~drizzle-trunk/drizzle/development

Hi!

* So why?

I've been wondering for a long time now about all of the changes post 4.1.

I believe there is a large market of users who never wanted them, and never

cared for them. I also wanted to question the foundations of what we built.

Do users want wrong data? How often is the query cache really valuable? If

everyone just has a root user with all privs, why carry the baggage of the

ACL code. Etc.

* What is the goal?

A micro-kernel that we then extend to add what we need (all additions come

through interfaces that can be compiled/loaded in as needed).  The target

for the project is web infrastructure backend and cloud components. 

* Is this a product of Sun/MySQL?

No, though several of the authors do work for Sun/MySQL. The development model is

one based around open collaboration.  Drizzle's license is the GPL v2.

* So what are the differences between is and MySQL?

No modes, views, triggers, prepared statements, stored procedures, query cache, 

data conversion inserts, ACL. Fewer data types.  Less engines, less code. 

Assume the primary engine is transactional.

* Why now?

Why not now :)

* "This is awesome, but I need you to add back..."

Forget it. Nothing is going back in at this time. As for the future? Maybe,

but at the moment this is not the target. If you want more features, go

use MySQL.

* What platforms?

I compile on OSX, Linux (Fedora), and Solaris Express. Windows is not

supported and will stay that way unless it gets a working posix layer +

autoconf system (aka becomes a platform that is reasonable to support).

Really, get a working GNU chain going on Windows and you can get Drizzle

working on it.

Drizzle relies on a C99 compliant compiler. Please do not ask us to support

older hardware, compilers, or OS'es.

* What is the future?

Less code, more modularity, and more active involvement. A few things that

still need to be done:

- Switch to stdint

- Walk through all of the replication code

- Re-implement information schema

- Modular logging system

- ...

* "This is not a SQL compliant relational..."

Very true, and we do not aim to be that. 

* What is left to be cut out?

Please ask on the mailing list or on IRC.

* Can I get involved?

Most certainly. There is plenty to do from refactoring code, design of interfaces, 

documentation and blueprints, etc... The best way to get involved it to join the 

mailing list at:

https://launchpad.net/~drizzle-discuss/

If you wish to suggest a refactoring project or an interface please email

the mailing list and keep an open mind. Do not expect anyone will work on

your idea though, you may influence someone to do that, but more then likely

you will need to rollup your sleeves and write some code. For very simple

bits you are welcome to ask others on #drizzle on Freenode, but please be

aware that you may be asked to email the mailing list. 

Showing up with a big block of code is probably the worst way of getting

your work accepted. This is unlikely to work. 

Right now we use a simple captain system for commits. Anyone can send in a

proprosal for merge via launchpad but your changes may be flowed first

through someone who has been around long enough to understand code

requirements to review your code. This system is based entirely on trust,

and individuals who have shown that they can provide three good patches gain

credibility. Starting small is fine, patches that are just comments or are

even two or three line cleanups are welcome and encouraged. I would really

recommend that anyone who wants to work on something first start with something 

of this size. Patches like these are valuable and teach you how to work with

the system.

The general rule is no new code in the core of the server, and any changes

to interfaces need to be code line neutral. AKA if you want to add an

interface you need to be able to remove at least the number of lines of code

you added. This is a rule of thumb, and does not apply to code cleanup.

It should be pointed out that we are more focused on code style,

performance, and over all maintenance then we are features.

* What is the coding style?

Please look here:

https://blueprints.launchpad.net/drizzle/+spec/style-cleanup

If you have a question on this, please email the mailing list so that we can

clarify the document. And when you get the answer? Please update the code

style document :)

* Should I email the mailing list with patches?

Please god no, we live in the age of the Internet :)

Create an account on launchpad.net. Create your own tree and let one of us

pull from it.

* What is the target?

Deliver a microkernel that we can use to build a database that meets the

needs of a web/cloud infrastructure. To this end we are exploring http

interfaces, sharding enhancements, etc... do not expect an Oracle, MySQL,

Postgres, or DB2. 

There is no GA date at the moment.

We are focusing on multi-core architecture. This is not designed to run on a

wrist watch (hint, go use SQLite). We support both 32bit and 64bit but the

class of machine we are targetting is 64bit. We are making design decisions

which assume very large amounts of RAM will be made available to the DB.

* Can I run a website with this?

No. We are still making incompatible changes, and I certainly do not believe

the code is production quality. Right now we are defaulting many configure

operations to generate debugging code for us so our binaries are not

optimal.  Therefore, do not go out and benchmark this and expect it to be one way or

the other. We are currently only doing benchmarks where it makes sense for

us to determine where bottlenecks are.

* Why drizzle?

I am from Seattle. Drizzle is our normal form of "rain" but it is not "rain",

it is drizzle. This was a bit of a rainy day project that finally found a

spot in my schedule :)

  -Brian

    Seattle, USA

    Sun, Microsystems