Re: Modelling Considered Harmful

mAsterdam wrote:
> dawn wrote:
>
> > mAsterdam wrote:
> >
> >>Kenneth Downs wrote:
<snip>
> 2. Language does more than just model thoughts.

I can't actually think of any model that only models. When using a model, you get the benefits of using that particular model. So, it makes complete sense to me that you might claim of any model that it does more than model.

<snip>
> >
> > While it is a good idea to model for a purpose, I don't think that
is
> > essential to the definition of the model.
>
> Ah! Here we diverge. I think as soon as the model steps out of
> the scope of the art-students she stops being a model.

This wasn't a "dig in my heels" opinion, but the reason I said that purpose was not critical to the definition is that there needs to be room for the model itself to exist because the process of modeling had a purpose, rather than the model itself having one. We could call that the purpose of the model, if you want to stretch it. But if we have students in a grade school diagram a sentence (I wish they still did that -- my kids didn't) then we have a model of a sentence (which, with my earlier points would make it a meta model), then (are you trying to diagram THIS sentence!) and that model itself really has no purpose, but the process of modeling did have a purpose.

On the other hand, if you want a statement about purpose in the definition of a model, I can live with that.

> > Instead one might write "we
> > are using a model to ...". If I'm working with Lincoln Logs (which
> > might be differently named outside the US), my purpose might be
"play"
> > or even "beauty" or as a creative act. The modeling need not be
for
> > the purpose of studying something except perhaps in the very
broadest
> > terms.
>
> What then, makes it a model? BTW should I google for Lincoln Logs?

I did and it would be better to go to amazon.com, select Toys & Games for the category and type in "Lincoln Logs" for the search. Have you ever seen a similar toy? (just curious) We have quite a collection of these in our basement.

> >>>It should seem almost painfully obvious that the standard examples
> >>>of employees, sales orders, inventory activity
> >>>and so forth fit far more the definitions for "records"
> >>>than they do for "model".
> >
> > They fit both because records, themselves, are modeling something.
>
> Not by themselves, IMO. The model lives outside the records.

I still think they model propositions in a similar way to language modeling thoughts. There can also be an abstracted model that all such records use that does exist outside of any one such record.

<snip>

> This is, I think, the same difference as noted above.
> Just records cannot make up a model. They may be part of a model,
> but more is needed. Along the line of "characters are not language."

I think of records as having meaning, of modeling propositions. Modeling occurs at multiple levels.

> >>>Nor is the meta-data a model.
> >
> > of course it is!
>
> Same difference again or is there more to this?

same, I suspect.

> >
> >>Indeed. It's just the record-keeping of the record-keeping
> >>mechanism - registering the forms.
> >
> > So, it is a model about the model.
>
> At most a part, so not _is_.

That something is a model is rarely the only statement about it, as mentioned above. So, anything where it is, in part, a model is a model as far as I'm concerned. It can surely be other things as well.

> > I think the term "model" is central to all software development.
That
> > is what we do from start to finish and also what we produce. It is
> > both the process and object of our work. So, I'm definitely not
with
> > you

(I was referring to Kenneth here)

> on this one!
>
> Agreed.

Good deal. --dawn Received on Thu May 05 2005 - 18:50:02 CEST