On Thu, 8 Oct 1998, Jackson, DeJuan wrote:
> So, has anyone tried the tutorials (with/with out success)? Let me know
> what did and didn't work for you.
I did not use tutorials too much since 1.0 or maybe even earlier, but those we have
now do not seem different to me. Those I tried before worked fine and were probably
more useful than other parts of documentation, especially in the beginning.
Occasionally, I use queries from tutorials as ready templates for something else (that
also includes the "Administrative Tasks" part of the manual), because I find it easier
to remember the task itself, rather than how it was done. I would contemplate
increasing the coverage of new features in tutorials and, like I already suggested
before, isolating example codes in a special sgml data element, so they could be
automatically processed and executed before distribution, in order to check for their
validity and such.
I no longer need that myself, but I was once disappointed to discover that the
tutorial on extended types, originally dedicated to boxes and polygons, disappeared
from the distribution because those types became built-ins. I don't think the complex
number type that is still there is sufficient for a novice to become a confident
extension writer. This type is too simple for that. I had to work all the way through
Hellerstein's GIST examples before R-trees started to make enough sense to me. Through
a fair amount of trial and error, I have written a number of useful objects and I am
willing to share the code, but I am afraid these will are too oversized for a
tutorial. The 2-dimensional box was just right. I would advocate putting it back,
under different name, so people could screw around with it without damaging their
databases. Because in my opinion, the whole subject of building extensions and
interfacing them to access methods is too complex to be just read about, without!
a hands-on experience.
--Gene
