sdavis2@[EMAIL PROTECTED]
("Sean Davis") writes:
> On Thu, Mar 27, 2008 at 2:13 PM, Martin Marcher <martin@[EMAIL PROTECTED]
>
wrote:
>> Hi,
>>
>> I'm investigating a Database with about 60 tables and 40 functions.
>> It's not that much but I'd like to be able to do***entation to it
>> (javadoc, doxygen, anything don't care as long as it runs on linux).
>> On pgfoundry I ran over http://pgfoundry.org/projects/autodoc/
but
>> that is quite undo***ented itself.
>
> Have you looked into the "comment" SQL command? It will let you
> do***ent to your heart's content all the pieces of your schema. You
> can then use SQL queries to generate "re****ts" that can be used to
> produce html do***entation, etc.
This actually "plays into" what autodoc can then do...
- The HTML + DocBook generator will extract those comments, and put
them into the schema descriptions
- What is, regrettably, less clear about autodoc is what to do with:
a) The .dia files, which take table definitions and put them into
the XML format used by the GNOME tool, "dia", which is something
of a Visio "clone."
Unfortunately, it doesn't do any autoarrangement of the UML
objects that it uses, so making that look "pretty" will take a
fair bit of hand work.
b) The .dot and .neato files, which may then be processed using
the AT&T tool, GraphViz, which *does* make a serious attempt to
turn them into well-laid-out diagrams.
(Aside, for Rod Taylor's benefit... I have added to my
share/postgresql_autodoc/dot.tmpl and
share/postgresql_autodoc/neato.tmpl files the line:
page="8.5,11"
That tends to allow it to generate better layouts in cases where
the objects don't fit well on a single page...
There would be some value to having some command line examples of
how to get .ps/.pdf output from GraphViz.
--
(format nil "~S@[EMAIL PROTECTED]
" "cbbrowne" "linuxfinances.info")
http://linuxfinances.info/info/unix.html
Science is like ***: sometimes something useful comes out, but that is
not the reason we are doing it. -- Richard Feynman
|
