whether that's the bot that has been scanning my website for three days by attempting to "crawl" through all session ids and causing more then 1 GByte of traffic.

"msnbot/0.11 (+http://search.msn.com/msnbot.htm)"

It was only stoppable by blocking the IP. (robots.txt was only read once before it started) Great, smart bot, really.

3 bad results.

[caluml]

Orange. No results for Orange, the mobile phone company.
Linux. No pointers to linux.org.
Google. Returns the Dutch/Belgian version of the page. Why?

Re:3 bad results.

[BuilderBob]

Orange. No results for Orange, the mobile phone company.
Linux. No pointers to linux.org.
Google. Returns the Dutch/Belgian version of the page. Why?

These are no longer true. I know it used to do this but now ...

'Orange' returns Orange.co.uk.
'Linux' returns linux.org
'google' returns google.com
'microsoft sucks' returns fuckmicrosoft.com
'abu graib' returns the photographs of inside the prison.
'lindows' returns lindows.com

This is from Firefox 0.8 on Redhat Linux.

BB

Re:3 bad results.

[caluml]

I ran those 3 searches a minute before I posted them. If they used to be true, they still are. I am in the UK. Maybe it's messed up over here?

Filtering out queries

[ttys00]

Will they be filtering out queries with this engine as well (eg. xfree86 being filtered as discussed here a while back)?

Of course. And while they do that, I won't be using it.

Search Results

[chia_monkey]

Oh my. I can just see the search results now:

Search entry: "Antarctic Penguin"

Search Result: A paperclip pops up on your screen and says "It appears you are searching for a penguin. Did you know Microsoft servers are cheaper to run than Linux? Would you like to buy one now?"

No faith in their own product

[BlackHawk-666]

I notice they're using Akamai instead of a cluster of cheap Windows servers. Nice of them to recommend to everyone else to use their technology, but then not trust it for their own stuff.

subversion:~# telnet techpreview.search.msn.com 80
Trying 213.253.9.73...
Connected to a213-253-9-73.deploy.akamaitechnologies.net.
Esca pe character is '^]'.
HEAD /

HTTP/1.0 400 Bad Request
Server: AkamaiGHost
Mime-Version: 1.0
Content-Type: text/html
Content-Length: 161
Expires: Wed, 10 Nov 2004 12:56:31 GMT
Date: Wed, 10 Nov 2004 12:56:31 GMT
Connection: close