Re: Ping: dawn, some mvl questions
Date: 22 May 2006 14:09:04 -0700
Message-ID: <1148332139.265535.172550_at_i39g2000cwa.googlegroups.com>
mAsterdam wrote:
> dawn wrote:
> > mAsterdam wrote:
> >
> >>Kenneth Downs wrote:
> >>
> >>>...
> >>>I am in the school of thought that says that if you need a list, make a
> >>>table, that's what the RM is for.
> >>
> >>(Not about MV but sets/lists)
> >>
> >>I am in school of thought that says that if you see a list, it might
> >>be set that is meant. People often don't need to distinguish
> >>between them for day to day purposes,
> >
> > <snip>
> >
> > Not only that, but it is impossible to enumerate a set without the
> > representation being a list. We write in lists, we speak in lists, and
> > we are sometimes unaware of the meaning we give to a set when we list
> > it. Does a grocery list refer to a set or would you lose something if
> > you treated it that way?
>
> Hm... I sound like a broken record. It depends on the context.
Agreed.
> Scenario:
> Say you are going to shop way to late (it's almost shop-closing time).
> You know you won't be able to get all items on the list.
> You are throwing a party tonight, and your partner (who always says:
> "First things first") wrote the list. Sure you'll be selective, and
> weighing necessities for the party (forget about the toothpaste and
> floorwax) - but for two items (/members) having the same party
> weight, having to guess, I'ld treat the list as a list, and look for
> the items in order. Then again if I found one item (5 baguettes)
> and another item way down in the list (2 paper table cloths) is
> in the immediate vicinity of the found one I'ld grab it anyway.
>
> > Retaining the order of something represented as a list might just
> > provide ongoing information never verbalized.
>
> Yes. You can't tell from just looking at the content.
I agree. However, one person decides to list something and someone else decides that since they were not explicit that the order had meaning, they would simply lose the information of the original ordering, that wouldn't be good either. When in doubt, retain the original order. If designing (LDM) with the RM, there is a tendency to assume order is meaningless unless the user gives an explicit requirement to retain such ordering. If designing with a model that permits LVAs there is a tendency to retain the ordering even if the user says it is irrelevant. It would be good to specify one or the other (or bags for the rare case, perhaps), but if there is no certainty, then we would want to err on the side of retaining the information provided by the user when they order the data.
> > If a user lists
> > something in an order, but we have defined it as a set because there
> > has been no overt statement of the meaning of the order, might be
> > losing information?
>
> Yes. OTOH, you might be adding misinformation be assuming
> the order relevant. My point was (and is):
Agreed that it would be misinformation if retained and the implication is that the ordering matters.
> You can't tell from just looking at the content.
> You'll have to investigate.
Agreed.
> P.S.
> Giving the scenario an unrealistic twist:
> Say you and your partner share the shopping data using
> a PDA-app where groceries are treated as a relation:
> now you know the order in the representaion is irrelevant.
> You don't have to guess, but you also have no guidance
> from the order of the grocery-list.
And perhaps the order is relevant to the person writing the list who is able to glean some additional information from the way they ordered it, but not relevant to the other person. There is no harm in the other person receiving the data the way the original person ordered it, but there is something to be gained by the person who had the brain that ordered it in the first place, getting that order back. There really might be some subtle, but real, productivity gains by users who enter data and have their data retain their original ordering if they ask for it again without specifying a new order. They can still do the sort by alpha or whatever other ordering, but no information, whether understood to be relevant or not up front, was lost if the data were modeled as a list.
Cheers! --dawn Received on Mon May 22 2006 - 23:09:04 CEST