Обсуждение: Hi
Hi, despite bein chronically short of time I'd like to be involved with some database development in free software. It seems to me that postgresql is the most complete package out there and it also has some nice features like object identity. I consider myself fairly knowledgable when it comes to databases (at least it brought me my Ph.D.) so I think I could do some. But before I do so I'd like to ask some questions about the status of development: 1) Is there an embedded SQL preprocessor available? I read about an alpha version, but is it still maintained? 2) How about ODBC? I know there is a driver but apparently I cannot use all ODBC features. I have an application on Windows that I'd like to use to stress test postgresql a little bit, but I get an SQLError -1 everytime I try to fetch a record. Is this a knwon bug? 3) How would you describe the general goal of this project: Offer as good a database system as you can create as free software? That is should it be able to handle huge amounts of data, or is it seen more as a tool for a private person/small business? 4) Are there areas/Which ares need help? Please tell me what you think Michael -- Dr. Michael Meskes, Project-Manager | topsystem Systemhaus GmbH meskes@topsystem.de | Europark A2, Adenauerstr. 20 meskes@debian.org | 52146 Wuerselen Go SF49ers! Go Rhein Fire! | Tel: (+49) 2405/4670-44 Use Debian GNU/Linux! | Fax: (+49) 2405/4670-10
On Wed, 4 Feb 1998, Michael Meskes wrote:
> Hi,
>
> despite bein chronically short of time I'd like to be involved with some
> database development in free software. It seems to me that postgresql is
> the most complete package out there and it also has some nice features
> like object identity. I consider myself fairly knowledgable when it
> comes to databases (at least it brought me my Ph.D.) so I think I could
> do some. But before I do so I'd like to ask some questions about the
> status of development:
Welcome :)
> 2) How about ODBC? I know there is a driver but apparently I cannot use
> all ODBC features. I have an application on Windows that I'd like to use
> to stress test postgresql a little bit, but I get an SQLError -1
> everytime I try to fetch a record. Is this a knwon bug?
Julie (maintainer of PostODBC(src/interfaces/odbc)) has just
joined this list, and will hopefully be able to answer more completely,
but the PostODBC driver is currently listed as v0.21, so is still young...
Any aid you can provide in helping to mature it are most welcome,
both on our part, and, I imagine, Julie's...
> 3) How would you describe the general goal of this project: Offer as
> good a database system as you can create as free software? That is
> should it be able to handle huge amounts of data, or is it seen more as
> a tool for a private person/small business?
Both...We have at least two ppl so far, that I'm aware of, that
are dealing with systems of several GB of data through it, including one
of the core developers...like *BSD/Linux "competing" with the more
entrenched OSs like Solaris (and Microsloth), we're moving towards the
concept of being a viable alternative to such monolithic RDBMS systems
such as Oracle/Informix...but we are also working at making sure that we
don't loss site of the majority of our users, which tend to be ppl using
it for WWW based projects.
> 4) Are there areas/Which ares need help?
http://www.postgresql.org/docs/todo.shtml :)
There is a long list of things that are "Wish list" items...if you
decide to dive into something, all we ask is that you post something here
so that others know and to prevent "recreating the wheel"...just in case
someone else is already working on it.
> 1) Is there an embedded SQL preprocessor available? I read about an
> alpha version, but is it still maintained?
I remember mention of it on the list a while ago, but have not heard
anything since. Also, no code for this was contributed for the
distribution, and it should be included. It would be _great_ to have a real
preprocessor.
> 3) How would you describe the general goal of this project: Offer as
> good a database system as you can create as free software? That is
> should it be able to handle huge amounts of data, or is it seen more as
> a tool for a private person/small business?
As scrappy pointed out, we would like to not forget either kind of user.
However, afaik PostgreSQL is the closest to a "heavyweight" RDBMS available
in the freeware world, so if we had to choose we would probably put
ourselves on the "huge amounts of data" end of the spectrum.
> 4) Are there areas/Which areas need help?
The ToDo has lots of projects. Also, as you get started and learn the
details of Postgres internals the docs project could use some contributions
:).
There are other areas which are not yet called out in the ToDo list,
including
- recovery of munged databases (I think disk full conditions may cause
damage)
- transaction-only and session-only tables, variables, etc.
- re-think the table/type interface to allow more extensibility of types
But, picking simpler items from the ToDo is probably the best way to get
familiar with pieces of the code.
Welcome.
- Tom
> > Hi, > > despite bein chronically short of time I'd like to be involved with some > database development in free software. It seems to me that postgresql is > the most complete package out there and it also has some nice features > like object identity. I consider myself fairly knowledgable when it > comes to databases (at least it brought me my Ph.D.) so I think I could > do some. But before I do so I'd like to ask some questions about the > status of development: > > 3) How would you describe the general goal of this project: Offer as > good a database system as you can create as free software? That is > should it be able to handle huge amounts of data, or is it seen more as > a tool for a private person/small business? About a year ago, our goal was to make it more reliable. At this point, our goal is SQL compliance, speed, and features. See the TODO list in the distribution or on our web page for an idea of our open items. > 4) Are there areas/Which ares need help? What we really need are more people who can get a handle on the entire source tree, so we can implement some of the more complex SQL features, or improve some of the older code. -- Bruce Momjian maillist@candle.pha.pa.us
> About a year ago, our goal was to make it more reliable. At this point,
> our goal is SQL compliance, speed, and features. See the TODO list in
> the distribution or on our web page for an idea of our open items.
Good summary Bruce. We haven't emphasized long term goals, since these
shorter-term ones don't happen overnight. In the long run, though, other
topics of advanced interest, such as storage managers, distributed databases,
logging and error recovery, more OO features, adaptation to emerging
standards (SQL3 and beyond), certainly will be pertinent.
> > 4) Are there areas/Which ares need help?
>
> What we really need are more people who can get a handle on the entire
> source tree, so we can implement some of the more complex SQL features,
> or improve some of the older code.
We have just recently gotten to the point where a very few of the current
developers have an understanding of _most_ of the backend code. Achieving
longer-term advances require that the development be planned by folks with
that level of understanding, though others can and do contribute
substantially.
Everyone starts on Postgres without understanding the code. As we do small
projects, we learn as we go and become more capable of dealing with larger
problems. Enjoy.
- Tom
Thomas G. Lockhart wrote: > > - recovery of munged databases (I think disk full conditions may cause > damage) I have filled up the disk lots of times, but never had any problems with corrupted data. When inserting, the backend simply says: "mytable: cannot extend" and the insert fails. The only problem I've had is with SELECT and ORDER BY, where the sort files fill up the disk, and the backend either hangs or returns zero rows. This can of course cause serious problems if an application does something bad when a select returns zero rows (like re-initializing the whole database or something... ooooh noooo...). /* m */