Thread: ...

...

From
Pedro Abelleira Seco
Date:
>> - Phppgadmin is a web based tool. You need a php
>> enabled web server. Most end users/admins don't
want
>> to have to configure a web server, PHP ("what is
>> PHP?") and to have a poor interface (I'm talking
about
>> web based interfaces in general, not the phppgadmin
in
>> particular).

> Maybe, but then you are platform independent.

With Java too. Well, you need Java (instead of PHP).
Java is not free and PHP is. I know there are people
for who this can be a problem. In my original post I
asked this, I'd like to know what developers think
about it.

>> - Both of them have limitations of what they can
>> manage. You can't use them to backup/restore the
>> database, to edit/see the postgresql configuration,
to
>> monitor the server(s), to start/stop server(s), ...

> Correct. You need another sort of tool for that. If
you go web based webmin
> can do most of this, or at least aims at this.

I'm looking for a complete solution. A tool that you
can use as your only interface to Postgres, if you
like. And I'm talking about a tool designed only for
Postgres. There are things that become bad if they
want too much generality. A Postgres admin tool.

>> To say it briefly if an average IT manager asks you
to
>> "show him PostgreSQL" and you open pgsql or
pgaccess
>> you are done. Sad but true.

> [I assume you mean psql with pgsql.]

Right

>
> Yes, but could show him that there are such tools
with Oracle too. Sqlplus
> is no better than psql for that matter. If you want
to add your commands
> inside a graphic tool you could use mpsql/kpsql
which are unfortunately not
> maintained anymore. Note, that I do not say the
tools are there for all your
> needs, but that I think there are quite some tools
worth extending. I don't
> think what we need is another tool that does parts
of the job, but an effort
> to build the one tool you can use for all of this.
If this has to start from
> scratch so be it. But maybe it's good idea to
improve some other tool.

Well, it would be great to pick code and ideas from
diferent projects, and to collaborate with them, but
my idea is:

1.- Have a tool for the administration of Postgres,
totally capable and integrated with the server.

In particular this means not to have to
download­compile hot-test a miriad of utilities, which
can have diferent dependencies and support for
versions of server or libs... Yes, a competent
sysadmin with enough time can do this right, but both
of this things are rare. And don't forget that a
sysadmin have to upgrade things from time to time, and
all of us know under what pressure you can be to do
it. If a sysadmin A with SQLServer can upgrade the
system in one day, only pressing "next" buttons and
you (B) are compiling, configuring, ..., your boss can
think how much is really costing that free database...
Not to mention what could he think about how much it
would cost to find another so competent guy if you
left.

2.- Have a beautiful and frienly face for those who
are used to other enviroments.

The manager and workers should feel confident in their
tools. This is a very subjective thing, not only
involving WAL's and transaction integrity. A blinking
cursor in a black screen is too much for many people.
Many, many people.
I'm yet impressed in how much difficult can be for
some people editing a config file. The world out the
University is very weird. And the corporative world
is, well, interesting.

3.- Have a well designed, extensible, interoperable,
mantainable, standard tool with a one to one relation
with Postgresql versións, easy to download precompiled
and ready to run.

A bit ambitious, and definitly I'm not that good. But
others are. We only need a clear goal about what is
needed in this area.

4.- Have a test case where the developers can inspire
about what things would be done in the server to easy
administration. 

Things like a complete interface for monitorizing
activity, have a configuration API, ... This things
can be used for other tools, if they want. The idea is
think about what is needed to provide a programming
interface for GUI's. The concrete tool is less
important.

Sorry for my English, and thanks for your atention.
Pedro

>
> Michael
> --
> Michael Meskes
> Michael@Fam-Meskes.De
> Go SF 49ers! Go Rhein Fire!
> Use Debian GNU/Linux! Use PostgreSQL!

_______________________________________________________________
Do You Yahoo!?
Yahoo! Messenger: Comunicación instantánea gratis con tu gente -
http://messenger.yahoo.es


Re:

From
Michael Meskes
Date:
On Wed, Jun 20, 2001 at 12:58:15PM +0200, Pedro Abelleira Seco wrote:
> With Java too. Well, you need Java (instead of PHP).
> Java is not free and PHP is. I know there are people
> for who this can be a problem. In my original post I

I do not care that much about the freeness of java, but be aware that due to
the lack thereof the newer java releases are not available on all platforms. 

I for one do not really like Java, because I have yet so Java developed
program to run stable and fast on my machine.

> 1.- Have a tool for the administration of Postgres,
> totally capable and integrated with the server.
> 
> In particular this means not to have to
> download­compile hot-test a miriad of utilities, which
> can have diferent dependencies and support for
> versions of server or libs... Yes, a competent

BTW how about adding some other tools that are available to the PostgreSQL
tree? The very same problem you have with all these tools will happen with
your tool unless you develop it as part of the PostgreSQL distribution. So
maybe we shoul try to merge in other tools too.

> I'm yet impressed in how much difficult can be for
> some people editing a config file. The world out the
> University is very weird. And the corporative world
> is, well, interesting.

Don't get me wrong, we really need that GUI so even the user we call DAU
(dümmster anzunehmender User = silliest user imaginable) can operate it
somehow. I'm all for your project. My problem is just that we need this tool
now or even better two days ago.

Maybe this sounds bitter, but I recently did a search for a groupware tool
to use in my company and also for a customer. We decided to go web based to
be independent of the OS. And of course we wanted open source. So I scanned
through sourceforge just to find at least a dozen such projects, but not
a single one finished enough to fill all our needs. 

While as a open source software developer I can understand why people set up
different projects, as a user I think this is a terrible waste of resources.

I hope you now understand what I meant to say.

Michael
-- 
Michael Meskes
Michael@Fam-Meskes.De
Go SF 49ers! Go Rhein Fire!
Use Debian GNU/Linux! Use PostgreSQL!