Re: Guidlines for a PostgreSQL Speech/Tutorial - Mailing list pgsql-advocacy
From | Scott Marlowe |
---|---|
Subject | Re: Guidlines for a PostgreSQL Speech/Tutorial |
Date | |
Msg-id | 1092779488.17763.16.camel@localhost.localdomain Whole thread Raw |
In response to | Re: Guidlines for a PostgreSQL Speech/Tutorial ("V i s h a l Kashyap @ [Sai Hertz And Control Systems]" <sank89@sancharnet.in>) |
List | pgsql-advocacy |
On Sat, 2004-08-14 at 13:10, V i s h a l Kashyap @ [Sai Hertz And Control Systems] wrote: > Hopeless and I thought I would get help. > Stupid me. > > > Dear all , > > > > I am an Electronics and Telecommunication engineer some how managed to > > come into > > development side. > > In PostgreSQL I have done a project which is related to Rural Banking. > > Here In India we have a annual Linux meet organized by one and only > > Asia's Linux magazine > > "Linux For You". > > I am planning to represent PostgreSQL people in this meet. > > But am stuck with following questions : > > 1. What stuff should I study the most. > > 2. What topics must I cover in the meet that will give PostgreSQL a > > big boost > > 3. Should I give a tutorial or just a speech about the features of > > PostgreSQL. > > 4. Scary topic for me is Object oriented part of PostgreSQL. > > 5. Their are many more questions but would ask them all once this > > thread grows. > > > > moreover I would be happy to receive guidelines and text of any > > previous PostgreSQL presentation from you kind people. > > > > The meet is in January 2005 > > > > With Warm Regards, > > Vishal Kashyap > > http://saihertz.com I'd be glad to help. I was on vacation over the weekend and didn't see your message. I hope it's ok with you if I take a vacation every now and then I know how much you pay me to sit here and answer your posts for you, oh wait a minute. Anyway... I've still got 600 emails from the lists I'm subscribed to that I need to slog through, so this won't be the long version. Here are some of the facts I generally cover when giving a postgresql speech: MVCC and the fact that postgresql indexes always point into the data store (i.e. you never get the value from just the index) means postgresql switches to seq scans quicker than other databases. PostgreSQL is quite programmable, with a dozen or so procedural languages. PostgreSQL is designed to have the power plug pulled while in the middle of transactions, and it should come back up coherent and ready to run in minutes. PostgreSQL's DDL as well as DML is fully transactable, except, of course, create / drop database. PostgreSQL is designed to handle massive parallel loads well, and can take advantage of large disk sets on RAID 1+0 / RAID 5 et. al. but tends to not be CPU bound much. aggregates tend to be slow for certain operations, and their work arounds (i.e. select id from table order by id limit 1 instead of max(id) ). I usually give a full blown demonstration, using the psql command prompt to show the simple yet powerful interface and what a user can do with it. Don't worry about administration so much, that's for later. I don't for the most part use any of the object relation stuff (i.e. inheritance) when using postgresql, as I'm a bit more of a relational purist. I also don't use arrayed types, since 99.9% of the time another table works better. There are some things online from Bruce's and Tom's and a few other folks presentations in the past at places like OSCON. However, PostgreSQL 8.0 just entered beta, so those guys are VERY busy. Take a look at the OSCON section of the O'Reilley site to see what's hanging there, and I'll see what I can dig up. But give the guys making this database a break for now, they're working hard on making it better and getting through beta, which is a breathtakingly fast moving process, leaving them little time to chat on the advocacy lists at the moment.
pgsql-advocacy by date: