[BozemanLUG] Upcoming Seminar Announcement / Blatant self promotion

[Rusty Conover]

Hi All,

Sadly, I don't yet find myself above a little self promotion on the  
rare occasion, my apologies in advance to those who find this off topic.

I'd like to let you all know I'll be giving a seminar on September  
15th (next Monday) at 4:10 pm in EPS 108 as part of CS 500 titled:

"Finding an answer to the questions of: What is the best price?  Does  
it come in my size and color?"  A survey of the experience of building  
a scalable, vertically targeted price comparison platform.

While not 100% Linux/Free/Open Source oriented but you may find it  
interesting because it touches a good amount on PostgreSQL, Perl,  
LibXML and grid computing.

If you're able to attend I'd more then happy to see all of you there.

Abstract:

This seminar will provide a survey of the techniques learned, problems
faced and solutions found to various challenges when building the
price comparison website GearBuyer.com.  It will contain sections
regarding:

* Web crawling and distributed storage on the scale of millions of
urls and hundreds gigabytes using low cost commodity hardware
* Low overhead RPC techniques that increase transaction speed by
utilizing vectorization
* An implementation of a PostgreSQL data type for URLs that offers
better indexing locality then hashing while still conserving space
* An overview of fast data extraction techniques from XML using XPath,
XSLT and Javascript evaluation.

It may be useful to read this paper for background knowledge regarding
search engines:

http://portal.acm.org/citation.cfm?id=988392.988407

Why Writing Your Own Search Engine Is Hard
Anna Patterson
ACM Queue - Volume 2 ,  Issue 2  (April 2004)

Available for free here:

http://www.acmqueue.com/modules.php?name=Content&pa=showpage&pid=143

Thanks,

Rusty
--
Rusty Conover
InfoGears Inc.
http://www.infogears.com / http://www.gearbuyer.com
http://www.footwearbuyer.com