I mentioned the lack of a schema builder mainly because it is the kind
of stuff that makes it attractive to all newcomers. If the resources
were available, it would be great to have a simple GUI to create the
XML file, and I believe that the number of users attracted to mondrian
will boost considerably.
Eventhough creating the schema file manually has a small learning
curve, it's the perception what usually determines wether to go with
it or not.
I don't think an intro to MDX is needed (there's plenty of tutorials
out there), but maybe a comprehensive how-to for mondrian schemas
would help the user to get an example on their own data up&running
I am not saying the XML schema files are difficult to learn, just
saying that the learning curve could be shorter if there were some
examples, a how-to or a few tutorials. I honestly think it would
attract more users that way.
Reading the forums I found out that a guy (Juan Gonzalez) wrote a PDF
with a tutorial that could be very useful to beginners, and it would
rock to have something like that available from the site.
As soon as I understand it well enough, I'll attempt to write one.
On 5/11/05, Kyle Cordes <kyle@...> wrote:
> >>Kyle Cordes wrote:
> >>What would really help:
> >>* A shorter path to try Mondrian out - ideally the distro
> >>* Tools, or at least guidance, or at least some links to
> Julian Hyde wrote:
> >Fair comment -- and I agree with you, that if all those resources were
> >available, Mondrian would be much more accessible.
> Well, I was describing an ideal world with unlimited resources :-)
> >If I might make an analogy, it would be like shipping an IDE and a
> >Java-programming tutorial with the JDK. It would be useful to beginners=
> Yes, certainly, it would be good, if all that great stuff existed, to
> keep it seperate from the core. I was just thinking of the "out of box
> experience" of someone going to the web site, click a big download
> button, and getting something ready to try out.
> >This project would be of immense value, but it would be a large undertak=
> >-- it is beyond my resources, I'm afraid. Typically this kind of project=
> You are absolutely right. However, I think that a scaled down version
> of all that wouldn't really be much more work than Mondrian is right
> now. Certainly if someone who understands ETL (not me, yet,
> unfortunately...) wrote a lengthy introduction to it, including that
> with the Mondrian distro / site would be trivial. If someone wrote an
> intro to MDX, that would be easy to include. The dev environment stuff,
> yes, you are right, would be a lot of work.
> >I expect that such projects/products will start to happen in the not so
> >distant future. Until then, I encourage you all to get involved and help=
> What would be great, is to get some of the momentum going around this,
> that is going around Jasper Reports. The big challenge, I think, is in
> getting "the masses" to realize some key points:
> * Getting a basic OLAP system running, isn't all that terribly hard, if
> you get Mondrian and some basic introductory material on OLAP.
> * It's generally much easier to make many commonly needed kinds of
> reports and analysis features, on an OLAP system, than on straight SQL
> to a production OLTP-schema database. It seems typical for a complex
> report (those that needs many lines of obtuse SQL, many joins, many
> Unions, many subqueries, and much performance struggle) to be
> expressible as a short, more clear bit of MDX. This is an immensely
> good thing, that many application architects are quite unaware of.
> We're working on using Mondrian in an application here, for the reasons
> above. We will certainly keep our eyes open for things to contribute bac=
> Kyle Cordes
> This SF.Net email is sponsored by Oracle Space Sweepstakes
> Want to be the first software developer in space?
> Enter now for the Oracle Space Sweepstakes!
> mondrian-users mailing list