Re: Adding structure, à la "multidimensional spread sheets"

[Jonathan DePeri <jonathan deperi gmail com>]

On Sun, Jul 18, 2010 at 4:19 AM, Andreas Guelzow <aguelzow pyrshep ca> wrote:
> Well, I had a look at the Quantrix website and if they have a useful
> product then they definitely have a bad marketing team since there intro
> essentially only causes yawns with respect to their simple pivot tables.
> Assuming they are showing their most exciting application, I would not
> find it useful to consider further.

Take a look at the "Tutorial", in addition to the intro. The intro
showed a relatively simple analysis, whereas the analysis in the
Tutorial got rather more sophisticated than that. If you can think of
an elegant way to do the analysis in the Tutorial -- especially if
different revenue/cost growth rate assumptions are used for different
regions or products, then I'd be very curious to know how you'd do it.
It seems like precisely the kind of problem I've dealt with in the
past where the lack of ability to impose structure makes formula
robustness and flexibility a nightmare.

> I would not deny that some users like to use the wrong tool for their
> problems. But just because some users use spreadsheets when they should
> be databases or statistical software does not make spreadsheets bad.

This is a fair point. The data I work with isalways stored in
databases, which means that a PivotTable often suffices for me. But
while it is easy to query data in relational format, it is hard to
work with them and do some kinds of multi-stage analysis. I'm a great
fan of the relational model for data storage and preparation, but I'm
just not sure how to use it effectively in actual data analysis or
financial model-building.

> Well, to me "Quantrix-like functionality" means reasonable data slicers
> (or pivot tables). Those are in the works. For me this has little to do
> with separation of presentation, formulas and data.
>
> Andreas

Maybe I am wrong and this is essentially all they have to offer. But I
think the Tutorial shows there is more -- or even just potentially
more. But 80/20 thinking might lead me to say you're correct, if the
potential user base for such features is small and we find it hard to
argue they're revolutionary. I think this is the fundamental question
-- is the idea just PivotTables or something bigger. I found out about
Quantrix because I started doing research after thinking about the
inefficiencies of spreadsheets. I experienced the flaws, reasoned
about them, and then tried to find products offering something better.
I'm not holding Quantrix out as a model, but just one example that
tries to address some of the flaws. I wish we could focus on
discussing the flaws themselves, not just Quantrix.

Jonathan