Postgres idea list - Mailing list pgsql-hackers

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


> Do people want an advocacy article written, like "How to choose a
> database?"  I could do that.

That would be good, as would an updated "postgres vs mysql" article 
to point people towards. Or a "postgres myths debunked" page.

> Basically, I am open to ideas.  Would it help to fly me out to meet IT
> leaders?  More books/articles?  What does it take?  What do successful
> companies and open source projects do that works?

Since you asked, here are some ideas and thoughts I've been batting 
around:

1. Start an advocacy mailing list, to help coordinate publicity, responses 
to mySQL FUD, ways to advertise, etc.

2. Stop using the name "postmaster" as our daemon. Seriously. I've seen 
many a person, some new to *nix and some not, take a look at ps -Af and 
say "what the heck is that?" Whereas mysql uses "mysqld", cron uses "crond", 
ssh uses "sshd", and apache uses "httpd", we (postgres) use "postmaster." The 
name seems to imply something to do with email, and should be abandoned in 
favor of postgresd or postgresqld or even pgsqld.

3. Combine pg_hba.conf, pg_ident.conf, and postgresql.conf into a single 
file, postgres.conf. Clean it up and simplify it. Have a command-line tool 
to make changes. Have a way to test out the changes, similar to 
"apachectl configtest"

4. Fix the documentation. The interactive documentation on the website 
is particularly bad: try a search on "sequence", for example. The result 
is 65 matches, and each one a filename.

5. The website needs lots of improvement, on layout, navigation, and 
content. mySQL actually has this one right.

6. Moderate the lists better. There is a lot of traffic in general 
that should be going to other lists. Keep all the high-volume, nitty-gritty 
stuff on hackers, away from everyday users looking for help.

7. Stop underestimating mySQL. This is our competitor for the short-term 
at least, especially as both are open-source. Yes, we are better than 
mySQL on a technical level, but in all other areas they have us beat: 

* integration with other apps
* mindshare
* publicity
* ease of install
* ease of use
* documentation
* website navigation and appearance
* coolness

mySQL has the feel of an fun, open-source project. Postgres feels like 
a stuffy, academic project. At least that's the impression I get from 
asking people. All mySQL has to do at this point is improve their 
product, by adding things such as sub-selects and transactions. A tall 
order, but they are well on their way. We need to tackle all the 
items listed above. Not as easy, IMO, and we are not on our way.

8. Stop overestimating Oracle. Postgres is not a blip on their radar 
yet. We will probably never catch up to them. Focus instead on the 
shortcomings compared to our real rival (see above). Oracle should 
be emulated but not chased.

9. Have an easily accessible "todo" list that not only itemizes coding 
tasks, but documentation tasks, advocacy tasks, etc. so anyone can 
get involved and make contributions, no matter how minor.

10. Sign the source code (and other files) cryptographically. We are one of 
the last open-source projects that do not do this. What's to stop someone 
from breaking in to a mirror and replacing the tarball and md5 file? 
What if they did it on the main server? This is very easy to implement.

11. Consider an official name change to simply Postgres. Yes, there are 
historical reasons for this, but everyone I know ends up abbreviating it 
to postgres eventually anyway, and postgreSQL is a mouthful. 

12. Offer something "fun": a naming contest for the elephant (I know, 
I know), a bug squashing contest with prizes, a short interactive 
"find the best database for you" quiz, etc.

13. Solve the benchmarking problem. Find out what it takes to get us 
benchmarking to the same standards as the commercial DBs. Find a neutral 
third-party to compare Postgres and mySQL. Publicize our outstanding 
results. Start a debate on slashdot about it. :) Put the ball in mySQL's 
court for once.

14. Other things: Offer a bz2 download to save people time and $$. Put 
a favicon.ico on the site. Put in a site map. Consider using postgres.org. 
Publicize every little change as if it were the best thing since sliced 
bread. Solicit more lists like this. Release more often, even if more minor: 
stick to beta deadlines strictly. Offer success stories. 

15. Don't shoot the messenger. Some of this is my opinions, some is based 
on talking to "everyday users" and developers about Postgres.


Greg Sabino Mullane  greg@turnstep.com
PGP Key: 0x14964AC8 200206251441

-----BEGIN PGP SIGNATURE-----
Comment: http://www.turnstep.com/pgp.html

iD8DBQE9GL2ovJuQZxSWSsgRAgKLAJ9zBJhw0SzDu0eXUhGSPuncGXGGdQCgua14
IC2aWSjcSEHYxDU1hZXnZmA=
=GV5l
-----END PGP SIGNATURE-----